ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Social Sciences 144

Social causality in motion

Visual bias and categorization of social interactionsduring the observation of chasing in infancy

MARTYNA A. GALAZKA

ISSN 1652-9030ISBN 978-91-554-9943-3urn:nbn:se:uu:diva-321904

Dissertation presented at Uppsala University to be publicly examined in Auditorium Minus,Gustavianum, Akademigatan 3, Uppsala, Thursday, 31 August 2017 at 10:15 for the degree ofDoctor of Philosophy. The examination will be conducted in English. Faculty examiner: Prof.Dr. Moritz Daum (Universität Zürich, Psychologisches Institut Entwicklungspsychologie).

AbstractGalazka, M. A. 2017. Social causality in motion. Visual bias and categorization of socialinteractions during the observation of chasing in infancy. Digital Comprehensive Summariesof Uppsala Dissertations from the Faculty of Social Sciences 144. 101 pp. Uppsala: ActaUniversitatis Upsaliensis. ISBN 978-91-554-9943-3.

Since the seminal work of Fritz Heider and Marienne Simmel (1944) the study of animacyperception, or the perception and attribution of life from the motion of simple geometrical shapeshas intrigued researchers. The intrigue for psychologists and vision scientists then and todaycentered on the stark disconnect between the simplicity of the visual input and the universalrichness of the resulting percept.

Infant research in this domain has become critical in examining the ontological processesbehind the formation of animated percepts. To date, little is known about how infants processthese kinds of stimuli. While numerous habituation studies have shown sensitivity to animatemotion in general, none to date has examined whether infants actually perceive animate displaysas social interactions.

The overarching goal of the present thesis is to answer this question and further augmentknowledge about the mechanisms behind the formation of animated percepts in infancy. I, alongwith my collaborators, do so in three ways, in three separate studies. First, we examined visualattention during online observation of randomly moving geometrical shapes in adults and infants(Study I, using eye tracking). Second, we examine distribution of visual attention in infancyduring online observation of non-contact causal interactions, focusing on the most ubiquitous,fitness relevant of interactions – chasing (Study II, using eye tracking). Third, we answer thequestion whether infants perceive social content in chasing displays by measuring the neuralcorrelates in response to chasing (Study III, using EEG).

The collective contribution of the present work is also three fold. First, it demonstrates thatstarting at the end of the first year of life, human visual system is sensitive to cues that efficientlypredict an interaction. Second, at 5-months infants begins allocating attention differently acrossagents within interactions. Finally, attention to specific objects is not due to low-level saliencybut the social nature of the interaction. Subsequently, I present the case that perception of socialagents is fast, direct, and reflects the workings of a specialized learning mechanisms whosefunction is the detection of heat-seeking animates in motion.

Keywords: social causality; motion; animacy perception; chasing goal-directed motion; heat-seeking; EEG; P400; Nc; spatial proximity; non-contact causality; functional specialization;specialized perception; evolution

Martyna A. Galazka, Department of Psychology, Box 1225, Uppsala University, SE-75142Uppsala, Sweden.

© Martyna A. Galazka 2017

ISSN 1652-9030ISBN 978-91-554-9943-3urn:nbn:se:uu:diva-321904 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-321904)

For my babcia

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Galazka, M., & Nyström, P. (2016). Visual attention to dynamic spatial relations in infants and adults. Infancy, 21(1), 90-103.

II Galazka, M., & Nyström, P. (2016). Infants’ preference for individual

agents within chasing interactions. Journal of Experimental Child Psy-chology, 147, 53-70.

III Galazka, M., Bakker, M., Gredebäck, G., & Nyström, P. (2016). How

social is the chaser? Neural correlates of chasing perception in 9-month-old infants. Developmental Cognitive Neuroscience, 19, 270- 278.

Reprints were made with permission from the respective publishers.

Contribution

The contributions of Martyna Galazka to the papers included in this thesis were as follows: Studies I and II: Designed and planned the study in collaboration with su-pervisors and the co-author. Collected eye-tracking data, performed the sta-tistical analyses, had the main responsibility of writing and revising the manuscript in collaboration with the co-author. Study III: Designed and planned the study in collaboration with supervisors and co-authors. Created animated stimuli, was responsible for collecting the EEG data, performed the statistical analyses, and had the main responsibility for writing and revising the manuscript in collaboration with the co-authors.

Abbreviations

A-I distinction Animacy-Inanimacy distinction EEG Electroencephalography AOI Area of Interest STS Superior Temporal Sulcus pSTS pre- Superior Temporal Sulcus ppf pixels per frame AA Agent Ahead AB Agent Behind D Distractor fps frames per second BMS Baysian Model Selection V5 Visual area 5 HSP High Spatial Proximity LSP Low Spatial Proximity PL Point light N170 negative ERP component 170ms post stimulus P400 positive ERP component 400ms post stimulus Nc negative central ERP component NIRS Near-Infrared Spectroscopy fMRI functional Magnetic Resonance Imaging ERP Event-Related Potential

Contents

Introduction ..................................................................................................... 9Observing motion ..................................................................................... 10Causality ................................................................................................... 12

Perceiving physical causality ............................................................... 12Perceiving social causality ........................................................................ 17

Examination of motion cues in the formation of animated percepts ... 18But do we really perceive animate motion? ......................................... 21Neural correlates of animate motion .................................................... 23Neural correlates of animate motion in infancy ................................... 24

Chasing ..................................................................................................... 26Chasing perception in adults ................................................................ 26Chasing perception in infancy ............................................................. 28Neural correlates of chasing ................................................................. 30

Theoretical considerations ........................................................................ 31Modularity ............................................................................................ 32Information-processing ........................................................................ 33The possibility of functional specialization ......................................... 34

Aims of the thesis .......................................................................................... 36

Methods ......................................................................................................... 38Participants ............................................................................................... 38Stimuli ...................................................................................................... 39

Video presentations .............................................................................. 41Test stimuli – Still images .................................................................... 42

Apparatus .................................................................................................. 42General procedure ..................................................................................... 43Data analysis ............................................................................................. 44

Study I – Visual attention to dynamic spatial relations in infants and adults48Design ....................................................................................................... 49Results ...................................................................................................... 49Discussion Study I .................................................................................... 51

Study II – Infants’ preference for individual agents within chasing interactions .................................................................................................... 53

Design ....................................................................................................... 53

Results ...................................................................................................... 55Experiment 1- Chase, Follow, No-interaction ..................................... 55Experiment 2 – Chase - No Switch, Follow - No Switch, No-interaction-Chase, No-interaction-Follow ........................................... 57Experiment 3 – Chase-no initiation, Follow-no initiation ................... 58

Discussion Study II ................................................................................... 60

Study III- Neural correlates of chasing motion ............................................. 62Design ....................................................................................................... 62Results ...................................................................................................... 64

Experiment 1 – Chasing vs. Inanimate motion .................................... 64Experiment 2 – Chasing vs. Random motion ...................................... 65

Discussion Study III ................................................................................. 67

General discussion ......................................................................................... 69Summary of findings ................................................................................ 70Functional specialization .......................................................................... 73

The function of specialized perceptual biases ..................................... 73Functional specialization and modularity ............................................ 74

Visual attention ......................................................................................... 75Event related potentials ............................................................................ 76Limitations ................................................................................................ 77Future directions ....................................................................................... 79Final conclusions ...................................................................................... 81

Svensk sammanfattning ................................................................................. 82

Streszczenie w języku polskim ..................................................................... 84

Acknowledgements ....................................................................................... 86

References ..................................................................................................... 89

9

Introduction

One of the most essential tasks of human cognition is differentiating be-tween living and non-living entities. Such distinction serves as the basis for identifying conspecifics, allocating attentional sources, forming adaptive interactions - all that consequently foster social learning (Cosmides & Tooby, 1992; Tooby, Cosmides, & Barrett, 2003). Recognizing one’s social partner as such, further allows for interpreting their behavior within a psy-chological framework and conceptualizing it in terms of abstract ideas such as agency, intent and goals. In short, it permits for a more profound interpre-tation of behavior beyond simple motion mechanics.

Animates, our most typical social partners, are defined as entities en-dowed with an internally-controlled source of energy, made of biological matter that honor principles of basic biology. Modern research has shown that the ability to categorize entities into animates and artifacts is present in diverse and remote native cultures (Atran, 1999), across a range of lan-guages (Diesendruck, 2003), is associated with distinct neurophysiological correlates (Saffran & Schwartz, 1994; Caramazza & Shelton, 1998), and is central to a broader array of more complex abilities such as word learning (Childers & Echols, 2004), attribution of mental states (Baron-Cohen, 1995) and understanding of biological processes (Carey, 1985).

Taking from the principles of evolutionary biology, contemporary theo-retical position proposes that our visual system has evolved a functionally specialized phenotype for the detection and monitoring of animates in mo-tion (Barrett, 2006; Barrett & Kurzban, 2006; New, Cosmides, & Tooby, 2007). Animates are not only vital static features of the visual environment, but they tend to change their status more often than do plants, artifacts or other features of the terrain, such as rocks and mountains. Animate motion in particular has a quality of function or purpose. It is not confined by laws of mechanics, and it responds to changes in the environment in order to adapt to new circumstances (Gelman, Durgin, & Kaufman, 1996). Animates follow frequently changing trajectories toward frequently changing goals. Their motion may be smooth or rigid, while the manner in which the move-ment initiates may not always be evident. Thus, the constant re-inspection, and online processing of movement may be equally if not more, as important as the initial detection of an animate.

The present thesis reports two studies designed to examine visual atten-tion during online observation of motion, and one study examining neural

10

correlates in response to animate motion. In Study I we examined both adult and infant attention during online observation of random motion. This was done to determine visual bias to motion cues associated with an interaction such as the gradual reduction of spatial distance between moving objects. In Studies II and III we focused on the most relevant, in the evolutionary sense animated percept, a chasing interaction (Frankenhuis & Barrett, 2013), in which one object moves toward another in a direct path while its target speeds up to move away. Chasing interaction, as I will discuss toward the end of the Introduction, is useful in the study of animacy perception in in-fants because it is a relatively simple non-contact causal relation, which has been found to result in rich percept of animate behavior in adults. Study II of the present thesis reports how infants distribute their attention during online observation of chasing. Finally, in Study III we examined whether neural correlates to a chasing object differ from those associated with non-interacting or inanimate objects.

Observing motion Unlike color or binocular vision, the ability to see motion is present in all species and is the oldest and most basic visual capability (Nakayama, 1985). Even newborn infants seem to have a special affinity for motion as their visual field range extends when presented with dynamic versus static stimuli (Tronick, 1972). Physically, the perception of motion begins when the cor-nea and then the lens of the eye focuses an image onto retina. It is there that the pattern of light gets transduced into neural impulses, which get transmit-ted to central ganglia in the brain. From then, motion perception proceeds via two pathways: the primary visual pathway through the lateral geniculate nucleus (LGN) to the primary visual area V1, to V2 and finally to MT+/V5. The second pathway follows through the superior colliculus (SC) and the pulvinar or LGN, and is found to dominate infant perception of motion. This latter pathway is suggested to be phylogenetically older motion perception system, highlighting the fundamental importance of motion perception be-fore proper maturation of cortical pathways are set to occur after 5 months of age (Rosander, Nyström, Gredebäck, & Hofsten, 2007).

Beyond physicality, the outstanding question in regard to motion percep-tion concerns the integration of information during the formation of per-cepts. One efficient and empirically constructive way of examining how the formation of animated percepts takes place is to strip down the visual scene to the proverbial ‘bare bones’ of the visual input, exclude morphology as a factor, and focus on how the entity interacts with the environment. Such as during the observation of motion of simple two-dimensional geometrical shapes - a method initiated by Heider and Simmel in 1944 and since used prolifically by psychologists, vision scientists and developmentalists alike.

11

The success of this method in studying animacy relies simply on the fact that animates engage in a rich spectrum of social interactions, while artifacts do not.

To date, literature is replete with research examining adult reactions to simple visual displays. Almost uniformly, adult observers derive rich per-cepts of animacy in reaction to relatively limited visual input (Castelli, Happé, Frith, & Frith, 2000; Gergely, Nádasdy, Csibra, & Bíró, 1995; Heider & Simmel, 1944) often, in spite of being aware they are watching lifeless geometric figures (for review see Scholl & Tremoulet, 2000). In recent years, studies focused more and more on identifying neural correlates pertaining to animacy perception (Blakemore, Boyer, & Meltzoff, 2003; (Martin & Weisberg, 2003; Schultz, Imamizu, Kawato, & Frith, 2004, for review see Heberlein, 2008) with equally uniform findings.

But while the field has benefited from advancing methods of inquiry, cur-rent literature is still lacking in several important aspects. First, there is still a limited number of findings examining perception of animacy from a devel-opmental perspective – an area that bears critical implications regarding ontology. Infancy research in general provides its own set of challenges, mostly because one cannot explicitly ask what it is that is being perceived. These challenges, however, also provide a way of avoiding some common pitfalls that are characteristic of adult research, a point I return to later in the thesis. Second, there is a continued lack of studies examining oculomotor behavior as a proxy for visual attention during the online observation of motion in both adults and infants. Without knowing how our visual system prioritizes motion information, mechanisms behind the formation of animate percepts are based on conjecture at best. Here, current advances in eye track-ing technology and data analysis from our lab help to fill in this important gap. Finally, there is a need in examining how the developing nervous sys-tem categorizes animated percepts based on motion. While there are few elegant studies on animate motion in young infants (Kaduk et al., 2013; Poulin-Dubois, Lepage, & Ferland, 1996; for review see Scholl & Tremoulet, 2000), we provide initial studies that show how infants catego-rize animate interactions and how this categorization translates on neural level.

Prior to presenting the empirical studies, the subsequent sections of the Introduction provide necessary background. First is the discussion on physi-cal and social causal events, with special focus on the perceptual processes. The discussion then turns to literature on chasing in particular as a type of non-contact causal event. Toward the end of the introduction two theoretical perspectives are introduced, as well as the overarching theoretical and em-pirical strategy for solving some of the remaining issues regarding the ontol-ogy of the percept.

12

Causality Understanding causality is important for determining the relationship be-tween any two objects and how one can bring about change in the other. Importantly, perceiving causal structure can occur independent from know-ing the status of the objects involved, but depending on the type of causal event, it allows to infer the status of the entity in it (Scholl & Tremoulet, 2000).

There are two types of causal events, which have been subject to count-less empirical and philosophical inquiries. In one scenario (Figure 1(a)) ob-ject A begins to move toward the stationary object B, contacts it, and object B begins to move in a straight line opposite of impact while object A stops. In the second scenario (Figure 1(b)), object A moves toward object B but before they touch, object B begins to move away in opposite direction. In the most rudimentary sense, both these events present instances of causality in that object A causes object B to change position. The categorical difference is in how they are perceived. While the first scenario, the launching event, can be explained in terms of motion mechanics, the second draws on inter-pretive framework in which observers tend to refer to objects' intent and goals (object A tries to get close, while object B moves away). In other words, it exemplifies a relation between animates beyond motion mechanics (Schlottmann, Ray, Mitchell, & Demetriou, 2006; Scholl & Gao, 2013; Scholl & Tremoulet, 2000). Let us turn discussion to the first scenario repre-senting physical causality.

Figure 1. A schematic depiction of (a) Michotte’s (1946/1963) launch event and (b) Kaniza and Vicario’s (1968) reaction event showing contact causal event and a non-contact/action-reaction event.

Perceiving physical causality Interest in these types of phenomena is long lived. In Treatise on Human Nature (1740), David Hume describes a familiar event. He writes,

Here is a billiard ball lying on the table, and another ball moving toward it with rapidity. They strike, and the ball which was formally at rest now ac-quires a motion. This is as perfect an instance of the relation of a cause and effect as any which we know either by sensation or reflection.

(Hume, 1740 p. 292)

13

For philosophers such as Hume, this type of direct launch event represented an inquiry into how the physical world operates. He argued that motion of the first billiard ball is physically distinct and separate from the movement of the second, and our impression of the relation between the two is due to the process of induction, resulting from statistical association formed within recurring spatiotemporal pattern. Simply put, because the world is replete with these types of cause-and-effect relations, our impression of causality is nothing more than an inference from repeated visual exposure. As we will later see, the distinction between experience and innately driven inferences plays a key role for how causality is studied.

Unlike philosophical inquiries, the psychology of causality as we know today turned the focus away from the world and toward perceptual processing of the event. Belgian psychologist Albert Michotte was the first to study and consistently report on the perceptual processes during mechanical collision events akin to Hume's billard balls. In his book, The Perception of Causality (1946/1963) he concluded that what we see is not two distinct motions, but rather one motion causing another, something he referred to as the causal illusion. Michotte took an anti-Humean stance in noting that the illusion is phenomenological in nature in that it is direct, immediate and, most importantly, immune to repeated visual exposure. He futher reasoned that causal illusion is the result of appropriate event stimulation, and varying stimulus conditions would undoubtedly result in changes to the illusion of causality. As a consequence, Michotte's work and the work of those who followed him, focused on determining the right stimulus conditions requrired for perception of causality. For example, researchers found that low-level changes to the pattern of the stimulus sequence such as adding a spatial or temporal delay before the second object begins to move (Michotte, 1963; Yela, 1952), varying the speed and trajectories of the objects (Natsoulas, 1961), changes in the smoothness and continuity of motion (Gordon, Day, & Stecher, 1990) as well as the effects of perceptual grouping, disintegration and attention (Choi & Scholl, 2004; White & Milne, 1999) all influence the universal impression of causality as reported in observers’ subjective judgements.

In the years that followed, some critics have questioned the universality of the reports (Beasley, 1968; Boyle, 1960; Gemelli & Cappellini, 1958; Rips, 2011) while other researchers moved away from focusing on how changes to the event change the impression of causality to examining the degree of interpretation. That is, whether causal perception is distinct from higher-level causal inference, with the implication that if causality is a higher-level causal inference, it would not be perceptual but one requiring learning. In an attempt to answer this question, research turned to infancy.

14

Physical causality in infancy The study of perceptual causality in infancy has almost exclusively relied on the habituation paradigm. Using the reverse presentation of the launching event as a control, Leslie & Keeble (1987) found that 6-month-old infants effectively recovered attention when presented with the reversal of the caus-al, rather than the spatiotemporal, structure of the event. They attributed their findings to a perceptual process that is present early in development and is specific to certain type of information. Later theoretical writings (Scholl & Leslie, 1999; Leslie, 1988, also Leslie, 1986) extended the idea that the perception of causality involves the workings of an early-developing, possibly innate, visual mechanism whose function is to produce a description of the event’s causal structure.

Alternatively, physical causality involves a conceptual process that de-velops with infants’ growing ability to interact with and act upon the world (Piaget, 1936/1963). In fact, following Leslie & Keeble (1987), researchers began reporting evidence for a developmental progression in how causality is perceived during the first year of life. These findings reported varying results, depending on methodology and interpretation (Oakes & Cohen, 1990; Cohen & Oakes, 1993), but eventually even those opposed to Leslie's modular claim confirmed his initial findings with 6-month-olds (Oakes, 1994), but, like Leslie himself, failing to find evidence before 6-months. Only recently studies have suggested sensitivity to physical causality to be evident in newborn infants (Mascalzoni et al., 2013). Using global preferential looking, Mascalzoni and colleages (2013) found that newborn infants orient toward a direct launching event more than the non-causal control. Resonating with Leslie's initial claim, the authors argue that newborn's visual system possesses innate mechanisms for computing causality independent of any previous visual experience.

Is physical causality perceptual? In addition to the developmental time frame, other research examined whether perception of physical causality is in fact a perceptual process. Two lines of evidence seem to suggest that might be the case. The first comes from studies showing that the perception of causal structure is separate from other cognitive processes. The second comes from studies showing the direct influence of visual adaptation on the perception of causality.

In addressing the first, one study (Schlottmann and Shanks, 1992) presented adult observers with a launch event in which the launched object’s gradual color change, rather than a collision with the first object, predicted its motion. They found that although color change was a better predictor of motion, adults’ ratings of perceived causality were unaffected, even when they were forced to focus on that predictive relationship. The results demon-strate that factors such as associative learning and inference influence adult’s

15

judgment of causal relation but not the phenomenal experience of it, sug-gesting that perception of causality is a distinct perceptual mechanism.

Second, processes that do involve perception affect ratings. In a recent study, Rolfs and colleages (2013) presented adult observers with series of short animations depicting a stationary disc being approached by another. As the second disc approached, the discs eventually overlapped, at which point the second disc took off. The degree of overlap varied - when the overlap was a small one approaching disc appeared to (causally) launch the other’s motion, but when the overlap was large, one disc appeared to (non-causally) pass over the other. After each trial, the observers had to categorize the event as either a ‘launch’ or a ‘pass’. In the second phase of the study, par-ticipants were then shown an adapting stream of causal launch events fol-lowed by another presentation of the initial animations, which they again had to rate. The authors found that if the animations were presented in the same location as the adaptation stream, observers became stricter in catego-rizing the event as a ‘launch’ when they observed the animations again. Consequently, previously ambiguous events were judged as non-causal passes, while previously reported non-causal passes were now judged as ambiguous.

Having established that visual adaptation influences causality judgments, the authors tested where the adaptation to perceptual causality occurs. They reasoned that if it occurs at early stages of visual processing, the adaptation would be expected to occur in a retinotopic frame of reference. That is, the strongest aftereffects of the adaptation stream would be observed at loca-tions falling on the same patch of the retina and the corresponding retinotop-ically- organized brain areas. On the other hand, if perception is simply as-sociated with a specific space in the external world, the strongest aftereffects would be expected at a fixed location in space. By varying gaze position between adaptation and test, the authors found that indeed the strongest af-tereffects are when the retinal locations, but not the location in physical space, match. This shows that at least in adults, perception of causality is an early visual process, resulting in a reduction in responsiveness of particular neural populations specific to retinal location.

Neural correlates of physical causality The involvement of the visual process is also reflected in the brain areas mediated when observing these types of events. In general, neuroimaging findings seem to support the idea that perception and cognition are tied to-gether in a bottom-up-top-down-loop, and different levels of processing affect each other. A simplified presentation of the research on physical cau-sality in particular seems to show that depending on the type of presented contrasts results vary in the degree and location of neural activation. For instance, contrasting a launch event with an event in which one object passes under the second stationary object, resulted in significant activation in the

16

MT/V5, medial, and superior temporal lobes bilaterally, as well as regions of the left superior temporal (STS) and intraparietal sulcus (IPS) (Blakemore et al., 2001; Fonlupt, 2003). Since these areas have been previously implicated in the visual analysis of complex stimuli and motion processing, their re-cruitment suggests the involvement of the visual system in recovering the causal structure from dynamic displays.

In other studies, when launch events were contrasted with (non causal) events containing spatial or temporal gaps, causal events invoked regions of the right middle frontal gyrus and the right inferior parietal lobule (Fugelsang, Roser, Corballis, Gazzaniga, & Dunbar, 2005). Those regions have previously been implicated in attention (Banich et al., 2000), retrieving spatial context (Burgess, Maguire, Spiers, & O’Keefe, 2001), working memory representations (Prabhakaran, Narayanan, Zhao, & Gabrieli, 2000) as well as during tasks involving judgments of other’s intent such as theory of mind tasks (Gallagher et al., 2000; Vogeley et al., 2001).

Finally, eliciting inferential processes such as in studies in which partici-pants were asked to judge how causal the event appears, as opposed to judg-ing movement direction in causal events resulted in activation of the medial frontal cortex (Fonlupt, 2003).

Taken together, the existing studies examining neural correlates of per-ceptual causality suggest that perceiving causal events recruits areas respon-sible for complex visual analysis and motion perception such as MT/V5, STS and IPS, lending support to the notion that perception of causality is largely processed by the visual system. But while the perception of causality in launch events is the function of the visual system, the output from these visual regions may recruit regions associated with executive processes be-yond those afforded by the visual system. Furthermore, while making infer-ences about the observed causality seems to recruit specific areas such as the medial frontal cortex, separate from the areas involved in perception.

In all, studies here reported highlight that perception of cause is aptly named. Differentiating causal from non-causal structures suggests the in-volvement of a visual rather than an inferential mechanism. This is support-ed by studies showing that the ratings of perceived causality in adults are directly affected by changes to the perceptual input (adaptive stream) but not other inferential processes (learning). This claim is further strengthened by the fact that these processes have been localized to specific retinal locations and recruit brain areas specialized for processing of visual input.

However, while many observable causal interactions in the world are physical, such as famously described by Hume and empirically replicated by Michotte, equally numerous are causal interactions that do not require physi-cal contact. These types of interactions form the basis for social causality between animates, and research on this specific type of causality gives a slightly different perspective, but one that ultimately centers on the question of whether it too is a product of perception.

17

Perceiving social causality As described at the beginning of the previous section, there are two main types of causal events. In one, object A moves toward and contacts object B, which then immediately moves in direction opposite of impact. In the se-cond, object A approaches object B, but before they contact, object B begins to move away, resulting in an impression of purposeful movement away from object A. This type of action-and-reaction event is viewed as causal as long as the reactive object (object B) begins to move before the first stops. And so, given appropriate event configuration, causal impression can be maintained without direct collision between objects in the Michottian sense.

The causal phenomenology in action-and-reaction events was first empir-ically examined in the work of two Italian researchers, Gaetano Kanizsa and Giovanni Vicario (1968). Non-contact causal relations of this kind, they claimed, can result from one agent physically reacting to another’s motion (e.g. a mouse runs away when the cat approaches) or as a result of internal states such as intentions, desires and beliefs (e.g. mouse wants to get away from the cat). More often than not, the two are not separate but are conjoined in a single action (e.g. the mouse is running away because it has an intent of getting away from the approaching cat). In this way, while contact causality promotes learning about material bodies abiding by laws of mechanics, non-contact causality promotes understanding of psychological causes of animates (Schlottmann & Surian, 1999). Simply put, contact causality gives rise to the impression of cause, non-contact causal relations give rise to the impression of animacy.

Because contact causality abides by laws of mechanics it is the more direct, predictable and uniform form of causality. Animacy, on the other hand, is more challenging in that it encompasses many varied perceptual cues, and as a concept can be studied on a continuum from more focused to more general and abstract. For instance, animacy perception may refer to the perception of animate motion defined as motion which violates Newtonian laws of physics, can be a study of goal-directed motion or intentionality perception, or more generally, can be the study of the perception of social cause or meaning (Gao & Scholl, 2011). But however one defines animate motion, it ultimately comprises of a combination of particular motion cues that can also function independently. Individually, these motion cues may promote animate status of the entity to some degree. In combination, they may form percepts of animacy that can be parsed into perceptual units reflecting animate entity’s ’approach’, another’s ’avoidance’ or a type of complex social exchange between the two described as ’chasing.’

18

Examination of motion cues in the formation of animated percepts

Of the proper subjects of motion some are moved by themselves and others by something not themselves, and some have a movement natural to them-selves and others have a movement forced upon them which is not natural to them. Thus the self-moved has a natural motion. Take, for instance, any ani-mal: the animal moves itself, and we call every movement natural, the prin-ciple of which is internal to the body in motion.

Aristotle, Physics (vol. V, p. 307) Wealth of observation of infant behavior and empirical data suggest that young infants are sensitive to animate as opposed to inanimate motion, as each obey certain principles which form part of our core knowledge about the world (Carey & Spelke, 1996; Spelke & Kinzler, 2007; Spelke, Philips, & Woodward, 1995). Inanimate and animate motion differ considerably. For one, inanimate objects are governed by a set of physical laws, which means that the source of the object’s motion is always external to the natural com-position of the object and the manner in which an object moves is deter-mined by external forces. Research examining infant looking at events that either follow or violate these laws have suggested an early intuitive concep-tion of object motion (Carey & Spelke, 1996; Molina, Van de Walle, Condry, & Spelke, 2004; Spelke et al., 1995). This research established sev-eral principles. First is the principle of cohesion (Spelke, Breinlinger, Macomber, & Jacobson, 1992), which states that inanimate objects are con-nected and bounded bodies that maintain their cohesion as well as their sepa-ration from the environment. Second is the principle of continuous motion, which states that any single object traces one path over space and time, and two solid objects cannot occupy the exact same space at the same time. Fi-nally, objects act upon each other only through contact and, in the absence of other objects, gravity.

For animates, while the principles of cohesion and continuous motion are retained, limitations through collision or gravitational forces are not. Instead, cues pertaining to the quality of motion aid in categorizing entities as ani-mate. Although there are many, most motion cues can be generalized into three that form the basis for animacy categorization. These are: object’s autonomous motion, its goal-directed motion and contingency with other objects (Opfer, 2002).

Autonomous motion Autonomous movement, or self-initiated movement, is the most palpable cue to animacy. In fact, some theorists (Premack, 1990) claim self-propulsion alone is sufficient to judge an entity as an intentional agent. It is a distinctive quality not possessed by objects that move only as a result of

19

contact with others (Di Giorgio, Lunghi, Simion, & Vallortigara, 2016), it has to do with internally stored sources of energy (Gelman, 1990; Gelman et al., 1996; Leslie, 1996) that are channeled and released during motion by biological entities. Empirical studies have shown that 5-month-old infants perceive self-propelled movement and form expectations about events initi-ated as such (Baillargeon, Wu, Yuan, Li, & Luo, 2012; Luo, Kaufman, & Baillargeon, 2009). Some researchers suggest it to be an innate predisposi-tion of the perceptual system, as preference toward animations depicting self-propelled geometrical shapes is present even in newly-hatched (and therefore visually naïve) domestic chicks (Mascalzoni, Regolin, & Vallortigara, 2010).

At a perceptual level, however, autonomous movement alone is not suffi-cient to deem the object 'alive.'1 A plastic bag can be swooshed around mid-air and an airplane can pass overhead but a typical adult would not attribute either type of self-propulsion to the object’s animate nature. Instead, adults rely on prior knowledge of these objects to make the judgment. But even when faced with an unfamiliar entity, research has shown that self-propulsion alone does not always guarantee the object to be judged as an animate (Gelman et al., 1996; Poulin-Dubois & Heroux, 1994; Richards & Siegler, 1986). Rather, self-propulsion is often at best an ambiguous cue of animacy, requiring knowledge of energy source as well the object's biologi-cal or non-biological composition. Some of the earliest work on categoriza-tion of unfamiliar entities based on motion was done by Judith Steward in a series of unpublished work as part of her doctoral thesis at the University of Pennsylvania (Steward, 1982). Guided by the hypothesis that animacy can be perceived in any object that violates Newtonian laws of physics, Steward designed computer-generated animations of a dot that systematically dif-fered its movement. In Experiment 1, she examined adult's judgment of a dot that was self-propelled versus one that moved as a consequence of another object exerting force on it. The findings (later re-examined by Gelman and colleagues in 1995) show that autonomously moving object was judged as more animate compared with one that was externally propelled, but only in 25% of the observers, suggesting it to be a rather weak cue indicating ani-macy.

Thus, while self-propulsion allows making animacy judgments when compared to objects moved by external force, it is not by itself a sufficient cue for judging an entity as animate. Rather, as some theorists have pro-posed (Gelman et al., 1996; Opfer, 2002) it is the combination with other cues that allows for the distinction when the object is unfamiliar. As Gelman 1 Some of Michotte’s original experiments also included studies on what he described as ’animal locomotion’. In these, he created a rectangle that expanded and contracted resulting in a perception of self-propulsion. Interestingly, while increase in speed of this object in-creased reports of object displacement report, animacy categorization (observers calling the object a catepillar) was virtually non existent independent of speed (pg. 189)

20

writes, animate motion 'has a quality of function…[which is] a direct conse-quence of their governance by control mechanisms that make it possible for them to respond to the environments and adapt to unforeseeable changes in circumstances' (Gelman et al., 1996, pg. 151, emphasis added).

Goal-directedness Goal-directed motion of which all living things are capable can be defined as an ‘autonomous motion in which an agent contingently directs its movement toward (or away from) another object, state, or location’ (Opfer, 2002, pg. 100). Extending Opfer’s initial definition of goal-directedness, the present thesis defines goal-directedness as a feature of agency that has to do with motion that is (I) direct in terms of motion trajectory, is (II) rational in terms of present contextual constraints and is (III) contingent on environmental variation (Csibra, Gergely, Bíró, Koós, & Brockbank, 1999; Gergely Csibra, 2008; Gergely et al., 1995; Kamewari, Kato, Kanda, Ishiguro, & Hiraki, 2005; Sodian, Schoeppner, & Metz, 2004). Furthermore, the definition can be subdivided into directionality-based components. Specifically, while mo-tion toward and away may be construed as ‘goal-directed’ as long as it is direct, rational and contingent, only the object whose motion is directed toward, rather than away from, another can be described as ‘heat-seeking’ (a term first employed by Gao, Newman & Scholl, 2009).

In a chasing interaction, for instance, two self-propelled agents are theo-retically goal-directed in a sense that the chaser’s goal is the target, while the target’s goal is to get away from the chaser. The difference between the two is their direction of motion (one is moving toward, another away) and in one case the goal is external (i.e. the target) and salient, while in the second case it is internal, more abstract and less obvious. Therefore, while both objects are goal-directed, only the chaser is heat-seeking, an important distinction examined in the present thesis.

In either case, like self-propulsion, goal-directedness is a feature of ani-mate action, and the same assumption about goal-directedness applies to any entity even a geometrical shape (Luo & Baillargeon, 2005; Schlottmann & Ray, 2010; Shimizu & Johnson, 2004). For instance, when presented with an ambigious, autonomously-moving entity, both adults and children identify it as a living organism when it moved toward a salient goal, but were less likely to categorize it as living when the goal was removed (Opfer, 2002). Consequently, aimlessly moving autonomous objects were less likely to be judged as living, but goal-directed agents were more often judged to be ani-mates. In fact, even superficial impression of goal-directedness (when there is none) results in the perception of animacy and effects visuomotor behav-ior of the observer (Buren, Uddenberg, & Scholl, 2015; Gao, McCarthy, & Scholl, 2010).

21

Contingency Another motion cue that helps in making the distinction between an animate and inanimate object is contingency. That is, whether action and behavior of an entity is time-linked to those of another.

There are two types of contingencies. One is mechanical contingency such as when one object collides with and effectively displaces another or when objects change in predetermined sequence (such as musical lights turn-ing on and off in sequence). The second type is an intentional contingency reserved for animates and forms the basis for all social exchanges. This type of causal contingency can involve the self, of which the most prominent example comprises a highly orchestrated back-and-forth exchange between infants and their parents (Brazelton, Koslowski, & Main, 1974; Stern 1974; Trevarthen 1977, 1979; Cohn and Tronick 1988). It can also involve observ-ing motion between objects such as when motion pattern of one object re-sembles another over time (turn taking) or through spatiotemporal configura-tion when one agent follows motion trajectory of another or reacts to another object’s shifts in directionality and accelerations such as in a chasing interac-tion.

In all, causal contingencies are varied but like goal-directed motion, mo-tion contingency enhances the perception of animacy. Even an amorphous shape can be endowed with attentional and perceptual abilities after being observed interacting contingently with another (Johnson, 2003; Johnson, Alpha Shimizu, & Ok, 2007; Johnson, Slaughter, & Carey, 1998).

To summarize, judging animate status of an entity is one of the basic abil-

ities and knowledge about objects and agents has been suggested to form the core aspects of human cognition (Spelke & Kinzler, 2007). While mor-phology allows making such judgments relatively easy, motion cues are equally, if not more, important in making the final call. Motion confirms (or contradicts) object’s animacy. Motion cues that help in categorizing entities as animates can be agent-specific, relating to the object's nature (self-propulsion) and relational (goal-directedness and contingency). The combi-nation of these motion cues results in inferences about object’s animacy because just as people appear 'subject to mechanical forces' so can 'inani-mate objects be seen as agents in social interactions' (Schlottmann & Surian, 1999, pg. 1107).

But do we really perceive animate motion? Much like the perception of cause, a critical question regarding social cau-sality in animate displays concerns the nature of the impression. Specifical-ly, whether the perception of animacy is inherent to the processing of the visual system. Answering this question has proven to be difficult mainly due

22

to how animacy perception has been measured, an important theoretical criticism brought forth by Brian Scholl and Tao Gao (Scholl & Gao, 2013). As they point out, for years the dependent variable in measuring percepts of animacy (much like measures of cause discussed earlier) has been either based on observer’s free reports coded for words pertaining to sociality in more or less elaborate ways (Heider & Simmel, 1944; Heberlein & Adolphs, 2004; Klin, 2000) or scaled ratings where observers are asked to rate their perception. Most of these ratings are numerical – for example, rating anima-cy on a 7-point scale, from 1 = ‘‘definitely not alive”, to 7 = ‘‘definitely alive” (Tremoulet & Feldman, 2000). Others, assign numbers post-hoc to various categorizations – for example rating animacy on a 4-point scale that consists of ‘‘physical”, ‘‘rather physical”, ‘‘rather personal”, and ‘‘personal” (Santos et al., 2008) options. However, as Scholl rightly points out (Gao & Scholl, 2013; Gao, Newman, & Scholl, 2009) the sheer act of judging impli-cates (or as he writes ‘contaminates’) percepts with higher-level cognition. In an exemplary case (Tremoulet & Feldman, 2000), participants were asked to rate motion of a single dot on a seven-point scale from ‘definitely not alive’ to ‘definitely alive’, with low ratings equating to motion that seemed ‘artificial, mechanical, or strange’ (pg. 945). Unsurprisingly, the results demonstrated that the more extreme the variations in direction and speed the higher the ratings, which although in line with the standing hypothesis, ques-tions whether observers really perceived the motion to be more animate or whether they varied their responses according to the observed variation of the motion itself. Simply put, whether the observers had a different impres-sion of animate motion or whether they responded in the way they did be-cause the motion varied across trials.

Asking observers to judge their percepts, implicates the person’s belief systems and inferences, not to mention risking the participants to be influ-enced by experimenters’ own intentions coupled with their own assumption that they should vary their responses across hundreds of trials. The key in addressing these concerns is to move toward more objective ways of meas-uring perception. One way is to examine visual performance through eye-tracking. Another is to examine neural correlates during observation of ani-mate motion. Let us consider each in turn.

One approach of controlling for the effects of self-reports is to measure oculomotor behavior and visuomotor performance during the observation of motion. In accomplishing this task, some studies required the participants to respond as soon as they detected a change. In one study, they responded as soon as they detected an object’s disappearance when presented with multi-ple animatedly interacting or inanimately (bumping) moving shapes (Pratt, Radulescu, Guo, & Abrams, 2010) or as soon as they detected a probe (small vertical or horizontal line) appear on one of the moving objects (Meyerhoff, Huff, & Schwan, 2013; Meyerhoff, Schwan, & Huff, 2014b). In still other studies, observers had to indicate (by pressing a button) whether they ob-

23

served a chasing interaction (Find-the-Chase), a type of non-contact causal relation resulting in high degree of animacy perception. In still others, partic-ipants were asked to manipulate ‘their’ shape which was being chased by another (Don’t-Get-Caught!) (Gao et al., 2010; 2009). Independent of the method, these studies show that objects involved in a chase are detected more quickly than inanimately or randomly moving shapes, provide evi-dence that animate motion effectively biases visual attention, and in turn influences visuomotor performance.

In all, measures of oculomotor bias and visuomotor performances in adults, rather than reliance on self-reports aim to eliminate the fundamental issue of subjective self-report in an attempt to isolate perceptual processes during the observation of animate motion. This method also allows for a study of animacy perception in preverbal infants although this possibility has not been prolific. Importantly, attentional biases and effects of visuomotor performance alone do not inform about whether these types of causal rela-tions are interpreted as social (point brought forth here following Study III). Rather, they speak for the bias of animate motion by the visual system. In an effort to address how animate motion is processed, we turn to neural corre-lates during the observation of animate motion.

Neural correlates of animate motion Categorization of animate motion, without the reliance on subjective self-reports, can be accomplished by examining the brain regions that support it. General findings from these studies confirm that animate motion elicits acti-vation in the temporoparietal cortex, posterior superior temporal sulcus (pSTS) and the angular gyrus, areas normally involved in interpreting social information (Blakemore, 2008; Frith & Frith, 2007). In majority of these studies, adult observers were presented with a broad array of interactions, often in visually simple displays depicting interactions between two-dimensional geometrical shapes. The interactions range from simple motion contingency (Blackmore et al., 2003) to more complex type such as dancing, fishing, sharing, scaring each other, playing on a seesaw, swimming (Martin & Weisberg, 2003), persuading, bluffing, mocking and surprising one an-other (Castelli et al., 2000). These types of interactions are contrasted with more mechanical motions such as movement reminiscent of billiard balls, pinballs, a cannon, a crane, a conveyor belt and a paper shredder. Within this literature, one recent study (Hillebrandt, Friston, & Blakemore, 2014) de-serves more detailed description. This study examined neural connectivity during animacy perception using dynamic causal modeling (DCM) which modulates self-connections within a specified brain area as well as the direc-tionality of the connections between the brain regions, allowing to infer whether the experimental manipulations affect top-down (backward) or bot-tom-up (forward) connections. In a large-scale fMRI dataset, the authors

24

presented 132 participants with 20-second video clips depicting geometrical shapes moving in animate or inanimate way. In the animate condition, two shapes were interacting in more (persuading, bluffing, mocking and surpris-ing one another) or less (dancing, chasing, imitating and leading) elaborate way. In the inanimate condition, the shapes were bouncing off the walls like billiard balls, in no relation to each other.

The authors identified regional responses for the DCM by looking at each participant’s local maxima activations that were nearest to the peak voxel of the group of participants for all motion (animate and inanimate). Using this method, they identified the MT+/V5 or the human occipital lobe area 5 (hOC5) during the observation of all motion. The authors then compared activation during all motion and compared it with the activation during ani-mate-inanimate contrast and found activation in the pSTS (inferior parietal cortex IPC) bilaterally which, as previously discussed, has been implicated in animate motion processing. The findings further demonstrated support for a forward connection from MT+/V5 to the pSTS modulated by animacy. In addition, inhibitory self-connection was decreased to zero in pSTS during animacy perception meaning that activation in pSTS was increased. Authors concluded that observation of animate motion is a bottom-up process, which begins with motion processing (MT+/V5) that is then forwarded to area spe-cialized in the processing of social information (pSTS).

The study has shown that viewing animate motion, while processed in the visual areas, is ultimately categorized as social. How specific are these pro-cesses and their ontological beginnings, however, are yet to be determined. In examining infant early categorization of these types of events will inevi-tably shed necessary light on these particular questions.

Neural correlates of animate motion in infancy Infant research on motion perception has shown that even newborns are attuned to biological motion (Johansson, 1973; Johansson, von Hofsten, & Jansson, 1980) by orienting and attending to a point-light (PL) display ar-ranged in a form of a human or an animal in motion (Bardi, Regolin, & Simion, 2011; Bertenthal, 1993; Fox & McDaniel, 1982; Simion, Regolin, & Bulf, 2008). One possibility for this may have to do with perceiving objects that follow similar motion trajectory as a unified whole, known as the Ge-stalt law of ‘common fate’(Wagemans et al., 2012). In this way, the percep-tion of PL walkers may be the result of a global process whereby individual dots signify important joints and structures of the human or animal body bounded together into a single coherent whole, whose motion reflects this coherence. Unlike biological motion displays, animate displays rely on the fact that each dot is perceived as an individual object whose quality of mo-tion implies agency and whose relationship with others (rather than being physically bounded) relies on a relatively abstract notion of social intent.

25

And so, while PL walkers and certain animate displays share common fea-tures in terms of low-level aspects and processes involved, important differ-entiation is in the fact that while a single dot in animate displays is able to convey a rich array of mental states it typically is not the case for individual dots in PL walkers.

Keeping this point in mind, research on animacy perception relies on the assumption that, much like biological motion, animate motion is visually preferred. This assumption is based on the notion that animate motion is generally less predictable for humans than inanimate motion. It tends to be more complex, can but doesn’t have to be linear, involves objects that are self-propelled that usually do not follow Newtonian laws of physics.

Like the habituation paradigms employed in studying physical causality (Cohen & Oakes, 1993; Leslie & Keeble, 1987; Oakes & Cohen, 1990), preferential looking paradigms rest on the idea that infants are able to dis-criminate between two types of stimuli and attend one over the other (Gelman et al., 1996). Preference for animate motion may indicate that in-fants visually attend the less predictable motion trajectories, making the animate motion more salient of the two.

Following this logic, one ERP study found evidence for differential sensi-tivity to animate and inanimate motion in 9-month-olds (Kaduk, Elsner & Reid, 2013). In this particular study, infants were presented with a geomet-rical shape moving through a type of obstacle course in which it can either follow Newtonian laws of mechanics or move in violation thereof. For in-stance, a ball jumped up to get to another level (in violation of gravity force) or fall down (in accordance with forces of gravity). But somewhat contrary to the behavioral findings, this study suggested that by 9-months, in an at-tempt to assimilate motion information, infants allocate more attention to an object moving inanimately, rather than an animate object. Greater allocation of attention due to increased processing of motion information for the inan-imate object was defined as an increased negativity in the fronto-central region in the mid-latency component, known as the Nc. Previously, this component was found to reflect general attentional arousal (Richards, 2003; Richards, Reynolds, & Courage, 2010) as well as orientation to salient stim-uli (Courchesne, Ganz, & Norcia, 1981). The authors cite methodological differences to account for the inconsistency between their and behavioral findings. Further, they suggest that because animate motion is more easily identified as such, it may require less attentional resources. However, the general conclusion is that infants as young as 9 months do indeed discrimi-nate between animate and inanimate motion.

Overall, objective measures in studying animate motion seem to suggest the presence of a visual bias, reflecting a bottom-up process by which mo-tion is processed in the visual areas and then is forwarded to areas normally involved in the processing of social input. Furthermore, the attentional bias toward animate motion seems to be present early in development, and re-

26

quires less attentional resourses in comparison to processing inanimate mo-tion.

Chasing Although thus far the discussion centered on animate motion in general, the importance of present empirical work is its focus on one type of non-contact causal interaction, chasing. Like action-reaction event, chasing involves one object (chaser) moving at constant velocity toward the second (target) in a direct path. That is, it is heat-seeking. The target then reacts contingently to the motion of the chaser by moving away before the chaser gets too close. Unlike action-reaction event, however, chasing is often non-linear and the target typically does not follow a straight trajectory. In turn, the chasing object varies its motion rationally and efficiently with respect to the target’s motion and trajectory (Gergely et al., 1995). And so, because a typical chas-ing interaction involves autonomously moving objects, whose motion is evidently goal-directed and contingent, chasing interaction is a powerful cue of animacy.

Evolutionarily speaking, chasing detection has been critical to survival. Efficient identification of other animates within one’s visual field, as well as determining whether the object in motion is a potential predator or food source is undoubtably a powerful adaptive mechanism. Some theorists (Barrett, Todd, Miller, & Blythe, 2005; Blythe, Todd, & Miller, 1999) have suggested that chasing detection is critical because while it is one of few stereotyped types of interactions within the natural environment, it is one that can be displayed by a large range of different animates. Therefore, correctly recognizing important cues pertaining to chasing (rather than remembering who typically is the chaser per se) across varied animates has ensured survivial of our ancestors. Today, chasing may continue to be critical for survival but more often it is important for reasons of competition, social dominance and play (Thomsen, Frankenhuis, Ingold-Smith, & Carey, 2011; Steen & Owens, 2001).

Empirically, chasing is a simple type of an interaction that does not involve turn taking and can be extended over time indefinitely. At the same time, it can be manipulated to study the effects of specific spatio-temporal variations on perception (Nahin, 2007; Kaniza & Vicario, 1968). Because of this, chasing has been used in a number of adult and infant studies, the most relevant of which are reviewed below.

Chasing perception in adults Across a relatively large number of adult studies, findings suggest that chasing geometrical shapes are perceived as animates, and that this

27

impression is automatic, irresistable and at times, in contrast to the observers own beliefs. When asked to describe what they see, adults often refer to the geometrical shapes as anthropomorphic entities moving in an intentional, goal-directed manner, endowed with not only specific goals and intent (Barrett, Todd, Miller, & Blythe, 2005; Blythe, Todd, & Miller, 1999) but also, depending on the type of interaction, range of beliefs and emotions (Dittrich & Lea, 1994; Gao, Newman, & Scholl, 2009; Scholl & Gao, 2013; Schultz, Friston, O’Doherty, Wolpert, & Frith, 2005; for review see Tremoulet & Scholl, 2000).

Much like research on animate motion in general, studies focusing on chasing specifically have demonstrated a visual bias toward chasing interactions. When presented with sets displaying multiple identical discs, adult observers were faster at detecting discs involved in a chasing interaction from among randomly moving discs, but were significantly slower at identifying randomly moving discs from among chasing ones (Meyerhoff, Schwan, & Huff, 2014a, Exp. 2).

Authors conclude that identifying a chasing interaction involves visual attention to relational or individual object cues, depending on the task at hand. For instance, reduced spatial distance a feature of a chasing interaction was found to effectively bias visual attention, even when it was determinetal to the task at hand. In one study, adult observers were quicker at detecting a probe appearing on discs that were close, even though the probability of a probe appearing was much higher on objects that were far apart (Meyerhoff, Schwan, & Huff, 2016, Exp. 1a, 1b). They further found that this attentional bias seems to be mediated by oculomotor behavior, because when eye movements were controlled (by having adults focus on a central fixation point) the attentional bias was reversed (Meyerhoff et al., 2016, Exp. 2a, 2b). These results demonstrate that it is not that the probe goes undetected when appearing on distal objects, rather it is that adult oculomotor behavior mediates visual attention toward objects that come close together.

Much like sensitivity to relational cues that help in identifying chasing interactions, cues pertaining to individual objects involved in a chase attract visual attention. One study showed that adult observers were faster at detecting a probe, when it appeared on the chaser as opposed to its target (Scholl and Gao, 2013). This was especially evident at the moment the chaser switched the target of its chase (Gao et al., 2009; Scholl & Gao, 2013), that is, it started heat-seeking another target. Further, changes in the motion cues of the chaser were found to profoundly effect the detection of chase in general. In one study, varying chasing subtlety and directionality or the degree to which the chaser deviated from the direct heat-seeking path toward the target, effected how quickly adults recognized the interaction (Gao et al., 2009). In all, the more obvious heat-seeking, the easier it is to detect the chaser.

28

But while changes to chaser’s motion affect the detection of a chasing interaction, some have suggested that chaser detection in particular is not immediate, as it does not ’pop-out’. In support of this claim, one study (Meyerhoff et al., 2014b) reports that increasing the number of distractors (i.e., presenting more objects or distractors) affects how quickly adults detect and identify a chaser from among visually identical shapes. This means, that while detecting chasing objects is relatively easy and changes to the chaser motion affect its detection, it is not an effortless process and certain amount of attention is needed. Instead, the set-size demands may result from a perceptual serial visual search (Meyerhoff et al., 2013), whereby each object is scanned individually for cues verying its involvement in a chase. As Mey-erhoff discusses, two options may explain the findings. One is that the hu-man visual system is sensitive to certain motion invariants (Cutting, Proffitt, & Kozlowski, 1978; Gibson, 1979/2014) such as intentional motion patterns (heat-seeking). Second option is that spatial proximity biases attention to-ward the two objects. Once there, agent specific cues may further direct attention toward specific agents. To test these two possibilities, more direct test is necessary such as examining gaze latency toward the chaser specifically, rather than relying on button-press response.

To summarize, current findings seem to suggest that, at least in adults, the detection of non-contact causal interactions, such as chasing, is a conse-quence of a of visual bias toward relational cues inherent to interactions (reduced spatial distance), as well as agent specific motion patterns (heat-seeking). But exactly how these processes take place requires more direct tests such as examining gaze patterns using eye-tracking methodology dur-ing online observation of motion as well as aptly designed tests. This, how-ever, is not an easy task. For one, to determine which cues drive adult visual attention during chasing, it is necessary to dissociate spatial proximity and heat-seeking, where one necessarily causes the other, while both cues consti-tute a chasing interaction. Second, it is necessary to examine how chasing interactions are attended in general, whether chaser bias is evident in the oculomotor pattern and whether this pattern extends to other types of inter-actions. Finally, providing developmental framework would shed more light on the early perceptual processes behind visual attention to chasing.

Chasing perception in infancy Much like adults, young children seem to perceive chasing as an interaction between animates. When presented with animations such as those by Heider and Simmel (1944), children describe them using anthropomorphic language in reference to the agents and their behavior. In one study, Berry and Spring-er (1993) presented 3-, 4- and 5-year-olds with four different versions of chasing, in which they varied motion and form of the shapes. Much like adults (Berry, Misovich, Kean, & Baron, 1992) children of all age groups

29

interpreted the displays as animate and shapes as anthropomorphic entities in conditions where the motion was preserved irrespective of form, demonstrat-ing that it is the properties of movement, rather than the stimuli that are fun-damental to the complex attribution of animacy in young children as well as adults. But while preschool children may already take note of motion prop-erties and are able to verbalize what they see, infants cannot. This fact, how-ever, does not minimize the importance of research with pre-verbal infants. To the contrary, it highlights such need because examining oculomotor be-havior in pre-verbal infants may tap into initial perceptual mechanisms that may form the basis of animacy attribution and interaction categorization later.

As with adults, one way of examining perception of animate motion is to focus on chasing. Using preferential looking paradigm, Rochat, Morgan, and Carpenter (1997) presented adults and infants between 3- and 6-months with two displays depicting two moving discs. In one display, a disc chased an-other while its target was trying to avoid being caught. In another, both discs moved haphazardly in no relation to each other, bumping into the boundaries of the screen (random/inanimate condition). The study showed that when presented with interacting and non-interacting motions 3-month-old infants looked longer at a display in which the discs were chasing than the control display, demonstrating early sensitivity to perceptual information for interac-tions.

Using chasing, a follow-up study (Rochat, Striano, & Morgan, 2004) found that slightly older infants were also able to differentiate between which one of the objects was the chaser, suggesting that they not only super-ficially discriminate between how the object interact, but encode two distinct objects displaying a particular set of motion cues. Later research (Frankenhuis, House, Barrett, & Johnson, 2013a) confirmed that for young infants, much of the visual attention directed toward chasing displays may have to do with the presence of individual motion cues such as object accel-erations and relational cues such as attraction (or heat-seeking) of one object toward another. These findings corroborate with adult research (Dittrich & Lea, 1994; Gao et al., 2009; Heider & Simmel, 1944) and the authors pro-pose a very basic property-based attentional filter that navigates attention toward cues heralding social interactions. Thus, whether a visual bias (Meyerhoff, Schwan, & Huff, 2014a; Meyerhoff et al., 2014b, 2016) or an attentional filter, the preference to attend to animate motion can be largely explained by sensitivity to specific motion cues informing about the type of interaction. However, in order to make definite conclusions about the atten-tional influence of a particular set of perceptual cues, it is important to ex-amine how infants attend to these types of events during online observation of chasing. To our knowledge, there is no study currently designed for that purpose.

30

Neural correlates of chasing Studies examining neural correlates involved in the perception of chasing are notably few (Lee, Gao & McCarthy, 2014; Schultz, Imamizu, Kawato, & Frith, 2004, Schultz at al., 2005). More often, the focus extends to examin-ing animate motion in general or intentional motion more specifically. Be-cause chasing is a type of animate motion that is goal-directed and intention-al it is often used as one of many types of goal-directed motions often con-trasted with inanimate, or mechanical motion exemplars (Castelli et al., 2003). But, independent of whether chasing is examined specifically or as an exemplar, the findings seem to implicate an activation of the posterior part of the superior temporal sulcus (pSTS) (Lee, Gao, & McCarthy, 2014; Schultz et al., 2005, 2004; Shultz, Lee, Pelphrey, & Mccarthy, 2011; Shultz & McCarthy, 2012) with some demonstrating stronger activation on the right hemisphere (Gao, Scholl, & McCarthy, 2012).

Similarly, adult ERP studies tend to focus on neural correlates of animate motion more generally. Findings from these studies also seem to implicate similar brain region. One study found that analyzing motion of an animate as opposed to an inanimate object results in a negative ERP response in the right occipito-temporal region around 250ms post stimulus onset (Fukuda & Ueda, 2013), while findings examining biological motion perception, a type of animate motion, find a similar N240 component found to generate in the STS (Hirai, Fukushima, & Hiraki, 2003; Hirai, Senju, Fukushima, & Hiraki, 2005).

The activation of the pSTS seems to be specifically linked to the percep-tion of goal-directed motion. In one study Gao, Scholl, & McCarthy (2012) presented adult observers with four displays that differed according to how goal-directed was the motion. In one display, one disc (wolf) chased another (sheep) (Single Intention condition). In another display, the wolf switched the target of its chase (Wavering Wolf condition). Adults also observed two control conditions aimed to control for low-level visual changes (Flashing condition and Phantom Chasing). A comparison between a Single Intention and the Wavering Wolf conditions indicated an increase in activation of the right pSTS and the adjacent right lateral occipitotemporal cortex (LOTC). Because animacy was the same across these two conditions, the activation of right pSTS seems to be specific to the perceptual analysis of shift in inten-tionality of the agent, or the change in goal-directed motion. In much the same way, pSTS was stronger in the Wavering Wolf condition compared to the two control conditions, eliminating the possibility the activation is due to low-level visual changes. Unfortunately and importantly, the comparison between Single Intention condition and the controls is not included in the paper, which begs the question whether any presence of goal-directed mo-tion or approach of one object toward another (rather than the switch) would result in similar activation.

31

In addition to goal-directedness, some studies examined neural correlates of interaction, or contingency of motion between objects (Georgescu et al., 2014; Santos et al., 2010; Santos, David, Bente, & Vogeley, 2008). In one such study (Santos et al., 2010), using a very simplistic design, a group of researchers from the University of Cologne presented adults with animations depicting one self-propelling sphere approach a stationary sphere. In some animations the second sphere ‘responded’ by approaching the first in return. In this way, the authors varied particular motion parameters, specifically goal-directed motion (approach by the first sphere) and contingency (re-sponse of the second sphere). They found that presence of both approach and responsiveness resulted in the highest impression of animacy, which in turn resulted in the strongest increase bilaterally in the insula extending to the posterior superior temporal sulcus.

In sum, several studies here reported found that goal-directed motion and contingent motion increases the perception of animacy and the degree of interaction between two agents selectively recruits areas normally associated with processing various types of social information most notably the pSTS.

Theoretical considerations The theoretical discussion normally associated with the perception of physical or social causality is often limiting in that it follows the dichotomy typical to developmental science, reminiscent of the age-old nature versus nurture debate in biology. This dichotomy centers on the idea of psychological dualism in which processes are either innate or learned, automatic or requiring conscious mentalizing, specific to certain kinds of information or general in the way information is processed. Consequently, physical causality is either innate, automatic, specific to one type of informational domain and perceptual, or is learned, conscious, resulting from a domain general learning and growing cognitive sophistication shaped by our immediate experience with the world. In much the same way, this dichotomy extends to social causality and animacy perception.

To review, the general findings from collision events and their various empirical variations, suggest that distinguishing physical cause from other types is present at birth (Mascalzoni et al., 2010) or is empirically evident within first few months of life (Oakes, 1994; Leslie & Keeble, 1987). Most recent findings (Kominsky et al., submitted) seems to suggest that perception of categorically distinct physically causal events is evident after 6 months. However, this ability seems not to be a binary construct, and depending on how it is tested, various types of empirical findings have shown developmental progression in how mechanical causality is perceived (Oakes & Cohen, 1990; Cohen & Oakes, 1993).

32

In much the same way, distinguishing animate from inanimate motion seems to be emerging early in much the same way (Di Giorgio et al., 2016; Rochat et al., 1997) and in the same way, there appears to be developmental progression in understanding individual object’s role within animate interactions (Frankenhuis, House, Barrett, & Johnson, 2013b; Rochat et al., 2004).

Related to the question of ontology is the nature of the resulting impression. That is, whether the impression of physical cause and animacy is indeed a product of perception. Some empirical evidence suggests that might be the case. Perceiving physical causality for instance seems to be effected by visual adaptation specific to a certain retinotopic locations (Rolfs, Dambacher, & Cavanagh, 2013). Similarly, animate motion has been shown to bias visual attention (Pratt et al., 2010; Meyerhoff et al., 2013, 2014a), it biases oculomotor behavior (Meyerhoff, Schwan, & Huff, 2016) and is greatly effected by changes in low-level visual input (Gao et al., 2009; Gao & Scholl, 2011)

Given current knowledge, I turn the discussion toward those two existing perspectives on the development of physical causality and animacy: the modularity approach and the information processing approach. However, steering away from the existing theoretical divide, I end by introducing a third, somewhat integrative, framework that highlights the role of special-ized perceptual mechanisms in promoting learning.

Modularity Historically, the idea of modularity reflects the notions put forth by Jerry A. Fodor, who in his seminal book The Modularity of Mind (1983) presented a thesis that the mind consists of genetically specified, independently func-tioning modules, or input systems, which serve a specific purpose. Accord-ing to Fodor, the information received from the environment initially passes through a system of ‘sensory transducers’ whose function is to transform information into formats catered to the special purpose module. Then, each module outputs data in a format that allows for the domain-general pro-cessing of the central system. The essential key to Fodor’s claim is the idea that modules are (I) innate, (II) domain-specific, (III) informationally-encapsulated, (IV) inaccessible to consciousness, (V) have a characteristic ontogenetic course, (VI) are fast, and (VII) have dedicated neural architec-ture (see Fodor, 1983, for a definition and description of each of these prop-erties).

Much support for the modular approach steams from studies examining early discrimination of causal events as well as the seeming phenomenology of the impression. Such discrimination and preferential orientation toward causal events seems to be evident in newborn infants (Mascalzoni, Regolin, Vallortigara, & Simion, 2013), in direct contradiction to the notion that un-

33

derstanding of causality is construed over an extensive period and is depend-ent on growing elaborations of infant’s actions (e.g. prehensile skills). Ra-ther than being linked to motor development, it is argued that causality re-sults from a fairly low-level perceptual mechanism reflecting vision’s spe-cialized internal structure. According to this view, visual system is equipped with mechanisms, which recognize object’s form and motion analysis that identify kinetic properties of observed events. In short, vision makes explicit the spatial information over time. Visual attention in turn is directed toward these events, allowing for the perception of causal chains that are then for-warded to the general conceptual developmental processes. As Leslie writes, ‘it characterizes those properties of the information processing system that provide the basis for development, as opposed to those properties that are the result of development’ (Leslie, 1996, pg. 121).

Following Leslie, Brian Scholl proposed that perception of causality is a product of the visual system being able to recover the social structure of the observed event, much like object’s color or shape (Scholl & Gao, 2013). Scholl deviates somewhat from the traditional modular claim in that in at-tributing animacy he argues against a monolithic agency or animacy detec-tor, but rather empathizes sensitivity to individual motion cues, taking stud-ies in which motion cues are varied subtly as well as the distinct activation of the visual areas, along with evidence separating higher level judgment from perception and visuomotor performance, as confirmation of his claim.

Information-processing In contrast to modular view, is the notion of causality as an inductive infer-ence. Piaget (1954) was the first to distinguish between stages of causal un-derstanding, derived from an increasing ability to control ones own body and to actively interact with the world. In the earliest stages of development, according to Piaget, a child has no understanding of the concept of object or of self, and thus has no meaningful understanding of physical causality. In later stages, the infant learns to distinguish between self and the environ-ment, understands the permanence of objects and begins to understand cause and effect relationships. Toward the end of the second year, infant comes to understand the distinction between psychological or internal causality de-rived from one’s own power over one’s actions, and the physical relation-ship between objects.

Support for the information processing view comes from studies dis-cussed earlier demonstrating a developmental trend in the understanding of causality. According to this claim, because infants are top-down processers, every time the input content deters them from forming a causal relationship, they drop to a lower and lower processing level. In general, when presented with Michotte’s launching event at 4-month-old infants continue to attend to a causal event despite being habituated to it (Cohen & Amsel, 1998), sug-

34

gesting they are not yet able to discriminate between the causal and non-causal events. Some have suggested infants are entrained by the smoothness of the causal event at such an early age. At slightly older ages, infants begin to process the launching event in terms of spatial and temporal aspects, and between 7-10 months begin to process it as an integrated causal event. At around 10 months, as they acquire the understanding of object permanence, they seem to understand cause and effect even when the collision is occlud-ed (Lucksinger, Cohen, and Madole, 1992; Van de Walle, Woodward, and Phillips, 1994). At around 14-months they begin to differentiate between the roles of each object in the launching event - one agent and one recipient (Golinkoff, 1975; Golinkoff & Kerr, 1978; Redford and Cohen, 1996). Fi-nally, at each stage some studies have shown, the level of processing de-pends on the complexity of the objects used, with more complex and realis-tic objects found to impair infants’ ability to perceive causality (Cohen & Oakes, 1993).

Although not explicitly discussed, a similar information processing de-velopment may take place in the formation of animated percepts. For in-stance, Rochat and colleagues (Rochat et al., 1997) found that while global discrimination of socially causal events is evident early, the role reversal specifying roles of one agent with another suggests developmental progres-sion in the type of information being processed, with positive findings to-ward the end of the first year (Rochat et al., 2004).

In sum, responding primarily in terms of something as simple as entrain-ment at one age, spatial and temporal features at a later age, and then in terms of causality at a still later age is problematic from an innate modularity perspective, but it makes perfect sense from a more traditional developmen-tal perspective (Cohen, 1988, 1991).

The possibility of functional specialization One of the major shortcomings in reasoning about cognition in terms of the dualism is that it doesn’t explain the totality of empirical evidence. That is, it doesn’t allow to account for evidence supporting developmental progression, while also acknowledging evidence for sensitivity to certain types of information in newborn infants. In addition, neither framework attempts to address the functionality of various cognitive features.

As a result, in recent years, others (Cosmides & Tooby, 2000; Tooby et al., 2003; Barrett & Kurzban, 2006; Barrett, 2015) influenced by core principles of evolution, proposed that modularity should be grounded in the notion of fuctional specialization. That is, on the idea that natural selection, created specialized learning mechanisms that help in the formation of cogni-tive concepts that are pertinent to the existing environmental parameters. The characteristic feature of these mechanisms is that they are hierarchal, deriving from ancestral ones, yet are adaptive and influenced by current

35

environment (Barrett, 2012; 2015). Importantly, these mechanisms have evolved because of the function they serve in that they are necessary for the organism to survive and reproduce. Identifying animates through motion may be one such mechanism. It may result in an attentional mechanism to-ward specific motion properties that are highly diagnostic of animacy, of which goal-directed contingencies may be one (Gao et al., 2010; Gergely & Csibra, 2003; Heider & Simmel, 1944; Scholl & Tremoulet, 2000). And so, because of the function they serve and have served in our evolutionary past, these mechanisms may be evident early but develop to adjust to current en-vironment.

The theoretical perspectives here discussed specify possible mechanisms behind the perception of chasing. However, as was highlighted throughout the introduction, current knowledge is still somewhat lacking in regard to the behavior during chasing observation. While there is evidence for early visual bias toward particular animacy cues, there is also evidence for developmen-tal change in the perception of physical and social causality. Further, and most importantly, current knowledge especially in infant population is lim-ited in that it is based on global measures such as habituation and preference studies. In order to make definite the cause of visual attention is to empiri-cally examine gaze within the interaction. The second, bigger question, con-cerns categorizations of animates. That is, whether objects within dynamic interactions are categorized as social, evidenced by specific neural corre-lates.

Given these theoretical perspectives and available empirical evidence, I argue that current knowledge about the perception of animacy is incomplete and only in providing a more comprehensive presentation of the behavior can we reexamine what is known in light of these mechanisms – a point we return to in the General Discussion.

36

Aims of the thesis

The general aim of the present thesis is to augment knowledge about ani-mate motion perception. More specific aim is to examine mechanisms be-hind the formation of animated percepts. In addressing this aim, the studies here presented examined visual bias across development to relational cues that efficiently predict social interactions as well as visual attention during observation of interactions. The third study examined neural correlates in response to objects previously involved in an interaction.

The primary objective of the two eye-tracking studies (Study I and II) was to examine oculomotor behavior as a proxy for visual attention in in-fants and adults during observation of dynamic events. Specifically, the goal was to demonstrate that much like the adult, infant visual system is specifi-cally attuned to relational motion properties that efficiently predict social interactions. Previous research suggests that adults’ visual attention is biased toward interacting objects because the reduction in spatial distance between them is an efficient cue of an interaction (Meyerhoff, Schwan, & Huff, 2014; 2016). Here, we examined whether reduction in spatial distance alone biases visual attention in adults during online observation of motion without the presence of an interaction. Similarly, we examined whether the same bias may be evident in two developmental groups: 5- and 12-month-olds. Exam-ining visual sensitivity across development, allows examining the source of the bias. Specifically, it is possible that the presence of the visual bias to-ward reduction in spatial proximity is due to low-level visual differences rather than because it is an efficient cue for interaction. In present case, ab-sence of a developmental trend would support this possibility. Alternatively, presence of age-related differences may suggest that visual bias results from a growing sensitivity to aspects of the display beyond low-level stimulus features.

In Study II we further examined visual attention in the developmental group by testing the same group of 5- and 12-month-olds during the obser-vation of various types of interactions. In so doing, we examined whether object-specific, rather than relational, motion cues are attended. This aim builds on previous research indicating that at 5-months infants are able to discriminate between random and interacting moving shapes (Rochat, Mor-gan, & Carpenter, 1997), show sensitivity to spatio-temporal parameters and by 6-months they dishabituate to the reversal of the causal structure in sim-ple action-reaction sequences (Schlottmann, Ray, & Surian, 2012). Toward

37

the end of the first year, they are not only able to discriminate between ran-dom movement and coordinated motion but also between the roles of each object within interactions (Rochat, Striano, & Morgan, 2004).

Finally, whereas the first two studies were designed to examine how vis-ual system reacts to various motion cues pertaining to animacy, the third study aimed to examine whether infants perceive the social content of inter-actions. This was evidenced by differential event-related potentials (ERPs) in response to chasing objects. Previous research (Kaduk et al., 2013) has indicated that by 9-months infants allocate more attentional resources, marked by higher amplitude of the negative attentional component (Nc), in response inanimate compared to animate shapes. Here, we ask whether in-fants of the same age respond differently toward animate objects whose mo-tion is heat-seeking.

All three studies bear theoretical weight by addressing the question of on-tology and cognitive structure in the ability to perceive animates. More spe-cifically, whether the perception of socially causal events such as chasing reflects the workings of a specialized perceptual process that prioritizes cer-tain type of information, promotes learning and the existence of which serves specific evolutionary function.

38

Methods

Participants For Study I, 27 5-month-olds, 28 12-month-olds and 20 adult participants were recruited. Of these, six 5-month-olds, ten 12-month-olds and ten adults were tested but were excluded from the final sample due to not meeting nec-essary criterion. This criterion was defined a priori as > 50% gaze data. Consequently, the final sample in Study I consisted of 10 adults (8 female), 18 12-month-olds (9 female) and 21 5-month-olds (9 female).

For Study II, 31 12-month-olds and 70 5-month-olds were recruited (26 for Experiment 1, 25 for Experiment 2, and 19 for Experiment 3). In Exper-iment 1 two age groups were tested: 12- and 5-month-olds, but in Experi-ments 2 and 3 only 5-month-olds were tested. Of the recruited participants, 1 12-month-old and 2 5-month-olds were excluded due to failure to meet the set a priori criterion of > 20% gaze data. Two additional 12-month-olds were also excluded due to fussiness. The final sample for Study II was: 28 (16 female) 12-month-olds and 24 (12 female) 5-month-olds for Experiment 1; 24 (12 female) 5-month-olds for Experiment 2 and 18 (7 female) 5-month-olds for Experiment 3.

Some participants recruited for Study I were also participants for Study II. Specifically, the 28 12-month-olds recruited for Study I comprised the final sample in Study II (Experiment 1 only), meeting the set a priori criteri-on for Study II as > 20% gaze data. Of the 27 5-month-olds recruited in Study I, 24 were also included in the final sample of Study II (Experiment 1 only).

For Study III 67 9-month-old infants were recruited (34 for Experiment 1, 33 for Experiment 2). Of the recruited participants, 31 (46.2%) participants (16 in Experiment 1; 15 in Experiment 2) were tested but were excluded from the final analysis due to the failure in meeting the minimum inclusion-ary criterion set a priori as having 10 artifact-free trials per condition. This type of attrition rate has been found to be typical of this age range in study-ing ERP responses (Stets, Stahl, & Reid, 2012). Consequently, the final sample included 18 (6 female) 9-month-olds in Experiment 1 and 18 (6 fe-male) 9-month-olds in Experiment 2.

For all studies, all participants, infants and adults, reported no neurologi-cal or developmental disability and all had normal or corrected-to-normal vision.

39

Infant participants were recruited from local birth records in Uppsala. Upon birth of a child, parents received an introductory letter from Uppsala Child and Babylab at Uppsala University. The letter included voluntary sign-up sheet, which the parents sent back to the lab. Parents who indicated inter-est in participating in developmental research were then contacted via phone call or e-mail informing them about their child’s eligibility and purpose of the study.

Adult participants were also recruited through a voluntary sign-up process from a pool of university student body. All adult participants and parents of the infant participants received a gift certificate worth approximately 10 euros for participation. All studies were approved by the regional ethical committee and were conducted in accordance with the 1964 Declaration of Helsinki.

Stimuli All stimuli are available to view online: Study I: https://doi.org/10.6084/m9.figshare.1330197.v1 Study II: https://doi.org/10.6084/m9.figshare.4986185.v1 Study III: https://doi.org/10.6084/m9.figshare.4986182.v1 The stimuli for Studies I and II were 15-second animations depicting 3 iden-tical orange discs (1.26 visual angle in diameter; Hex color #ff6600) moving across black background. The animations were generated using Adobe Flash Professional CS5 and ActionScript 3 (Adobe Systems, San Jose, CA, USA). They were then edited using Adobe Premier Pro CS6, a video editing soft-ware. The animations for Study III were created by hand using Anime Stu-dio Debut 10, animation software.

For Study I, a set of video clips was generated with object trajectories turning rates of 0.1 radians per frame moving at a constant speed of 5 pixels per frame (ppf). Of this set, certain clips that fulfilled our criterion were chosen. The criterion was that only the video presentations, which did not include (1) overlapping objects, (2) objects that disappeared from the visible portions of the screen or (3) objects that interacted (bounced off) with the boundary. Consequently, although we refer to the motion in this display as ‘random’, they were not in fact such. That is, while the algorithm to generate the stimuli specified random, constant motion for all three objects, only those that satisfied a priori set criterion were chosen, effectively placing a restriction on what is ‘random.’ In so doing, we sacrificed complete empiri-cal purity in order to maintain the integrity of the goal of the study.

The random animations from Study I were also used in Study II as one of the control conditions, referred to as 'No interaction' events in Experiment 1. This experiment also had two additional conditions: chasing and following.

40

Chasing animations were generated in the same way as the random ani-mations except one of the orange discs (Agent Behind or AB) moved at a constant speed of 7.5 ppf toward second orange disc (Agent Ahead or AA) with 0o deviation from its path. Once the discs were within 90 pixels from each other, AA accelerated to 15 ppf until the distance between AA and AB increased, and then returned to its initial velocity. Throughout, AB moved toward AA at a constant speed, while AA continuously accelerated in order to get away from it. The third object, the distractor (D), moved randomly. Halfway through the animation (8 seconds after stimulus onset) AB changed the target of its chase and began approaching the distractor. At this point, previous D became AA.

The following animation was similar to the chase with one important dif-ference: once the AB got within 90 pixels of AA, the AA did not accelerate but both moved at a constant speed with AB decelerating to match the veloc-ity of AA.

Experiment 2 of Study II included chasing and following events without the target switch. Specifically, the motion algorithms for Chase-No Switch and for Follow-No Switch events were the same as those in Experiment 1 except in both cases AB never switched the target, and the D never became AB. In addition to the two no-switch events, the Experiment 2 included two no interaction events: No Interaction-Chase and No Interaction-Follow. Both events were created in order for the shapes to mimic velocity, acceleration and trajectory changes of the shapes in Chase and Follow events from Ex-periment 1, but there was no spatial relationship between the shapes. Specif-ically, in the No Interaction-Chase event, two discs moved at 5 ppf and the third disc at 7.5 ppf. After 2 seconds and again 6 seconds since the start of the stimulus, one of the slower discs accelerated to 15 ppf for approximately 1 second and then decelerated to 5 ppf. After 8 seconds since the start of the stimulus, the second slower-moving object accelerated and decelerated twice.

Similarly, in the No Interaction-Follow event all three discs moved ran-domly. Two moved at a speed of 5 ppf and one moved at a constant speed of 7.5 ppf. Two seconds from the stimulus onset, one of the slower moving discs and the faster moving disc decelerated to 4 and 3.5 ppf respectfully, thereby simulating the velocity changes in the following interaction. After 8 seconds since stimulus onset, the disc moving at 3.5 ppf accelerated back to 7.5 ppf for approximately 2 seconds, and then decelerated to 3.5 ppf. A third disc, which has been moving randomly at constant speed of 5 ppf now de-celerated to 4 ppf. The resulting event, therefore, was the same as the fol-lowing event in terms of velocity changes but without any spatial relation to each other.

All stimuli from Study I and Study II (Experiment 1 and 2) began in the same triangular configuration (we controlled for this configuration by run-ning an inverted condition for all studies). Experiment 3 of Study II, howev-

41

er, began differently. In this experiment, the stimuli were the same as those in Experiment 2 (Chase-No Switch and Follow-No Switch) but 2 seconds at the beginning of the video were trimmed out, in order to control for the ini-tial heat-seeking of the AB to AA. In this experiment, the stimuli lasted 13 seconds, and began with two discs on the lower left-hand side of the screen and the third shape (D) on the right-hand side.

For Study II, the stimuli were presented in three blocks, of which one trial was presented in inverted configuration. The order to presentation was coun-terbalanced for all experiments.

The stimuli for Study III differed from those of Studies I and II. Because Study III was measuring ERP response, rather than gaze, the stimuli were animations of an orange triangle or a blue rectangle and a gray circle fol-lowed by a still image of either the orange triangle or a blue rectangle. Spe-cifically, all infants initially saw 4 video animations (10 second each), then the static images, followed by a repeat of 2 initial animations. The logic behind this paradigm was that animation would establish object’s role in an interaction while the differences in ERPs would be measured as a response to the static image. The repeated instances of the initial animations served as a reminder of the initial interactions.

Video presentations Experiment 1 of Study III included a comparison between the chasing and inanimate event. Experiment 2 of Study III included chasing and no-interaction event. The background was a patterned green with two brown rectangles placed vertically on each side of the screen. Rectangles served as a context for the animation.

The Chasing interaction began with the main shape (orange triangle or blue rectangle) moving at a constant velocity of 0.4 visual degrees per frame (12ppf) toward the gray circle while it moved at the baseline velocity of visual degrees/frame (6ppf). When the chaser got within 2.5 visual degrees (80 pixels) from the target, the target accelerated to 4 times its initial veloci-ty (to 0.8 visual degrees/frame or 24ppf) for the duration of 24 frames after which it decelerated to its initial velocity. This type of acceleration and de-celeration of the target occurred 4 times during trial duration. A 4-note sound accompanied the animation and it corresponded to the increasing proximity of the chaser to the gray circle.

The Inanimate motion animation also depicted an orange triangle or a blue rectangle and a gray circle. Both objects moved at a constant velocity of 0.4 visual degrees/frame (12ppf) throughout the length of the trial. Both objects moved in a straight line, moving randomly and in no relation to each other. As the objects moved, they eventually bumped into the borders of the screen as well as the brown rectangles. When that occurred, the objects

42

bounced off in a random direction, maintaining the rule that the angle of the reflection equaled that of the incidence. As in chasing animation, the inani-mate stimuli were accompanied by a specific sound that happened when the main shaped bounced off of various surfaces.

The Random motion was similar to the chasing animation in that the main shape moved at a constant velocity of 12ppf and the gray circle moved at a baseline velocity of 6ppf, accelerating to 24ppf four times throughout the trial. In this animation, the shapes trajectories differed in that they each moved independently from each other. There was a 4-note sound that ac-companied the animation at the same time as in the chasing condition, mo-ments before the gray circle accelerated (thus it was contingent on the main shape’s 'approach' toward the target). The sound was the same as in the chas-ing condition, but in a lower pitch.

Test stimuli – Still images The animations were followed by static test images. The still images were 40 presentations (20 trials per condition) of a geometric shape (either an orange triangle or a blue square). Each presentation of the shape (1000 ms) with the corresponding sound (800 ms) was preceded by a presentation of a white fixation-cross against a black background (1000 ms). The shape was cen-tered against the same background used during the animations with no other objects present. The rectangle was 2.5 visual degrees (80 pixels) x 2.4 visual degrees (71 pixels); the triangle had a base of 3.1 visual degrees (98 pixels) and height of 2.4 visual degrees (77 pixels). Low-level perceptual cues such as luminosity were controlled and subject distribution between conditions was equal. An example of the shape images is also available to view online: https://doi.org/10.6084/m9.figshare.4986182.v1

Apparatus In Studies I and II, gaze was measured using a Tobii Model T120 corneal reflection eye tracker (Tobii Technology, Danderyd, Sweden). Gaze was recorded at 60 Hz (accuracy = 0.4 degree, precision = 0.16 degree) and pre-sented on a 17- inch (43-cm) flat screen monitor (1024 by 768 pixels) pre-sented at 30 fps.

In Study III, ERP signals were measured using an age appropriate 128-channel Geodesic Sensor Nets (HCGSN 130; EGI, Eugene, OR). The signal was referenced to the vertex and sampled at 250 Hz. It was then amplified by EGI Net amplifier (GES 300 Amp; EGI, Eugene, OR) and stored for off-line analysis. The stimuli were also presented on a 17-inch computer screen.

43

General procedure All participants were tested in the Uppsala Child and Babylab, at the De-partment of Psychology at Uppsala University in Uppsala, Sweden. After arriving to the lab, adult participants and the infants’ caregivers were pro-vided more detailed information about the purpose of the study as well as the procedure. They were then asked to sign a consent agreement. Adults were seated 60 cm from the Tobii screen, were asked to pay attention to the stimu-li and to limit their head movement. The 5-month-old infants were placed in a car seat that was placed on a chair 60 cm from the screen. Parents were seated behind the infants so as not to distract the infants from the screen. The older infants in Study II as well as all participants in Study III were seated on their parents’ lap approximately 60 cm from the screen. Parents were asked to keep their infants as still as possible. For Study III parents were also asked to hug their child slightly so as to prevent infants from pulling on the net. In both studies parents were also recruited to direct their infants attention back to the screen when the child began to look away, but to oth-erwise not comment or interfere with the presentation. All participants in Studies I and II passed the 5-point calibration procedure (Gredebäck, Johnson, & von Hofsten, 2010).

In Study I, participants observed three 15-second trials. Two trials were presented in an upright orientation and the middle trial was presented in an inverted orientation as a control (Figure 2).

Figure 2. Upright (2 trials) orientation of the starting positions for each trial. Invert-ed (1 trial) orientation was flipped vertically.

For Study II, 5 and 12 month old infants observed three blocks of trials (2 upright, 1 inverted) per condition. The order of the presentation was semi-random in that the same event was not identical to the last one.

The procedure for Study III was based on a procedure previously used to examine ERPs during still image based on previously-seen motion infor-mation (Gredebäck et al., 2015; Kaduk et al., 2013). For this study, all in-

44

fants observed 2 video animations for each condition (4 total) followed by still test images. Each infant was presented with 40 test trials (or until he/she stopped paying attention by looking away and moving). After the test imag-es, infants were presented with a reminder animation which was one presen-tation of each condition (2 total). Another set of test images followed the reminder animation.

On average in Experiment 1 of Study III, infants were presented 3-6 re-minder videos (M = 4.1) and approximately 56 static images (range: 39-74) of which 56 were chaser trials and 55.9 inanimate object trials. In experi-ment 2 of Study III, attended to 58.5 trials (range: 40-74) of which 58.6 were chaser trials and 58.4 were random agent trials.

Data analysis For Studies I and II recorded gaze data from Tobii were exported as text files, which were uploaded onto TimeStudio Version 2.8 (available at Figshare: https://figshare.com/articles/ TimeStudio_v2_7/1293476/2), a Matlab program specifically designed for analyzing time series data (Nyström, Falck-Ytter, & Gredebäck, 2016).

Deriving at High and Low Spatial Proximity intervals – Study I In Study I, for each frame, the TimeStudio program extracted the distances (measured from the center of the discs) between the objects. The program then calculated the ratio of the distance between the smallest and the next smallest distance. For example, if the distance between object A and object B was 1.4 cm, the distance between object B and object C was 1.2 cm, and the distance between object A and C was 0.9 cm, then the resulting ratio at this time frame was 1.2/0.9 = 1.34. The smallest possible ratio was 1.0 and it would result if all three discs were equally spaced forming an equilateral triangle. The higher the ratio, the closer two objects were relative to the third one. Please refer to the time series of distance ratios for the entire presented video (Figure 3).

45

Figure 3. Time series plot with high spatial proximity (in red) and low spatial prox-imity intervals (blue) identified across samples as the ratio of the second smallest distances between two objects by the smallest distance. When this ratio is high, two objects are close to each other and the third is distant. When this ratio is low, all three objects have the same distance to each other.

Once the distance ratio was known for each frame in the entire video, a threshold of 1.425 was used to indicate high and low ratio values (> 1.425 = ‘high spatial proximity’ [HSP]; < 1.425 = ‘low spatial proximity’ [LSP]). This threshold was used because it resulted in an even distribution of high and low proximity values. It effectively split the entire video presentation (15 sec) into two equal time frames (7.5 sec each), each with the total of eight intervals of low and high spatial proximity.

Dependent Variable – Study I Gaze was measured by counting fixations within a dynamic area of interest (AOI), manually defined as circular area 200 pixels in diameter (6.28 visual degrees) around each of the discs. Gaze falling outside the defined AOIs was excluded from the analysis, and gaze falling within more than one AOI at one time point was counted toward the AOI with the least distance to its center. Given this criterion, as long as the gaze was within defined AOI, it was included in the analysis, always in one of three AOIs. Randomly dis-tributed gaze would result in 33% (ratio 0.333) for each disc. The resulting dependent variable was oculomotor behavior operationalized as the ratio of looking time at any of the two proximate objects. For both HSP and LSP, gaze points in two proximal AOIs were calculated and divided by the total number of gaze points in all three AOIs. Consequently, at chance looking would result in a ratio of 0.667 (66%), higher value would signify a prefer-ence for proximate objects, and lower value would signify preference for

46

distal objects. Dynamic gaze plots for one subject from each age group are available online https://doi.org/10.6084/m9.figshare.1330223.v1

Dependent Variable – Study II Similar to Study I, in Study II, dynamic AOIs were defined, with a slightly smaller diameter of 150 pixels (4.73 visual degrees) around each of the discs. Percentage of looking time within each AOI was generated as a frac-tion of total fixation duration within each AOI divided by total trial length (15 sec). Trials < 10% total looking were excluded from the analysis. The dependent variable was a proportion of looking at individual AOI over total looking at all three AOIs for each trial. This proportion was then averaged across trials for each event type and used for the final statistical analysis.

The AOIs always reflected the role of the disc during the event. This means that the chasing disc was always called Agent Behind, the disc being chased or followed was Agent Ahead and the distractor disc was Distractor.

Dependent Variable – Study III For Study III, the continuous EEG data was first digitally filtered (0.3 – 30 Hz) and segmented from 300ms prior to the appearance of a still image to 1000ms after the image’s appearance.

Data from the most anterior and posterior electrodes (37 electrodes) were excluded from the analysis due to extremely high noise and artifact frequen-cy, often caused by poor contact with the infant’s scalp. Following segmen-tation, data were manually checked for artifacts in individual electrodes for each trial. These artifacts often resulted from channel glitches and strong drifts. Following the manual check, infants with less than 10 valid trials in each condition were excluded from further analysis. This minimum criterion has been used in previous studies and is deemed as the standard minimum in EEG studies using visual stimuli in infant population (Kaduk et al., 2013; Stets & Reid, 2011; Stets et al., 2012). Given this criterion, Experiment 1 of Study III averaged 13.3 trials (range: 10-23) for the chase condition, and 14.2 trials (range: 10-23) for the inanimate condition, Paired sample t-test = -1.25, p = 0.23. Experiment 2 of Study III included 13.4 trials (range: 10-24) for the chase condition, and 13.4 trials (range: 10-26) for the random condi-tion, Paired sample t-test = .039, p = .970.

Resulting segments were then re-referenced to average reference for all trials and were baseline corrected with the average amplitude between 0-300ms prior to the appearance of the test image. The data were then aggre-gated to individual averages for each trial type for each condition. All data sets were checked for outliers (± 3 z-score) but none were found.

Several regions of interest were chosen from the electrodes included in the final analysis. These regions were determined based on prior studies examining these specific ERPs covering areas of low occipital and temporal areas (Bakker, Kaduk, Elsner, Juvrud, & Gredebäck, 2015).

47

The first component to be examined was the Nc (Kaduk et al., 2013), which included 14 channels (channel numbers: 5, 6, 7, 12, 13, 20, 29, 36,104, 105 [C4], 106, 111, 112, 118). The average amplitude for this com-ponent was a time interval ranging from 400-600ms after the appearance of the image. The DV for the Nc component was measured in a Paired sample t-test with Agent type as an independent variable.

Second was examined the P400 component which included 13 channels over posterior area (all channel numbers: 62, 66, 67, 70 [01], 71, 72, 74, 75 [Oz], 76, 77, 82, 83 [02], 84; of which left channels were: 66, 67, 70, 71,74; and right channels were: 76, 77, 82, 83, 84). The average amplitude for this component was a time interval ranging from 350ms – 600ms after object appearance. The DV for the P400 was entered in the General Linear Model (GML) with the Agent type (Chaser, Inanimate object) and Hemisphere (Left, Right) as within-subject factors.

The focus on the Nc and the P400 components was motivated by the hy-pothesis that if the difference between the events is largely attentional, we would expect differences in the Nc. Specifically, if the Nc component corre-sponds to the presence of animate motion in general, as was previously re-ported (Kaduk et al., 2013), we would expect higher negativity in the Nc component corresponding to the inanimate object (Experiment 1). Alterna-tively, if infants process the chasing object in terms of the social content, we would expect differences in the P400 previously found in response to social-ly valanced information across Experiment 1 and 2 (Gredebäck et al., 2015; Melinder, Konijnenberg, Hermansen, Daum, & Gredebäck, 2015).

48

Study I – Visual attention to dynamic spatial relations in infants and adults

Study I was designed to examine gaze during online observation of dynamic events. Specifically, whether reduction in spatial proximity between moving geometrical shapes encourages visual bias in adults and infants.

As reported in the Introduction, animate motion perception has been stud-ied using limited visual displays such as dynamic presentation of geomet-rical shapes. Here, we combine this established method with novel analysis tool to determine online oculomotor behavior in both adults and infants at 5- and 12-months.

Previous research established that in multiple object displays, adults are visually biased and are therefore quicker at identifying shapes that are inter-acting in some way. One explanation is that adults use relational cues, such as the reduced spatial distance between the objects, as an efficient cue pre-dicting presence of an interaction (Meyerhoff, Schwan, & Huff, 2014; 2016). The outstanding question is whether relational properties outside the context of an interaction similarly bias attention. By examining online visual attention during observation of random motion in adults, we examine the extent of the visual bias toward objects with reduced spatial distance.

As a secondary hypothesis, we examine the developmental trajectory of the bias by testing 5- and 12-month-old infants. Previous research has shown that infant visual attention, especially in the first months of life, is marked by orientation toward stimuli of increased saliency. For instance, moving as opposed to static stimuli, stimuli of high-contrast or high visual complexity (Cohen, 1972; Dannemiller, 2000; Frantz & Fagan, 1975; Hood, 1995; Slat-er, 1995). Examining gaze across development allows speculating about the mechanisms behind it. Specifically, whether the attentional bias toward spa-tially close objects reflects bias toward low-level saliency. Absence of age-related differences and a preference for spatially close objects across age groups would support this possibility. Otherwise, presence of a developmen-tal trend would allow the possibility that visual bias toward spatially proxi-mal objects in adults results from growing sensitivity to relational properties of the display rather than low-level stimulus features. Developmental pro-gression would reflect attentional processes for the interpretation of motion, such as the anticipation of an interaction.

49

Design Participants were 5-, 12-month-olds and adults presented with a dynamic display depicting three identical two-dimensional discs. The stimulus videos (15-sec long) depicted discs moving in no relation to each other, never con-tacting each other or the boundary of the screen. The discs moved at a con-stant speed of 5ppf with trajectories generated randomly with turning rates of 0.1 radians per frame. All participants viewed the event 3 times.

First, data analysis included analysis of the stimuli. The distance between the discs was extracted for every time point, ratios between the second smallest and the smallest distance was calculated. High and low spatial prox-imity intervals (those more and less than 1.425) were determined. Prefer-ence for two proximal objects would be >.667, while a preference for the distant object would be <.667.

Second, was the examination of the participants’ gaze falling within cir-cular AOIs (200 pixels in diameter) corresponding to the three dynamic discs. The dependent variable was an average ratio of looking time on the two proximal objects divided by the sum of looking at all three AOIs.

Results A 2(HSP, LSP interval) x 3(Age group) Repeated Measures ANOVA re-vealed main effect of Interval type (F(1, 46) = 23.459, p < .001, η2 = .338) indicating that all participants looked more at the adjacent objects during HSP interval than during LSP interval MDifference = .079, p < .001. There was also main effect of Age group, (F(2, 46) = 15.205, p < .001, η2

= .398), indicating that 5-month-olds looked significantly less at the adjacent objects than the 12-month-olds MDifference = .09, p = .002 and the adults, MDifference = .149, p < .001. The interaction effect between the two factors was not sig-nificant.

Follow-up analysis (Figure 4) revealed that both adults (p = .004) and 12-month-olds (p ≤ .001), differentiated between the two interval types, looking more during intervals of HSP than LSP. During intervals of HSP, both groups also looked significantly above chance (.667) at the adjacent objects over objects that were distal. 2

2 During spring 2017 we had the possibility to test 16 more adult participants with this para-digm. The findings confirm the initial findings in showing that adults look more at objects that are closer relative to the third object during instances of HSP, (M = .79, SE =.02), t(15) = 5.68, p < .001, as well as LSP, (M = .77, SE =.02), t(15) = 6.98, p < .001.

50

The difference between intervals was not significant for the 5-month-olds (p = .052), and during HSP, 5-month-olds looked at chance at adjacent ob-jects (p = .359). However, during intervals of LSP, 5-month-olds looked significantly below chance at the adjacent objects (p = .007). This means that they failed to detect adjacent objects when the distance was relatively equal, and in fact, they looked more at the objects that were further apart.

Control analysis revealed that across the three groups, 5-month-olds seemed to spend less time looking within the specified AOIs compared to the adults and the 12-month-olds. However, differences in total looking did not correlate with the found distribution patterns within each age group.

Figure 4. Percent looking time at the two most spatially close objects during high spatial proximity intervals and low spatial proximity intervals in 5-, 12-month-olds and adults. Looking randomly between all three objects would yield a percentage of 66.67% (dashed line).

These findings suggest that while there are differences in total looking be-tween groups, this difference does not influence the selective looking at the target objects during HSP.

51

Discussion Study I Present eye-tracking study was designed to (I) investigate the presence of a visual bias toward spatially proximal objects outside the context of an inter-action in adults as well as (II) the developmental trajectory of this ability.

In line with previous research (Meyerhoff, Schwan, & Huff, 2014; 2016) we found that adults’ visual attention is biased toward relational properties such as the reduction in spatial distance. That is, when the two objects come closer together, adults direct their gaze toward those objects. Importantly, unlike previous studies, we found this bias to be evident during online ob-servation of motion when the objects were not interacting.

Results further indicate that at 12-months infants, much like adults, are sensitive to a reduction in spatial proximity. During intervals of high spatial proximity, adults and 12-month-olds direct their gaze toward objects that are closest to each other. During intervals of low spatial proximity, in which the objects are relatively equally distant from each other, 12-month-olds’ look-ing time was no greater for any of the objects. Adults were better than 12-month-olds in detecting even the slightest difference in proximity. That is, even during instances of low spatial proximity, they looked more at objects that were closer than those that were further apart.

For the 5-month-olds, the bias was not evident. While the 5-month-olds were able to detect a difference between instances of high and low spatial proximity, this difference was not significant and in neither case was the percentage of looking at adjacent objects above chance. In fact, during in-stances of low spatial proximity, they looked significantly below chance at objects that were close.

The findings from this study are two fold. First, results show that the presence of an interaction between objects is not necessary for the attention-al bias in adults. Randomly moving objects elicit specific looking pattern based on dynamically changing relational properties. That is, adults’ visual system seems to be sensitive to very subtle relational cues, such as cues that normally predict interaction. The 12-month-olds are equally as sensitive to changes in spatial proximity but only when these are more evident. That is, during instances of high but not low spatial proximity. Five-month-olds did not seem to show the same looking profile as the other two age groups, fail-ing to differentiate between intervals and showing no preference for the spa-tially proximal objects.

Second, the presence of a developmental trend suggests that the visual bi-as is not due to low-level visual saliency. Rather, it may reflect attention toward objects in anticipation of an interaction, with a possible caveat that for the 5-month-olds, more obvious cues or potentially longer stimuli presentation is necessary. In line with this interpretation, others have pro-posed that certain attentional processes reflected in oculomotor behavior can occur prior to recognition of an interaction and in turn effective recognition

52

of motion patterns can only occur at the locus of visual attention (Meyerhoff, Huff, & Schwan, 2013). Based on our findings, we suggest that a visual bias for reduced spatial proximity occurs because spatial proximity is an efficient cue for interaction and in many cases predicts an interaction.

To further understand how other motion cues influence infant looking it is important to examine infant visual attention during observation of interac-tions. This was the purpose of Study II.

53

Study II – Infants’ preference for individual agents within chasing interactions

Findings from Study I suggest that visual system is sensitive to certain rela-tional properties that efficiently predict interaction when no interaction is present. This sensitivity is marked by development. While for the older in-fants and adults this trend was evident, it was not so for the 5-month-olds. Given the similarity between adults and 12-month-olds in Study I, Study II focused on infancy. Specifically, in Study II gaze in 5- and 12-month-old infants was examined during observation of interacting and non-interacting objects. Using global measures of preference, previous research found that by 5-months infants effectively differentiate between geometrical shapes that are interacting and those that move randomly (Rochat et al., 1997). Old-er infants toward the end of their first year of life are also found to differen-tiate between a change in the relationship between interacting objects sug-gesting increased sensitivity to the event’s causal structure (Rochat et al., 2004).

To our knowledge, no study examined how interacting shapes influence online visual attention. In examining this question, present study focused on a chasing interaction, a type of socially causal event that is relevant in evolu-tionary sense (Frankenhuis & Barrett, 2013; Frankenhuis et al., 2013b). We hypothesize that a resulting visual profile may reflect phylogenetically old attention mechanism designed to focus on the most reliable features in mo-tion patterns. In light of this perspective, we predict that although at 5-months infants do not attend to certain relational properties (such as reduced spatial proximity) they may attend to other object-specific motion cues. In chasing, these may reflect focus on the goal-directed motion of the most informative object, the chaser. In turn, interacting objects whose motion is synchronized but that does not rely on goal-directedness, we predict undif-ferentiated looking.

Design In the three experiments, gaze was examined while watching non-contact causal interactions between three two-dimensional discs. In Experiment 1, we examined how infants observe objects that are chasing, using a ‘Waver-

54

ing wolf’ paradigm (Gao et al., 2009) in which a single chasing object (the ‘wolf’) chases and changes the target of its chase (the ‘sheep). This type of paradigm allows examining how objects are tracked and how gaze reacts to target changes.

In this experiment, 5- and 12-month-old infants were presented with a chasing interaction (goal-changing/wavering wolf) and two control condi-tions: following (also goal-changing/wavering wolf) and no-interaction event. To recap, in both chasing and following conditions, the chas-ing/following object (Agent Behind or the ‘wolf’) initially accelerated to-ward one of two targets (the ‘sheep’). In both interactions, the ‘wolf’ switched the target mid-presentation by accelerating in the direct path to-ward the second target. The qualitative difference between the chasing in following condition was that while in the chasing interaction the target ac-celerated away from the chaser, the target did not do so in the following condition. The ‘wolf’ in the chasing interaction moved at a constant speed (7.5ppf), while the ‘wolf’ in the following interaction adjusted its velocity to that of the target (5ppf).

In Experiment 2 and 3 only the younger group, the 5-month-old infants, were tested. In Experiment 2 they were presented with chasing and follow-ing interactions in which there was a single goal (Chase-No switch, Follow-No switch) and two no interaction events in which all low-level motion cues were identical to chasing and following interactions (No interaction-Chase event, No interaction-Follow event respectively) except for the contingency between the objects.

In Experiment 3, the initial heat-seeking motion was removed and infants observed a chasing interaction without the initial acceleration of the ‘wolf’ toward the ‘sheep’ (Chase-No initiation) and an equivalent following inter-action (Follow-No initiation). In all three experiments, participants observed three presentations of each event type, controlling for orientation (Figure 5).

55

Figure 5. Starting positions of the discs with transparent overlays indicating areas of interest (not included in the actual stimulus presentation) for the Agent Behind, Agent Ahead, and the Distractor in the upright (left) and inverted (right) orienta-tions.

Results Experiment 1- Chase, Follow, No-interaction Proportion of looking was examined for each event type and within each AOI corresponding to the three observed objects. The analysis revealed a significant main effect of AOI type F(2,24) = 36.70, p < .001, η2 = .75, and a significant Event by AOI type interaction, F(4,48) = 11.19, p < .001, η2 = .48.

Follow-up analyses showed that across event types, both 12- and 5-month-olds preferred to attend to the Agent Behind more than the Agent Ahead, MDifference = .166, p < .001, but looked at the Agent Ahead more than the Distractor, MDifference = .21, p < .001.

While there was no main effect of age, we examined looking within each age group due to our a priori hypothesis based on age-related differences we found in Study I.

For the 12-month-olds (Figure 6), analysis showed that during the Chase event, infants tracked the Agent Behind more than the Agent Ahead, MDifference = .19, p < .001, and looked more at the Agent Ahead than the Dis-tractor, MDifference = .09, p = .002. The pattern was similar during the Follow event where the 12-month-olds looked more at the Agent Behind than the Agent Ahead MDifference = .14, p < .001, but they did not differentiate between the Agent Ahead and the Distractor. In the No-Interaction event, there were no significant differences between the AOIs, p = .815.

56

Figure 6. Average proportion of looking time at the Agent Behind, Agent Ahead and Distractor areas of interest by Event type in 12-month-olds (Experiment 1). Vertical lines indicate mean standard error.

The 5-month-olds’ looking pattern differed slightly (Figure 7). During the Chase event, 5-month-olds preferred to look at the Agent Behind more than the Agent Ahead (MDifference = .27, p = .001) but unlike the older group, they looked equally at the Agent Ahead and the Distractor (MDifference = .10, p = .187). Importantly, in the Follow event, the 5-month-olds did not seem to prefer to look at the Agent Behind over the Agent Ahead MDifference = .16, p =.056. They also looked similarly at the Agent Ahead and the Distractor, MDifference = .114, p = .204. And finally, in the No-Interaction event, there were no differences between the AOIs, p = .870.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Chase Follow No-interaction

Aver

age

prop

ortio

n of

look

ing

time

Event type 12-month-olds

Agent Behind Agent Ahead Distractor

57

Figure 7. Average proportion of looking time at the Agent Behind, Agent Ahead and Distractor areas of interest by Event type in 5 month olds (Experiment 1). Verti-cal lines indicate mean standard error.

Experiment 2 – Chase - No Switch, Follow - No Switch, No-interaction-Chase, No-interaction-Follow As in Experiment 1, the main analysis revealed a significant main effect of AOI type F(2,26) = 27.54, p < .001, η2 = .68 and a significant Event type by AOI type interaction F(4,52) = 30.94, p < .001, η2 = .70.

Follow-up analysis (Figure 9) showed that in general 5-month-old infants looked more at the Agent Behind than the Agent Ahead, MDifference = .14, p = .004, but more at the Agent Ahead than the Distractor, MDifference = .08, p = .001.

When observing Chase-No switch event (Figure 8) in which the chaser (Agent Behind) had only one target (Agent Ahead), infants preferred to look at the Agent Behind more than the Agent Ahead, MDifference = .31, p < .001, but did not prefer the Agent Ahead over the Distractor, MDifference = .05, p = .151.

When observing the Follow-No Switch event, infants preferred to attend to the Agent Behind over the Agent Ahead, MDifference = .15, p = .001. But unlike in the Chase-No switch event, here they preferred the Agent Ahead over the Distractor, MDifference = .19, p < .001. This means that infants attended more to the target when it was part of a follow interaction, but did not seem to attend to the target more than the distractor when the target was being chased.

There were no differences between the AOIs in the two No-Interaction events.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Chase Follow No-interaction

Aver

age

prop

ortio

n of

look

ing

time

Event type 5-month-olds

Agent Behind Agent Ahead Distractor

58

Figure 8. Average proportion of looking time at the Agent Behind, Agent Ahead and Distractor areas of interest by Event type in 5 month olds (Study 2). Vertical lines indicate mean standard error.

Comparison between Experiment 1 and 2 Finally, a comparison between No-Interaction event in Experiment 1 and the two No-Interaction events in Experiment 2 did not reveal any differences in distribution of looking between the AOIs (for all AOIs ps > .20), nor were there any differences in the total percentage of looking at the No-Interaction events between the two studies (p = .074).

Experiment 3 – Chase-no initiation, Follow-no initiation As in Experiment 1 and 2, the main analysis revealed main effect of AOI type, F(2,28) = 43.86, p < .001, η2 = .76 and a significant Event type by AOI type interaction F(2,28) = 13.95, p < .001, η2 = .49. Like in previous experi-ments, the main effects indicated that in general, infants looked more at the Agent Behind than the Agent Ahead, MDifference = .24, p < .001, and looked more at the Agent Ahead than the Distractor, MDifference = .18, p < .001.

However, analyzing each event separately revealed some important dif-ferences. In the Chase-no initiation event (Figure 9), infants continued to look more at the Agent Behind than at the Agent Ahead, MDifference = .41, p < .001, but did not seem to prefer to look any more on the Agent Ahead than the Distractor, MDifference = .04, p = .357.

Critically to the hypothesis of the paper, in the Follow-no initiation event, infants looked equally at the Agent Behind and the Agent Ahead MDifference =

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Chase - No switch Follow - No switch No-interaction-Chase No-interaction-Follow

Aver

age

prop

ortio

n of

look

ing

time

Event type

Agent Behind Agent Ahead Distractor

59

.05, p = 1.00, but continued to look more at the Agent Ahead than the Dis-tractor, MDifference = .29, p < .001. This pattern indicates that the removal of the initial heat-seeking effectively eliminated the preference for the Agent Behind when it followed the Agent Ahead, but not when the Agent Behind was a chaser.

Figure 9. Average proportion of looking time at the Agent, Agent Ahead and Dis-tractor areas of interest by Event type in 5-month-olds (Study 3). Vertical lines indi-cate mean standard error.

Comparison between Experiment 2 and 3 Finally, we compared findings from Experiments 2 and 3. We did this in order to examine whether elimination of total heat seeking influenced look-ing preferences. By performing a 2(Event Type) x 3(AOI Type) repeated measures ANOVA with Experiment (1 or 2) as a between-subject factor. This analysis revealed a significant three-way interaction with Experiment type F(2,74) = 3.86, p = .025, η2 = .095, but no main effect of Experiment type F(1,38) = 1.480, p = .231, η2 = .037

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Chase-no initiation Follow-no initiation

Aver

age

prop

ortio

n of

look

ing

time

Event type

Agent Behind Agent Ahead Distractor

60

Comparison between Experiment 1, 2 and 33 For the purpose of this thesis, I ran a comparison between all three experiments. By performing a 2(Event Type) x 3(AOI Type) Repeated Measures ANOVA with Experiment (1, 2, or 3) as a between-subject factor, the analysis demonstrated a significant three-way interaction F(4,114) = 3.464, p = .010, η2 = .108, but no main effect of Experiment type F(2,57) = 3.167, p = .05, η2 = .10

In sum, across the three experiments we find that when presented with a chasing interaction, infants show a preference for the agent displaying heat-seeking motion. To a lesser degree, they attend to the target of the chase, and least to the object not involved in the interaction. By gradually modulating the amount of heat-seeking motion and eliminating it completely in Experi-ment 3, we find that the preference for the chasing object disappears and infants attend to synchronously moving objects equally.

Discussion Study II Findings from Study I indicate that even in the absence of a causal relation-ship, adults and older infants are biased toward object that momentarily have reduced spatial proximity. We reason this bias is due to bias toward relation-al cues that efficiently predict an interaction between objects. Here, we ex-amine whether presence of object-specific motion cues such as goal-directed motion through heat-seeking influence gaze in young infants.

Across the three experiments in Study II, we find a distinctive preferential looking pattern during the observation of interacting geometrical shapes. Non-interacting moving shapes did not elicit any preferences. In contrast, chasing as well as following interactions in Experiments 1 and 2 have demonstrated object-specific preference for the agent whose motion was distinctly target-directed and heat-seeking. To a lesser degree, infants looked at the other object involved in an interaction, the target. This pattern re-mained across events in which the chasing object switched the target (Exper-iment 1) or maintained the same target throughout the trial (Experiment 2). What seemed to effect the preference for the chaser was the heat-seeking motion. In Experiment 3, when heat-seeking motion was completely elimi-nated from the following interaction, there was no preference between the two interacting objects. Instead, the two objects were tracked equally, possi-bly viewed as a single unit composed of two synchronously moving compo-nents. Finally, across all conditions, least visual attention was directed to an object in the display that did not interact with other objects (the Distractor).

3 Not included in the original article

61

Previous research found that more than self-propulsion, goal-directed motion is a feature of animacy, and objects that move in goal-directed way are more likely to be judged as alive (Opfer, 2002). Here, along with others (Frankenhuis et al., 2013b), we extend these findings in showing that this cue contributes to the general preference for animate motion. Unlike previ-ous studies we distinguish between generally conceptualized goal-directed motion and motion that is specifically heat-seeking. That is, motion effi-ciently directed toward a salient goal.

Given this distinction, we present a case for early-emerging preference for agent-specific motion cues. Namely, we find that the bias toward objects whose motion is heat-seeking is evident starting at 5-months when other relational properties (such as reduction in spatial proximity) are not readily detected. With the current level of analysis and the developmental stability evident in Study II Experiment 1, it is possible to conclude that sensitivity to heat-seeking motion reflects the workings of a phylogenetically old percep-tual mechanism which prioritizes heat-seeking motion.

However, although the visual bias informs about how interactions are at-tended, it does not inform about whether the visual input effects later catego-rization. To do so, a more direct examination of the neural processes in-volved is a necessary next step, one that we have undertaken in Study III.

62

Study III- Neural correlates of chasing motion

Whereas Studies I and II examined oculomotor behavior during the observa-tion of dynamic events, the overarching aim of Study III was to examine whether animate displays are categorized as social events. That is, whether observation of certain motion patterns elicits specific neural responses con-sistent with responses attributed to other social information. Growing re-search area suggests this might be the case. Adult studies, for instance, have consistently shown that animate motion elicits activation of the area normal-ly responsive to socially-valanced information (Gredebäck et al., 2015; Melinder, Konijnenberg, Hermansen, Daum, & Gredebäck, 2015). At the same time, relatively few studies have been devoted to examining neural correlates to animate motion in infancy.

One previous finding with 9-month-old infants (Kaduk et al., 2013) sug-gest that differentiation between animate and inanimate motion is a percep-tual, rather than a conceptual, process. They found that infants allocate more attentional resources (evidenced by the increased negative amplitude of the Nc component) to objects that move inanimately, suggesting that animate motion is more salient (Elsner & Gerber, 2010) and fewer cues are necessary to identify animate motion as such (Scholl & Tremoulet, 2000). In Experi-ment 1 we examine whether indeed animate motion is more salient than inanimate motion evidenced by differential amplitude of the Nc component.

Alternatively, animate motion may form the basis for social categoriza-tion. Specifically, by contrasting interacting (chasing) and non-interacting (random) motion, we examine whether presence of an interaction elicits specific ERP component found to be responsive to social information, the P400. Building on previous findings (Study II) we test whether objects whose motion is heat-seeking in a chasing interaction are categorized as social agents.

Design In Experiments 1 and 2, 9-month-old infants were presented with 2 presenta-tions of 10-second video displays from two conditions. The videos depicted two dynamic geometrical shapes. The two shapes were the main shape (ei-ther a blue square or an orange triangle) and a gray ball.

63

In Experiment 1, chasing interaction was compared to inanimate motion. In the chasing interaction, the main shape (for instance, the blue square) always accelerated toward the gray ball, while the gray ball accelerated con-tingently away from the main shape. In the inanimate condition, the main shape (here, the orange triangle) and the gray ball did not move autonomous-ly. Instead, they bounced off the screen boundaries.

In Experiment 2, chasing motion was compared to random animate mo-tion. Random motion included the main shape and the gray ball moving autonomously but in no relation to each other – they did not bounce off the boundaries of the screen but moved autonomously in no relation to each other.

For each experiment, after two initial video presentations, infants were presented with 40 still images of the main shape from the chasing condition and the corresponding shape in the comparison motion (an inanimate main shape in Experiment 1 and random main shape in Experiment 2). The pur-pose of presenting a still image following video presentation allows to time-lock ERPs within a specified segment. After the test images, infants saw one presentation of each motion video again as a reminder.

Please refer to Figure 10 for the schematic description of the design.

Figure 10. Stimuli used during the initial video presentation (initial and reminder) and test images. The identity of the chaser (here a blue rectangle) and the inanimate object(here an orange triangle) in Study 1 and random object in Study 2 was coun-terbalanceld across participants. The trajectory represented here reflects the first 4 s of the video. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

64

Results Experiment 1 – Chasing vs. Inanimate motion Two components of interest were examined. P400 component comprised of the lateral posterior region including low occipital and post-temporal chan-nels. The second was the Nc component, which included areas in the central frontal channels. Given previous findings suggesting hemispheric lateriza-tion in adult samples (Gao, Scholl, & McCarthy, 2012), we included the hemispheres as a factor. Results are demonstrated in Figure 11.

Nc component Since the Nc component includes centrally located effect, laterization was not examined. Instead, a Paired sampled t-test indicated a higher amplitude when viewing the Chaser (3.38 μV, SE = 4.45) compared to the Inanimate object (.28 μV, SE = 4.75), but this difference was not significant, t(17) = 1.79, p = .091.

P400 component A 2(Agent Type: Chaser, Inanimate Agent) x 2(Hemisphere: Left, Right) Repeated Measures ANOVA demonstrated a significant main effect of Agent type F(1, 17) = 5.05, p =.038, η2 = .229, as well as Hemisphere F(1, 17) = 48.39, p <.001, η2 = .740. The follow-up analysis revealed higher am-plitude when viewing the Chaser (9.45 μV, SE = 2.25) rather than the Inan-imate object (4.32 μV, SE = 1.86), especially on the right (12.48 μV, SE = 1.92) rather than left (1.28 μV, SE = 1.88) hemisphere. The interaction be-tween the Agent Type and Hemisphere was not significant F(1, 17) = .944, p =.345, η2 = .053. Following visual inspection, we further discovered two trends in the wave-form. The first was a positive peak around 200ms after stimulus onset (P200) and the second was a negative dip around 290ms following stimulus onset (N290). We analyzed these two and found that for the P200 there was a slightly higher amplitude for the Chaser (4.87 μV, SE = 2.42) than the Inanimate object (.19 μV, SE = 2.45), but this difference was not significant, t(17) = 1.99, p =.062. For the N290, however, we found a significant differ-ence between the agents t(17) = 2.30, p = .034, with a positive amplitude to the Chaser (1.67 μV, SE = 9.27) and negative to the Inanimate agent (- 2.87 μV, SE = 9.38).

Finally, we examined whether the number of reminder video presentations correlated with any of the tested components. We found that none correlated significantly (all ps > .05).

65

Figure 11. Grand average ERP data for selected channels for the Nc component (in green; top graph) and P400 (in red) component of the Left (bottom left graph) and Right(bottom right graph) hemisphere with time of interest shaded in gray. (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)

Experiment 2 – Chasing vs. Random motion As in Experiment 1, we again examined the Nc component found to signify attentional differences to the animate and inanimate motion. We also consid-ered the P400 component found to relate with the activation of the pSTS. In addition, we tested the presence of two components found in Experiment 1 – the N290 and P200. Results are demonstrated in Figure 12.

Nc component As in Experiment 1, there were no differences in the Nc component, t(17) = .586, p = .566 with positive amplitudes for both agents and slightly higher positive amplitude for the Chaser (1.26 μV, SE = .88) than the Random ani-mate agent (.87 μV, SE = .75).

P400 component A 2(Agent Type: Chaser, Random Agent) x 2(Hemisphere: Left, Right) Repeated measures ANOVA demonstrated a significant main effect of Agent Type F(1, 17) = 12.20, p =.003, η2 = .418 as well as Hemisphere F(1, 17) = 44.47, p <.001, η2 = .723. As in Experiment 1, follow-up analysis indi-

66

cated higher amplitude for the Chaser (8.48 μV, SE = 2.70) than the Random animate agent (-.221 μV, SE = 2.69), and just like in Experiment 1 this effect seemed to be laterized with higher amplitude in the right (13.26 μV, SE = 3.05) than left (-5.0 μV, SE = 2.44) channels. The interaction between Agent Type and Hemisphere was not significant, F(1, 17) = .472, p =.501, η2 = .027.

N290 component Similar to Experiment 1, the difference at the N290 was also significant, t(17) = 3.60, p = .002 with a positive amplitude for the Chaser (3.38 μV, SE = 3.21) and a negative amplitude for the Random animate agent, (- 6.56 μV, SE = 3.33).

P200 component In contrast to Experiment 1, there was a significant difference between the two agents at the P200 t(17) = 2.88, p =.010, with slightly positive amplitude for the Chaser (.59 μV, SE = 2.42), and a negative amplitude for the Random animate agent, (-6.85 μV, SE = 2.64).

Finally, as in Experiment 1, for all the components tested, none correlated significantly with the number of reminder videos (all ps > .05).

67

Figure 12. Grand average ERP data for selected channels for the Nc component (in green; top graph) and P400 component (in red) of the Left (bottom left graph) and Right(bottom right graph) hemisphere with time of interest shaded in gray. (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)

Discussion Study III The present study examined whether neural correlates in response to the agent previously observed chasing another differs for the neural correlates in response to an object previously seen moving inanimately.

The study design was such so as to allowed time-locked ERP response to geometrical shape previously involved in the dynamic event. Consequently, across two experiments, chasing object was contrasted with an object whose motion was inanimate (Experiment 1) and an object whose motion was ani-mate but non-interacting (Experiment 2).

Unlike in previous studies (Kaduk et al., 2013), findings here reported did not find attentional differences as measured in the amplitude of the Nc com-ponent. The null effect here, seems to suggest no differences in the low-level visual input from these motion types, as neither seemed to have elicited ad-ditional attentional resources.

Rather, findings from the two types of comparisons suggests that the agent previously involved in a chasing interaction elicits higher positive

68

amplitude in the low occipital and post temporal areas at around 400ms post stimulus onset, an ERP component known as the P400. This particular com-ponent was present during both types of comparisons, meaning that it is particularly involved in response to motion that signifies interaction – goal-directedness, such as heat-seeking, and contingency, rather than as a re-sponse to self-propulsion alone.

Along with the P400, present findings indicate a specific waveform pat-tern in response to agents previously involved in the inanimate and random motion. For both these objects, there was a more negative drop around 290ms (N290) and a lower P400. The N290 component has been associated with higher novel stimuli and longer visual processing (Key & Stone, 2012; Key, Stone, & Williams, 2009; Scott, Shannon, & Nelson, 2006).

The N290/P400 component has been suggested to be a precursor to the adult N170, localized to the fusiform gyrus and the STS activated during the perception of social stimuli. Furthermore, the P400 has been found to index adult pSTS activity which has been evident in response to chasing interac-tions and other animate and intentional motion (Castelli et al., 2000; Lee et al., 2012; Martin and Weinsberg, 2003; Shultz et al., 2004; 2005; Gao, Scholl, & McCarthy, 2012). Infant P400 we find here may in parallel reflect activation elicited from the same type of information. This effect was espe-cially prominent in the right hemisphere, which corroborates with previous findings in adults (Gao, Schultz & McCarthy, 2012) and has been found to be sensitive to goals and intentions (Gao et al., 2012), as well as successful goal-directed outcomes (Schultz et al., 2011).

In addition to the N290/P400 component, the chaser was responsible for an increase in the P200 component, especially evident when it was com-pared to the randomly moving object in Experiment 2. Previous research attributed an early P200 component to be responsible during the formation of perceptual categories as well as the visual discrimination and recollection between previously seen items.

Present findings suggest the perception of chasing reflects different pro-cessing compared to inanimate and random motion. We argue it parallels in response to other socially-valenced information.

Notably, the presence of differential neural responses in the ERPs nor-mally associated with social information suggest very fast social categoriza-tion based on motion information in infants at 9-months. It contrasts the view that the observed preference for social motion in infants reflects low-level perceptual processing of individual motion parameters, which should have been reflected in the differential attentional component, the Nc.

69

General discussion

Present work comprises of three empirical studies whose aim was to exam-ine mechanisms behind the formation of animated percepts. Benefiting from a novel method of analyzing gaze, Studies I and II inform about oculomotor behavior during observation of animate motion. Using measures of event related potentials, Study III informs about neural correlates for object cate-gorization based on motion.

The goal of Study I was to examine visual bias in response to cues that ef-ficiently predict the presence of an interaction. Previous research has shown that adult observers’ visual attention toward interactions may result from a bias toward objects with reduced spatial proximity (Meyerhoff et al., 2014b). Building on these findings, Study I examined the presence of a visual bias in adults during online observation of motion. Unlike previous efforts, the re-duction in spatial proximity was presented as an isolated cue outside the context of an interaction, effectively disentangling the empirical confound in studying the effect of spatial distance and interactions. Additionally, visual bias toward spatial proximity was examined developmentally in infants at 5- and 12-months of age to determine whether the presence of the bias can be attributed to low-level saliency.

In Study II, gaze was examined during observation of socially causal non-contact interactions such as chasing and following by infants at 5- and 12-months. Examining gaze during observation of interactions allows determin-ing visual attention to object-specific, rather than relational (spatial proximi-ty), motion properties. Previous research suggests that by 5-months infants are able to discriminate between displays in which objects are chasing, from displays in which objects move in no relation to each other (Rochat, Morgan & Carpenter, 1997). By the end of their first year, infants also detect a dif-ference in object-specific role within the chasing interaction – that is, when the chaser becomes the target of the chase (Rochat, Striano & Morgan, 2004). This indicates a growing understanding of the causal structure of social interactions. Unlike these studies, Study II examined visual attention within the interaction during observation of a dynamic event rather than as a global preference between motion displays, shedding light on how infants track individual objects during interactions.

Finally, Study III was designed to examine neural response toward ob-jects based on motion. More specifically, whether infants at 9-months cate-gorize objects differently based on whether they were previously observed

70

interacting with another object. Background literature is limited in this re-gard but one study seems to suggest that at 9-months infants allocate more attention, evident in the increased amplitude of the Nc component, in re-sponse to an object that moved inanimately (Kaduk et al., 2003). Authors reasoned that because animate motion is more easily detected, it does not require additional attentional resources. In parallel to this finding is research examining sensitivity to social content in young infants (Bakker et al., 2015; Bakker et al., 2014, 2016; Gredebäck et al., 2010; Melinder et al., 2015). This research suggests the presence of the P400 component, which has been argued to index pSTS activity (Gredebäck et al., 2010; Gredebäck and Daum, 2015), an area found to consistently respond to coordinated, inten-tional interactions such as chasing in adult observers (Castelli et al., 2000; Lee et al., 2012; Martin and Weinsberg, 2003; Shultz et al., 2004; 2005; Gao, Scholl, & McCarthy, 2012). Focusing on the attentional Nc component as well as the P400, Study III compared ERP activation in response to an object previously seen interacting with another (chaser) and an object that moved inanimately (Experiment 1). As a follow-up, we examined whether presence of an interaction, rather than animacy per se, would result in a dif-ferential ERPs. In this case, the interacting object (chaser) was compared to an animate object that did not interact with another (Experiment 2).

Summary of findings To reiterate specific findings, results from Study I show that adult observers readily direct their gaze toward two randomly moving identical shapes when their relative distance is smaller compared to a third object. Similar visual bias is evident in infants toward the end of their first year, but not earlier. For the 5-month-olds, looking toward objects that are approaching was not statistically above chance level. An important conclusion from Study I is that the presence of an interaction is not necessary for a visual bias toward proximal objects. Spatial proximity alone may elicit looking, possibly in anticipation of an interaction after the first year of life.

Given that spatial proximity biases looking in older infants when present-ed with randomly moving shapes, the follow-up inquiry aimed to consider gaze behavior during online observation of interactions. In so doing, we used one of the most ubiquitous, fitness-relevant in evolutionary sense inter-action that has been shown to be a powerful cue of animacy - chasing. Across the three experiments in Study II, findings suggest that indeed as infants observe motion, majority of looking is directed toward the interact-ing objects that are also proximally close, in line with findings from Study I. At the same time, results show differential looking within the interacting pair of objects based on which object’s motion is heat-seeking. In a chasing in-teraction, infants at 5- and 12-months preferentially attended to the chaser.

71

To a lesser degree, they attended to the target of the chase, and least to an object outside the interaction. When observing an interaction in which one object follows behind another, the preference remained for the agent behind but only when it also was the object that initiated the interaction through heat-seeking. This preference effectively disappeared with the elimination of initial heat-seeking, as two spatially close objects that moved synchronously were attended equally.

Findings from Studies I and II further motivated the research effort by asking in what way the preference for the heat-seeking agent, or the chaser, translates in measurements other than gaze. More specifically, the final study included in the present thesis was motivated by two outstanding hy-potheses. One hypothesis suggests that animated interactions such as chas-ing, and agents within them, are conceptualized as social entities. An alter-native hypothesis attributes the visual preference toward chasing interactions to low-level perceptual processing of individual motion parameters causing the attentional shift towards them. Utilizing a verified paradigm in extracting time-locked ERP responses following motion observation (Gredebäck et al., 2015; Kaduk et al., 2013), Study III examined 9-month-old infants’ response to the chasing agent. In two experiments, we have effectively compared the presence of animacy (Experiment 1) with the presence of an interaction (Ex-periment 2). That is, we examined differential ERP response to the chasing agent compared with a response to an agent that moved inanimately (Exper-iment 1) or a self-propelled agent that moved randomly (Experiment 2).

Results from Study III lend support to the first hypothesis in indicating no attentional differences between objects in the attentional Nc component, but corroborating with previous research examining perception of social infor-mation (Bakker et al., 2015; Gredebäck et al., 2015; Melinder et al., 2015; Schultz et al., 2005; Senju, Southgate, Snape, Leonard, & Csibra, 2011). Specifically, both comparisons indicated strong positive amplitude for the chasing agent in the low occipital and post temporal areas 400 ms post stim-ulus onset. Presence of the P400 component in infants has been linked to pSTS activity in adults in response to animate motion (Gao, Scholl, & McCarthy, 2012) and chasing in particular (Schultz et al., 2005).

Across the three studies, we find that already by 5-months of age infants’ vision is biased toward specific, meaningful social cues such as heat-seeking motion. Infants prioritize objects displaying those cues while watching a dynamic event unfold. At the same time, less obvious and less specific rela-tional properties of interactions, such as spatial proximity, do not bias visual attention in 5-month-olds. Importantly, visual preference toward heat-seeking objects is beyond low-level saliency, giving way to a more meaning-ful interpretation. This interpretation is evident by 9-months, but perhaps earlier, when infants begin to process the chasing object as a social agent. Finally, toward the end of the first year, priotization of objects that are heat-seeking remains, and even less obvious cues such as spatial proximity are

72

attended in manner similar to adults, highlighting the importance of ongoing visual experience.

Given current findings, it becomes increasingly important to ask where they fit theoretically. In the introduction, two approaches were discussed. A gross characterization of each approach is that one, a modular framework, proposes specific early mechanisms specialized for processing specific type of input. In terms of perceptual organization, it is consistent with continuity thesis, which proposes that infants’ perceptions are the same as adults’ ‘alt-hough infants’ perceptions often may be indistinct and indeterminate where those of adults are sharp and clear’ (Condry, Smith & Spelke, 2001, pg. 25). Support for this view comes from studies examining sensitivity to mechani-cal physicality in visually naïve human babies (Mascalzoni et al., 2013), as well as many studies providing evidence for automatic and universal impres-sion of animacy (Gao et al., 2010; Heider & Simmel, 1944).

In contrast, here and elsewhere (Cohen, Amsel, Redford, & Casasola, 2014; Leslie, 1996) modularity has been presented as an alternative to in-formation processing based on extensive and rapid learning. In terms of perceptual organization, these types of theories embrace the discontinuity thesis. Discontinuity thesis postulates fundamental changes to the way in-fants perceive the world across development with the emphasis on the gen-eral learning mechanisms fueled by increased visual and motor experience with the world. Support for information processing approach comes from studies indicating developmental change. These, as discussed at length in the introduction, are based on habituation and global preference paradigms ra-ther than measures of online motion observation.

Considering these two perspectives, current findings are difficult to place. On the one hand, present research shows that the ability to detect and track a heat-seeking object is evident already at 5-months and this seems to be an invariant developmental feature. On the other hand, present findings suggest development in the way infants attend to less obvious motion cues such as spatial proximity. Consequently, present findings reflect continuity aspects of perceptual organization in some regard and discontinuity in others.

In addition, a notable point concerns the nature of what is tested. By fo-cusing on animate motion and a chasing interaction in particular, the re-search program here outlined examines persistent task humans have faced with respect to animals across evolutionary history. Given the findings as well as the nature of the problem, I argue that the results are consistent with predictions derived from the evolutionary framework and represent a critical junction for a novel area of inquiry: that of a specialized perceptual learning mechanisms for animate motion.

73

Functional specialization In the introduction of On the Origin of Species by means of Natural Selec-tion (1859) Charles Darwin outlines four postulates on which he bases the rest of his famous text. These are simply, that (1) individuals within any particular species vary, (2) that some of these variations are passed on to their offspring, (3) and because not all offspring survive into adulthood (4) those that do are consequently endowed with variations that are most favor-able given the environment in which they live. The surviving offspring are simply naturally selected. One such important variation is the ability to de-tect other animates. Visual processes that help in identifying other animates have proven to be critical for survival and have been passed on to the subse-quent generations (Barrett et al., 2016).

Reflecting on Darwin’s initial postulates, natural selection may have se-lected those particular variations of the visual system, which promote mech-anisms that are essential to survival. These variations, in turn, have evolved based on the input the visual system has received over the course of evolu-tionary history forming fitness-relevant adaptations. These types of adapta-tions may have influenced the very cognitive architecture in distinct ways creating functionally specialized phenotypes or specialized learning process-es that ultimately promote survival and reproduction (Barrett, 2006; Barrett & Kurzban, 2006). Detection of a chaser from the rich array of visual infor-mation in the environment may have been one such important adaptation.

The function of specialized perceptual biases Certain aspects of the visual input are more salient than others and natural human response to these types of stimuli is to attend to them. In adults it is evident in a fast detection and monitoring of animates from complex visual input (New et al., 2007). In infants these perceptual biases are evident across a range of stimuli including attention to human speech sounds (Vouloumanos & Werker, 2007), configurations that resemble human faces (Macchi, Turati, & Simion, 2004), causal structures (Mascalzoni et al., 2013), and the detection of biological motion (Simion et al., 2008). In both, adults and infants, the overarching function of these attention biases is to promote rapid acquisition and retention of knowledge.

A stimulus that signifies presence of threat is especially salient and re-membered. For instance, 8-month-old and 14-month-old infants have been found to look faster at images of dangerous animates such as snakes (Lobue & Deloache, 2010; Hoehl, Hellmer, Johansson, & Gredebäck, submitted), and 5-month-olds attended to these types of animates for much longer than any other stimuli (Rakison & Derringer, 2008). Young children also quickly remember facts about dangerous animals, independent of whether this type

74

of animal is an actual, everyday threat. For instance, city-dwelling children from Los Angeles as well as Shuar children from the Amazon region of Ecuador were able to learn and retain information about dangerous animals that were only a real threat to the latter group (Barrett & Broesch, 2012). In addition, non-animate stimuli that may be life threatening, such as the inges-tion of unknown plants, are also regarded with caution by young infants as evidenced by their increase in social-referencing behaviors prior to interact-ing with them (Wertz & Wynn, 2014a, 2014b; Elsner & Wertz, submitted).

Differential attention toward stimuli that may be harmful suggests a sys-tem that is prepared for learning about these types of stimuli (Lobue, 2013; Lobue et al., 2017; Nairne, Pandeirada, Gregory, & Van Arsdall, 2009; Sherry, Jacobs, & Gaulin, 1992). Attending to and tracking heat-seeking object for instance, may promote learning about the entity, inform about the type of observed event, which in turn may form basis for categorization.

Functional specialization and modularity Because perceptual processes are triggered by specific type of visual input, it is said that they are modular as opposed to resulting from a single undiffer-entiated process. They are ‘modular’ in the sense that they are specialized, are apparent very early in development and are similar across individuals. In light of present findings, as soon as the ability to smoothly pursue a moving target is evident, infants pursue the chaser, and continue to do so despite increased visual and motor experience with the world.

An alternative to the traditional view of modularity is the view that modu-larity in the sense of specialty is grounded in the notion of functional spe-cialization (Barrett, 2006; Cosmides & Tooby, 1992; Pinker, 2005; Sperber, 2005), but one that does not claim its final state. Rather, functional speciali-zation promotes a system that is prepared to promote learning about ani-mates. A functional feature of modules in this sense is that within new con-texts requiring novel skills, these specialized processes ‘piggyback’ on older ones, a process referred to as bifurcation (Barrett, 2012) allowing for pheno-typically distinct modules to be spawned for specific, highly practiced or important skills. In the case of chasing geometrical shapes, it is possible that the developmental system recruits phylogenetically old modules specialized for the detection of predators or more specifically for processing motion information.

This pattern of specialization suggests that a final phenotype may include a combination of ancestral as well as newly derived features that interact in their contribution to the adaptation’s function (Barrett, 2012). We can ex-trapolate in suggesting that if environmental input were to undergo drastic deviation from the norm (for example, if infants were to be -unethically- prevented from seeing many types of causal events) then the resulting final phenotype of that individual would change, perhaps reflected by the way

75

they observe causal events. At the same time, certain ancestral aspects of the phenotype would reliably remain the same despite poor visual experience, such as sensitivity to animates that are heat-seeking because the function it serves for predator detection. And so, while some aspects might vary across individuals (e.g. heart’s propensity to be effected by individual’s poor diet), others will remain invariant (such as the heart’s function of pumping blood). And so, it is not that modules are identical across individuals but, given typ-ically normal environmental circumstances and input, these modules will develop in a way that will exhibit specialization to process certain kinds of information in the same way across individuals. One way to examine the specificity of these perceptual processes is by examining whether growing sophistication of one type of function (e.g. language or working memory) influences attentional biases to cues such as heat-seeking as evidenced dur-ing online motion observation – a goal for the future research effort.

Another way to reason about these mechanisms is to examine mecha-nisms behind visual attention, a key dependent variable, where ancestral and derived aspects of the phenotype are reflected in its cortical structures.

Visual attention The first two of the three empirical studies that form the basis of present thesis, rely on the measures of oculomotor behavior as a proxy for visual attention of which there are two distinctions (Egeth & Yantis, 1997; Theeuwes, Kramer, Hahn, & Irwin, 1998). One reflects top-down control mechanisms that are directed by volitional and attentional factors character-ized in slow cortical saccades. The second is a bottom-up or stimulus driven mechanism that is rapid, automatic and largely outside volitional control reflected in subcortical saccades (Johnson, 1989; 2001). The two processes are also the products of two distinct neural pathways. During cortical sac-cades, receptive cells in the parietal cortex respond to a combination of eye and head position as well as the retinal distance from the fovea to the target. During subcortical saccades, on the other hand, receptive cells commonly respond according to the retinal distance and the direction of the target from the fovea (Andersen & Zipster, 1988; Zipster & Andersen, 1988). In terms of ontology, it has been proposed that subcortical saccades are present at birth, while cortical saccades reflect increasing specialization of micro cir-cuitry in response to experience, ‘scaffolded’ by subcortical saccades (John-son, 1989).

Empirical evidence suggests that the developmental time course of the cortical saccades is relatively quick. For instance, one study (Gilmore & Johnson, 1995) found that 4-month-olds directed their gaze to the retinal location of the target produced by subcortical saccades, while already at 6-months their gaze was made to the target’s actual spatial location reflecting

76

the workings of cortical saccades. If we accept that before 6-months infant visual attention is largely stimulus-driven and gaze behavior is largely com-posed of subcortical saccades, then our findings on spatial proximity reliably coincide with previous research: at 5-months infants do not direct gaze to-ward objects that approach each other outside an interaction, but after 5-months they do. This supports the view that attention to subtle isolated cues such as relative spatial proximity may involve volitional control or top-down processes. As we propose, such subtle cues may be attended simply because change in the spatial structure usually, but not always, informs about a po-tential interaction.

At the same time, the attentional mechanisms in chaser preference demonstrated in Study II are more difficult to explain. On the one hand, 5-month-old chaser preference may again be result of stimulus-driven visual bias, but by 12-months it is a top-down process. Alternatively, initial bias toward the chaser might result from subcortical structures in both age groups but the continued monitoring of the interaction might involve more top-down processing, also reflected in the later categorization (Study III). In either case, it is possible that the very initial attention to the agent whose motion is distinctly heat-seeking reflects the workings of perceptual process-es involving subcortical neural pathways, while in the course of the stimulus presentation other inferential processes might be involved. One way to sepa-rate the two might be to examine an immediate reaction or first saccades toward the chaser during extremely short presentation time, another im-portant goal for future research.

Event related potentials Along with the measures of visual attention, Study III added analytical depth to the behavioral data by focusing on the event related potential (ERP) com-ponents. While indeed complementing behavior, it is important to acknowledge that measuring ERPs nevertheless lacks in precision and di-rectness akin to measuring neural circuitry, such as when using NIRS or the fMRI. It is evident that the developmental literature has clearly benefited in the use of ERPs. However, conclusions drawn from these types of studies rely primarily on functional reactivity to certain types of stimuli, while their validity relies on the replicability of the findings to a much greater degree than do other methodologies.

In particular, when considering developmental ERPs, two layers of paral-lelism must be drawn. The first requires finding the adult analogue to the observed infant ERP. The second requires finding neural circuitry responsi-ble for the adult ERPs. Only then it is possible to speculate on the neural circuitry as reflected in the infant ERPs. Given these parallels, it is evident that caution must always be exercised when interpreting the neural genera-

77

tors of the scalp-recorded ERPs, especially in young populations. But be-cause critical questions in cognitive neuroscience always ultimately concern brain circuitry and structures that mediate the differential responses, these types of parallels permeate current literature, with Study III among them.

While fully acknowledging interpretive limitations, Study III demonstrat-ed the presence of a N290/P400 complex in 9-month-old infants in response to the chasing object. As previously discussed at great length here and in the original manuscript, this ERP component has been implicated in processing of various types of social information such as goal directed actions (Bakker et al., 2015), pointing (Gredebäck, Melinder, & Daum, 2010; Melinder et al., 2015), biological motion (Reid, Hoehl, & Striano, 2006) and gaze direction (Senju, Johnson, & Csibra, 2006). This component has been found to be an analogue for the adult N170, suggested to reflect increased vigilance to so-cial information (Whalen et al., 2004) and face perception (Leppänen, Moulson, Vogel-Farley, & Nelson, 2007). Given physiological differences in the head, skull, and the underlying brain tissue, infant and adult ERP waveforms differ in timing, amplitude, directionality and scalp-topography (DeBoer, Scott, & Nelson, 2005). Importantly, however, presence of these comparisons relies on similar response properties reflecting functionally equivalent processes (Halit, De Haan, & Johnson, 2003).

Drawing on the second parallelism, source localization techniques have found evidence linking adult N170 with the STS activity (Allison, Puce, Spencer, McCarthy, & Belger, 1999; Dalrymple et al., 2011; Itier & Taylor, 2004), a structure linked to social information processing, considered to be the hub of the social brain network (Brothers, 2002; Dasgupta, Tyler, Wicks, Srinivasan, & Grossman, 2016).

Therefore, the ERPs observed in infants may be linked to the STS activity and its posterior area in particular, suggesting that response to the chasing object, as opposed to inanimate or random object, is interpreted in terms of goals akin to human actions, ultimately categorized as social. In sum, while motion interpretation may involve primary visual areas, it possible that their categorization may involve evolutionarily newer structures involved in pro-cessing various types of social cognitive cues.

Limitations As with any studies, there is always room for improvement. In our case, all three studies could benefit from an increased number of participants. This is especially true for Study III, which, due to practical difficulties measuring ERPs in infants at 9 months, resulted in relatively high drop out rate (alt-hough still within acceptable limits).

Other limitations include short stimuli presentations in Study I. An out-standing hypothesis discussed earlier may be that the 5-month-olds’ detec-

78

tion of the reduction in spatial proximity may result from limited ability to process visual input. Longer presentation of the stimuli may have resulted in significant results similar to that of older infants and adults. In Study II on the other hand, it may have been useful to examine not only the ability to track and fixate on the objects but also latency toward the chaser. Current design did not allow us to do this, but future research should focus on de-signing stimuli in such a way as to allow the measurement of first fixations toward specific objects. Finally for Studies I and II it would be interesting to trace the developmental trajectory in more detail by examining visual biases in groups at different ages or in a longitudinal design.

Still other limitations concern Study III. One has to do with the methods and the second with interpretation. First, the paradigm used in Study III was a valid method for the measurement of ERPs following motion observation. However, because we measured ERPs during still image, we effectively measured time-locked responses based on previously seen motion infor-mation that categorized objects into ‘social/animate’ or not. While this par-ticular method has been used in our lab with similar-aged population previ-ously, the ultimate findings indicate not motion perception per se – whether chasing interaction is categorized as social – but how infants categorize still-images of the objects based on previously seen motion. Another way of ex-amining the effects of motion would be to perform frequency analysis dur-ing motion observation, a goal for our future studies.

The second limitation of Study III has to do with the interpretation of the findings. Specifically, we argue that the N290/P400 complex demonstrated in our study is the result of activity in the pSTS. The specific limitation here is that activity in the pSTS is not isolated to animate motion and in fact has been implicated across a wide and varied range of tasks from biological motion perception (Allison, Puce, & McCarthy, 2000; Puce & Perrett, 2003) to face perception (Haxby, Hoffman, & Gobbini, 2000), gaze recognition (Itier & Batty, 2009) and interpretation of human actions (Bakker et al., 2015; Pelphrey, Morris, & McCarthy, 2004). Further, it has been implicated in nonvisual tasks, which require higher level mentalizing, such as theory of mind attributions (Gallagher & Frith, 2003; Saxe, 2006; Zilbovicius et al., 2006), speech perception (Price, 2012) and audiovisual integration (Amedi, Von Kriegstein, Van Atteveldt, Beauchamp, & Naumer, 2005; Beauchamp, 2005; Calvert, 2001). Rightfully so, it has been dubbed ‘the chameleon of the human brain’ (Hein & Knight, 2008, pg. 2125). While true, a recent study (Dasgupta, Tyler, Wicks, Srinivasan, & Grossman, 2016) has shown that across a variety of social perception tasks, pSTS is the central pro-cessing area for social information but the unique aspect of each task is evi-dent in the neural connectivity of the pSTS with other areas of the brain. The most informative therefore would have been examining the perception of chasing in infants using fMRI and NIRS. The very lack of this type of in-formation is a limitation not only in Study III but also in developmental re-search in general.

79

Future directions The focus of the present thesis centers on the ability to on the one hand rec-oncile evidence for developmental change with the notion of inborn sensitiv-ity and on the other the automaticity of the percept with voluntary inference. The point of reconciliation is the functional specialization approach, which suggests specialized mechanisms that can undergo developmental change – an integrative (and somewhat diplomatic) approach presented as a third op-tion at the end of the introduction. When it comes to chasing perception specifically and the perception of animacy more generally, much is yet to be explored in order to make clear the theoretical picture. For that purpose, there are several strategies, some more feasible than others, that reflect both the limitations currently present and ways in which to address them.

The first strategy is to study newborn infants. Much like with the percep-tion of physical causality, one way of testing modularity or continuity thesis is to demonstrate it at the very beginning of development. One important problem faced with research on newborn infants, however, is in the distinc-tion between competence and performance (Condry, Smith, & Spelke, 2001). While it is possible that mechanisms are present and functional, their performance is blocked by limits on infants’ sensitivity to the information on which they operate. For the purposes of chasing perception, it would be nec-essary to examine whether preference for the chaser is present at birth but it would not be possible to do with the current stimuli because tracking the chaser requires to smoothly pursue a target in motion. However, comprehen-sive research done by von Hofsten and Rosander (1996, 1997) has shown this is not possible with newborn infants. In their research von Hofsten and Rosander (1996;1997) found that the development of gaze stability and smooth pursuit is not evident until about 2 months. Before 5 months, track-ing moving objects is marred by substantial lag, which later disappeared. And so, given the stimuli, testing infants younger than 5 months in Studies I and II might have been problematic and potentially unreliable. It may be possible to bypass this problem by measuring first saccades toward the chas-er, rather than smooth pursuit, a possibility for the future research effort.

The second strategy is to examine cross-cultural variations in chaser de-tection. Focusing on individuals in remote parts of the world such as indige-nous people of Papua New Guinea or the Amazon, whose main source of food still includes daily hunting expeditions reminiscent of our evolutionary ancestors’. One argument here would be that the detection of animate mo-tion, and specifically heat-seeking, is the same despite the continued use of these abilities for survival in one peoples and not so much in the other. Al-ternatively, the daily experience of utilizing these skills would result in a different preferential pattern toward specific motion cues (Brislin, 1983; Kagitcibasi & Berry, 1989).

80

The third strategy, as discussed previously, is to focus on cognitive neu-roscience in infants. While in recent years there have been a growing num-ber of studies examining brain structures and processes involved in animacy perception, this research effort has been limited to adult population largely due to practical demands involved in infant studies. Years of adult research have indicated specific areas of the brain responsible for processing specific type of information (Changeux, 1985; Edelman, 1987). In other words, there are specialized cell clusters that share common response properties such as to moving visual stimuli, resulting in cortical representations. However, it cannot be stressed enough how important it is to study the developing brain. For our particular purposes, the pertinent question is whether neuroanatomi-cal structures responsible for the perception of animacy are present early in the developing brain or whether they involve maturation of the pathways found in adult studies. Given the nature of the stimuli, it is possible that are-as of the visual cortex are prespecified at birth to particular type of infor-mation (Schlaggar & O’Leary, 1991) even if, in general, cortical structures exhibit high degree of plasticity (Johnson, Oliver, & Shrager, 1998). How-ever, this information is not available for the current thesis.

Finally, present research effort would benefit in examining clinical popu-lation. In more general terms, a clinical comparison group would help in identifying the kinds of stimuli that are social, help in examining alternative strategies used to process social information, and to determine which strate-gies are developmentally related and which are independent of development (Rutherford & Kuhlmeier, 2013). Specifically, individuals with autism spec-trum disorders who have been shown to have a deficit in recognizing causal-ity in launching events (Ray & Schlottmann, 2007) and recognizing animacy in artificial stimuli in motion (Congiu, Schlottmann, & Ray, 2010). For in-stance, Klin (2000) found that in the classic Heider and Simmel task, indi-viduals with autism provided significantly less social elements, used signifi-cantly less pertinent theory of mind terms pertaining to psychologically based personality features of the shapes motion. Evidence from brain imag-ing research also suggests deficits in activity of specific brain structures such as the bilateral STS at the temporo-parietal junction, the basal temporal area, the fusiform gyrus, the right temporal pole and the medial prefrontal cortex involved during the performance of social cognitive tasks (Rutherford, Pennington, & Rogers, 2006). At the same time, when presented with chas-ing motion or other types of goal-directed motion, there seemed to have been no differences between the ratings of intentionality between typical controls and the ASD group (Castelli, Frith, Happe, & Frith, 2002; Abell, Happe, & Frith, 2000). Thus, question remains whether the ASD group’s deficit has to do solely with higher-level failure to mentalize during more complicated displays or failure in proper observation and attention distribution during simplified goal-directed motion displays.

81

Final conclusions To conclude, the work here presented has fulfilled the purpose for which it was set out to do, in that it augmented current knowledge about the percep-tion of animacy across development. Beyond that, present work has indicat-ed that phenomenological nature of animacy is a result of visual biases to-ward specific motion cues and objects that display them. This sensitivity, I argue, is a mechanism on which further event categorization is based. Final-ly I propose that visual bias and subsequent categorization here presented reflect the workings of a phylogenetically old attention mechanism that has been selected for this specific function.

82

Svensk sammanfattning

Sedan psykologerna Fritz Heider och Marienne Simmel gjorde sina banbry-tande studier på fyrtiotalet har forskare studerat hur och när människor upp-fattar något som en levande varelse baserat på dess rörelser. I den ursprung-liga studien av Heider och Simmel (1944) fick vuxna deltagare se en hand-gjord animation där två trianglar och en cirkel rörde sig runt en rektangel. Trots att instruktionen var att återberätta vad de såg tillskrev de allra flesta deltagarna de geometriska formerna en mänsklig komplexitet: formerna var levande, deras rörelser var meningsfulla och deras mål var komplexa. Lock-elsen för psykologer och andra forskare både förr och idag är att förstå detta universella intryck av liv och intentioner, och hur så enkla synintryck kan leda till så rika och komplexa tolkningar.

För att studera dessa fenomen är det viktigt att forska på spädbarn, ef-tersom de kan visa hur hjärnan utvecklas och ger upphov till vuxna funkt-ioner. Det är också viktigt att känna till den typiska utvecklingen för att kunna identifiera eventuella avvikelser från denna. Idag vet vi väldigt lite om hur spädbarn bearbetar och tolkar den här typen av enkla synintryck, trots många försök. Åtskilliga habituerings-experiment har visat att spädbarn ser skillnad på animationer med och utan kausala relationer mellan de geo-metriska formerna, men ingen har ännu kunnat påvisa att spädbarn faktiskt uppfattar kausala interaktioner som sociala.

Det övergripande målet för denna avhandling är att fördjupa kunskapen om de mekanismer som leder till att spädbarn tolkar en geometrisk form som levande. Vi gör detta på tre olika sätt, i tre olika studier. I studie 1 undersö-ker vi blickbeteendet hos vuxna och spädbarn med hjälp av ögonrörelsemät-ning, som ett mått på visuell uppmärksamhet, medan de tittar på geometriska figurer som rör sig utan inbördes kausala relationer. I studie 2, som också använder ögonrörelsemätningar, undersöker vi hur spädbarn fördelar sin visuella uppmärksamhet på olika objekt när de ser kausala interaktioner mellan dessa, med ett fokus på en grundläggande och tänkbar evolutionärt betingad relation – jagande. I studie 3 svarar vi på frågan om spädbarn upp-fattar ett socialt innehåll i den här typen av animeringar genom att använda EEG och mäta aktiveringen av områden i hjärnan som bearbetar sociala intryck.

Studie 1 visar att vuxna och äldre spädbarn (12 månader gamla), men inte 5-månadersbarn, uppmärksammar när två geometriska figurer kommer i närheten av varandra, även om de inte interagerar. Studie 2 visar att när 5-

83

månadersbarn ser kausala interaktioner så riktas deras uppmärksamhet till vissa särdrag hos objekten som beror på interaktionstypen, och att de speci-fikt tittar på objekt som rör sig bakom ett annat. Studie 3 visar för första gången att spädbarn tolkar en jakt-animering som en social interaktion, och att sociala områden i hjärnan aktiveras starkare när de ser en jägare än en icke-jägare, vilket antyder en viktig länk mellan rörelseperception och social kognition.

Det gemensamma bidraget från dessa studier är också trefaldigt. För det första visar vi att det visuella systemet hos människor tidigt och effektivt uppfattar tecken som predicerar interaktion. För det andra visar vi att denna förmåga inte bara beror på enkla rörelseintryck, utan att de bearbetas som sociala interaktioner mellan objekten. För det tredje visar vi att spädbarn aktivt fördelar sin uppmärksamhet till specifika objekt när de ser enkla geo-metriska animationer. Tillsammans skapar detta en bild där mänskliga pro-cesser för att kategorisera sociala agenter är tidig i utvecklingen, snabb och precis, vilket leder till frågan om social kategorisering verkligen kan anses vara en ”högre” kognitiv funktion eller om den är en ”lägre” automatisk och domänspecifik process liknande rörelse- eller form-seende.

84

Streszczenie w języku polskim

Od czasu twórczość psychologów Fritza Heidera i Marienne Simmel (1944) spostrzegania i odbioru reprezentacji bodźców ożywionych, intrygowało badaczy.

W swoich nowatorskich badaniach Heider i Simmel odkryli, ze dorośli, którzy obserwują ruch narysowanych ręcznie figur geometrycznych (dwa trójkąty i okrąg poruszający się wokół kwadratu) prawie jednoznacznie in-terpretują ich zmianę położenia przez pryzmat złożoności ludzkich zachow-an. Kształty były opisywane jako żywe obiekty, których ruchy są celowe i wielowymiarowe. Psychologowie i naukowcy zainteresowali się uniwersal-nością i bogactwem zależności widzenia rzeczy jako ożywionych i w ich „zachowaniu” celowych poprzez ukazanie ich w ruchu, zwłaszcza biorąc pod uwagę niewielka ilość dostarczonych informacji.

Badania obiektów ożywionych u niemowląt są kluczowe, ponieważ mogą ujawnić, w jaki sposób te procesy rozwijają się w złożone funkcje które możemy zauważyć u dorosłych. Dodatkowo, wydaje się ze poznanie jak wygląda rozwój spostrzegania takich zależności u normalnie rozwijającej się populacji jest niezbędne w celu zidentyfikowania odstępstw od normy. Do tej pory bardzo niewiele wiemy o tym jak niemowlęta przetwarzają tego typu informacje. Wiele badan wykazywało wrażliwość na związek przyczynowo-skutkowy pomiędzy kształtami geometrycznymi, ale aż do teraz nie wykazano, czy niemowlęta faktycznie postrzegają ruch obiektów nieożywionych jako interakcje społeczne.

Głównym celem obecnej pracy doktorskiej jest poszerzenie aktualnej wiedzy na temat mechanizmów spostrzegania bodźców ożywionych w okresie niemowlęctwa. Przeprowadziliśmy trzy badania. Najpierw zbada-liśmy jak dorośli i niemowlęta odbierają i przetwarzaja kształty geome-tryczne, które są wprawione w ruch (Badanie I, przy użyciu eye-tracking). Po drugie, analizowany był rozkład uwagi u niemowląt w wieku 5- i 12-miesiecy podczas obserwacji interakcji przyczynowej bez kontaktu fizycznego – tzw. gonitwie (Badanie II, używając eye-tracking). Po trzecie, zbadaliśmy, czy na poziomie neuronalnym będziemy mogli zauważyć ró-żnice w czasie obserwacji bodźców ożywionych w porównaniu z bodźcami nieożywionymi. (Badanie III, przy użyciu EEG).

Badanie I wykazało, że dorośli i starsze niemowlęta (12-miesieczne), ale nie 5-miesięczne, są wrażliwi na przypadki przestrzennej bliskości obiektów. Badanie II wykazało, że obserwując bezkontaktową interakcję

85

przyczynową, niemowlęta w wieku 5 miesięcy są szczególnie wrażliwe na właściwości ruchu pojedynczego bodźca w obrębie interakcji i jest to przede wszystkim wrażliwość na bodziec, który goni za innymi (chaser) bodźcami (target). Banie III było pierwszym, które wykazało, że niemowlęta przet-warzają obserwowana interakcje jako społeczne, a postrzeganie agresora (chaser) wywołuje silniejsze pobudzenie na poziomie neuronalnym, wykazując tym związek pomiędzy percepcją ruchu a poznaniem społec-znym.

Powyższe badania przyczyniają się do pogłębienia wiedzy. Najpierw pokazujemy, że ludzki system wizualny jest wrażliwy na sygnały, które skutecznie przewidują interakcję. Po drugie, pokazujemy, że wrażliwość na wspomniane bodźce ruchowe nie wynika z podstawowych bodźców per-cepcyjnych/wizualnych, ale raczej jest wynikiem przetwarzania ich jako elementy interakcji społecznej. Wreszcie wykazujemy, że postrzeganie agentów społecznych jest szybkie i bezpośrednie, co stawia pod znakiem zapytania czy kategoryzacja społeczna powinna być uznawana jako funkcje wyższe lub jako niższy proces percepcyjny.

86

Acknowledgements

I begin this all-time favorite section of the thesis by thanking my research supervisors Pär Nystrom and Gustaf Gredebäck for the crucial role they played in my graduate training. Sincere thanks go to Pär for his unwavering research enthusiasm, innumerable academic insights and incredible pro-gramming ability. But, perhaps most importantly, I thank you for your pro-longed personal support marked by consistent willingness to give advice on matters of intellect and of life. I thank you for your guidance and support.

Equally sincere thanks to Gustaf for his faith and incredible foresight, in answering that one strange e-mail from some student from Boston almost 6 years ago. Since that time, I thank you for your invaluable theoretical and practical input, verbal encouragement and genuine expressions of joy and pride after personal achievements in academia and life.

I’d like to also thank Claes von Hofsten and Kirstin Rosender. Although I was not directly supervised by you, I have undoubtedly benefited from your incredible academic work – you both have created a legacy that is our lab and through your hard work you have crafted a platform on which I, along with countless researchers, have been able to learn and perfect our skill. Thank you both.

The work here presented would not have been in the state it is, if it wasn’t for the invaluable input I had received from Leo Poom, Karin Brocki, and Benjamin Kenward during my halftime and final seminars. I thank you all for your insight, pointed questions and constructive criticism. In much the same way, I’d like to thank Gunilla Bohlin, Ann-Margret Rydell and Carin Marciszko for your numerous inputs and comments on earlier versions of the papers that make up this thesis. Your pertinent questions and comments during manuscript meetings and suggestions on ways to improve have been imperative and appreciated.

Sincere thanks go to Terje Falck-Ytter, Christine Fawcett, Marcus Lind-skog and Janny Stapel. Terje, thank you for your collaboration, your incred-ible research insight, and willingness to give honest and thoughtful advice. Christine, you’re an incredible researcher and person. I am so grateful to have learnt from your incredible statistical and writing skills. Marcus, thank you for always being so welcome in discussing chasing of all things and to always pose thoughtful comments and questions. Janny, you are an incredi-ble researcher and a friend. Thank you your work as a collaborator, your advice and support.

87

Next, I would like to thank my dear colleagues and friends Marta Bakker and Claudia Elsner. Marta, there are really no words to describe how grate-ful I am to have you in my life. It continues to amazes me that although you and I have had so little overlap working together we have done so much and have become so close! Not only did we author a paper together but you have advised me in innumerable ways with other aspects of academia, both prac-tical and theoretical. Most importantly, however, you have helped me in the transition of becoming an ‘academic mom’ and from the bottom of my heart I appreciate all the advice in this aspect of my life. Thank you for every-thing. Claudia, from the moment you and I met I knew there’s something very familiar in you - your energy, your ‘get-it-done’ attitude, your ability to make impossible possible. You are wise beyond your years and I cannot thank you enough for sharing your wisdom with me in immeasurable ways. I thank you for your profound academic and life insights, too pervasive to enumerate, that make it easy to forget your own youthful age.

I’d like to thank my colleague, officemate and friend Benjamin Koch. Thank you for wonderful conversations, full of laughter, tears, and jokes. Thank you for your consistent willingness to help – not just when it is con-venient but always. Thank you for your understanding, care and constant supply of chocolate for your moody officemate.

Much thanks to Dorota Green, Therese Ljunghammar and Hanna Skager-ström. Dorotka, you are such a wonderful person and such an incredible researcher. You have an amazing sense of humor, which often centers on your ability to not take yourself too seriously. Your calm demeanor, level-headedness and strong sense of self is truly inspiring and I thank you for all the advice, help and laughter. Therese, the most wonderful thing about you is your ability to not stress the small things and to be very practical when the need arises. I very much appreciate that about you and thank you for all the help. Hanna, your sense of humor is unprecedented. I thank you for all your help, creativity, laughter, support and always sharing in our joys. You are officially an honorary Uppsala Babylab member!

If there is one thing I have learned as a doctoral student, it is that your fel-low students, people with whom you share the ‘daily struggle’ are some-times greater sources of knowledge than any journal article can provide. It is your fellow students who help you solve numerous technical problems, who direct you to an important article you of course missed, who ask the tough questions and with whom you can always share joys and frustrations of re-search. In our lab, while big and a constantly growing, I was lucky to find those very valuable sources of knowledge and support in so many people. Collectively, I’d like to thank my fellow doktorander Emilia Thorup Viktor Granvald, Tommie Forslund, Matilda Frick, Johan Lundin Kleberg, Joshua Juvrud and Elisabeth Nilsson Jobs. I’d like to thank Kahl Hellmer. Kahl, thank you for being such a positive and happy person. I wholeheartedly ap-preciate for your calm attitude, positive demeanor, amazingly funny stories

88

about your kids and your help in creating many, many videos involving clementines and hammers.

I would also like to extend thanks to those doktorander who have already completed their ‘journey’ of obtaining their Ph.D.’s but who have been part of mine. Thank you Maria Johansson, Eva Stening and Andrea Handl. You all have been such wonderful co-workers and I am grateful for such positive learning and working environment your presence provided. Janna Gottwald, thank you for accompanying me on this incredible journey, thank you for the laughter, commiserating about research and ‘congratulations’ to us both for not giving into our pessimism. Clara Schmitow, thank you very much for your unending generosity, your kindness, your creativity and for inspiring me to make – and wear!- a ‘medieval’ wedding dress.

I thank Goncalo Barradas and Laura Sakka, my friends, office neighbors and amazing researchers. It has been truly a pleasure to have you as friends, to share in the joys of teaching about gender, sex and ethnicity and for many discussions about travel, research and life in general.

In my time at Uppsala Child and Babylab, I had the pleasure of meeting and working with many visiting students and researchers from who I was able to learn and share knowledge. I thank, Bahar Tuncgenc, Aurora de Bartoli Vizioli and Victoria Wesevich for the wonderful times, laughter, inspiring research talks and hallway chair racing.

I would also like to express my gratitude to Annika Landgren, Lars-Gunnar Karlsson, Cecilia Sundberg for all your practical help navigating the many twists and turns of the Swedish system.

Many thanks to Costis Chatzidakis, the artist and creator of the front cov-er, Pär Nystrom for help in the Swedish translation as well as Marta Bak-ker for help with the Polish translation of the text.

Finally, I’d like to thank my mentor and colleague from Boston Universi-ty. To Dr. Patricia Ganea whose intellect and academic prowess encouraged me to strive in emanating the same in myself. Dr. Caitlin Canfield whose intellect, maturity marked by kindness surpass her young age in many re-gards.

I end by extending my thanks to my family. My mom for numerous heat-ed discussions on evolution, help during the writing process, as well as, for instilling in me principles of hard work and humility. I thank my brother, Milosz and his wife Angel, for keeping me focused and motivated long be-fore and during graduate school. Thank you also to my other brother, Kevin, for his quiet support and willingness to run in endless circles for the sake of his sister and research. My husband, Joshua, for unending support, love, strangely-appropriate discussions on chasing with equally as strange notions of the possibility to apply my research to help people (pff, engineers!). Last but not least, I thank my incredibly smart and beautiful daughter, Izabella, whose very existence grounded me in unprecedented ways and helped me look beyond the computer screen. You are the light of my life.

89

References

Allison, T., Puce, A., & McCarthy, G. (2000). Social perception from visual cues: Role of the STS region. Trends in Cognitive Sciences, 4(7), 267–278. http://doi.org/10.1016/S1364-6613(00)01501-1

Allison, T., Puce, A., Spencer, D. D., McCarthy, G., & Belger, A. (1999). Electrophysiological studies of human face perception. I: Potential generated in occiptotemporal cortex by face and non-face stimuli. Cerebral Cortex, 9(5), 415–430. http://doi.org/10.1093/cercor/9.5.415

Amedi, A., Von Kriegstein, K., Van Atteveldt, N. M., Beauchamp, M. S., & Naumer, M. J. (2005). Functional imaging of human crossmodal identification and object recognition. Experimental Brain Research, 166(3–4), 559–571. http://doi.org/10.1007/s00221-005-2396-5

Baillargeon, R., Wu, D., Yuan, S., Li, J., & Luo, Y. (2012). Young infants’ expectations about self-propelled objects. The Origins of Object Knowledge, 1–84. http://doi.org/10.1093/acprof:oso/9780199216895.003.0012

Bakker, M., Kaduk, K., Elsner, C., Juvrud, J., & Gredebäck, G. (2015). The neural basis of non-verbal communication-enhanced processing of perceived give-me gestures in 9-month-old girls. Frontiers in Psychology, 6(FEB), 1–6. http://doi.org/10.3389/fpsyg.2015.00059

Banich, M. T., Milham, M. P., Atchley, R. a, Cohen, N. J., Webb, A., Wszalek, T., … Brown, C. (2000). Prefrontal regions play a predominant role in imposing an attentional “set”: evidence from fMRI, journal = {Brain Res Cogn Brain Res}, uuid = {0F27DB0D-E008-4ED8-861C-D644A90A0FB0}, volume = {10}, number = {1-2}, pages = {1-9}, month = {9}, year =, 10, 1–9.

Bardi, L., Regolin, L., & Simion, F. (2011). Biological motion preference in humans at birth: Role of dynamic and configural properties. Developmental Science, 14(2), 353–359. http://doi.org/10.1111/j.1467-7687.2010.00985.x

Barrett, C. (2006). Modularity and design reincarnation. The Innate Mind: Culture and Cognition, Ed. P. Carruthers, S. Laurence & S. Stich, 199–217. Retrieved from http://www.sscnet.ucla.edu/anthro/faculty/barrett/barrett 2006 modularity.pdf

Barrett, C. (2015). The Shape of Thought: How Mental Adaptations Evolve. Oxford Scholarship Online.

Barrett, C., & Kurzban, R. (2006). Modularity in cognition: framing the debate. Psychological Review, 113(3), 628. http://doi.org/10.1037/0033-295X.113.3.628

Barrett, C., Todd, P. M., Miller, G. F., & Blythe, P. W. (2005). Accurate judgments of intention from motion cues alone: A cross-cultural study. Evolution and Human Behavior, 26(4), 313–331. http://doi.org/10.1016/j.evolhumbehav.2004.08.015

90

Barrett, H. C. (2012). A hierarchical model of the evolution of human brain specializations. Proceedings of the National Academy of Sciences of the United States of America, 2012, 1–8. http://doi.org/10.1073/pnas.1201898109

Barrett, H. C., & Broesch, J. (2012). Prepared social learning about dangerous animals in children. Evolution and Human Behavior, 33(5), 499–508. http://doi.org/10.1016/j.evolhumbehav.2012.01.003

Beasley, N. A. (1968). The extent of individual differences in the perception of causality. Canadian Journal of Psychology, 22(5), 399–407. http://doi.org/10.1037/h0082779

Beauchamp, M. S. (2005). See me, hear me, touch me: Multisensory integration in lateral occipital-temporal cortex. Current Opinion in Neurobiology, 15(2), 145–153. http://doi.org/10.1016/j.conb.2005.03.011

Berry, D. S., Misovich, S. J., Kean, K. J., & Baron, R. M. (1992). Effects of disruption of structure and motion on perceptions of social causality. Personality and Social Bulletin, 18(2), 237–244.

Blakemore, S.-J. (2008). The social brain in adolescence. Nature Reviews. Neuroscience, 9(4), 267–277. http://doi.org/10.1038/nrn2353

Blakemore, S.-J., Boyer, P., & Meltzoff, A. (2003). The Detection of Contingency and Animacy from Simple Animations in the Human Brain, 837–844. http://doi.org/10.1093/cercor/13.8.837

Blakemore, S. J., Fonlupt, P., Pachot-Clouard, M., Darmon, C., Boyer, P., Meltzoff, a N., … Decety, J. (2001). How the brain perceives causality: an event-related fMRI study. Neuroreport, 12(17), 3741–3746. http://doi.org/10.1097/00001756-200112040-00027

Blythe, P. W., Todd, P. M., & Miller, G. F. (1999). How Motion Reveals Intention: Categorizing Social Interactions. Simple Heuristics That Make Us Smart.

Boyle, D. G. (1960). A contribution to the study of phenomenal causation. Quarterly Journal of Experimental Psychology, 12(3), 171–179. http://doi.org/10.1080/17470216008416721

Brislin, R. (1983). Cross-cultural research methods in psychology. Annual Review of Psychology, 34, 363–400.

Buren, B. Van, Uddenberg, S., & Scholl, B. J. (2015). The automaticity of perceiving animacy : Goal-directed motion in simple shapes influences visuomotor behavior even when task-irrelevant. http://doi.org/10.3758/s13423-015-0966-5

Burgess, N., Maguire, E. A., Spiers, H. J., & O’Keefe, J. (2001). A temporoparietal and prefrontal network for retrieving the spatial context of lifelike events. NeuroImage, 14(2), 439–53. http://doi.org/10.1006/nimg.2001.0806

Calvert, G. a, & Calvert, G. a. (2001). Crossmodal processing in the human brain: insights from functional neuroimaging studies. Cerebral Cortex (New York, N.Y. : 1991), 11, 1110–23. http://doi.org/10.1093/cercor/11.12.1110

Caramazza, A., & Shelton, J. R. (1998). Domain-specific knowledge systems in the brain the animate-inanimate distinction. Journal of Cognitive Neuroscience, 10(1), 1–34. http://doi.org/10.1162/089892998563752

Carey, S., & Spelke, E. S. (1996). Science and core knowledge. Philosophy of Science, 63(4), 515–533. http://doi.org/10.2307/188065

91

Cassia, V. M., Turati, C., & Simion, F. (2004). Can a nonspecific bias toward top-heavy patterns explain newborns’ Face preference? Psychological Science, 15(6), 379–383. http://doi.org/10.1111/j.0956-7976.2004.00688.x

Castelli, F., Happé, F., Frith, U., & Frith, C. (2000). Movement and mind: a functional imaging study of perception and interpretation of complex intentional movement patterns. NeuroImage, 12(3), 314–25. http://doi.org/10.1006/nimg.2000.0612

Childers, J. B., & Echols, C. H. (2004). 2.5-year-old children use animacy and syntax to learn a new noun. Infancy, 5(1), 109–125. http://doi.org/10.1207/s15327078in0501_5

Choi, H., & Scholl, B. J. (2004). Effects of grouping and attention on the perception of causality. Perception & Psychophysics, 66(6), 926–942. http://doi.org/10.3758/BF03194985

Cohen, L. B., & Amsel, G. (1998). Precursors to infants’ perception of the causality of a simple event, 21(4), 1–19. http://doi.org/10.1016/S0163-6383(98)90040-6

Cohen, L. B., Amsel, G., Redford, M. A., & Casasola, M. (2014). The Development of Infant Causal Perception. Igarss 2014, (1), 1–5. http://doi.org/10.1007/s13398-014-0173-7.2

Cohen, L. B., & Oakes, L. M. (1993). How infants perceive a simple causal event. Developmental Psychology, 29(3), 421–433. http://doi.org/10.1037/0012-1649.29.3.421

Condry, K., Smith, W. C., & Spelke, E. S. (2001). The Development of perceptual organization. In F. Lacerda, C. von Hofsten, & M. Heimann (Eds.), Emerging cognitive abilities in early infancy (pp. 1–28). Mahwah, New Jersey: Lawrence Erlbaum Associates.

Cosmides, L., & Tooby, J. (1992). Cognitive adaptations for social exchange. The Adapted Mind: Evolutionary Psychology and the Generation of Culture, 163–228. http://doi.org/10.1098/rstb.2006.1991

Courchesne, E., Ganz, L., & Norcia, A. M. (1981). Event-Related Brain Potentials to Human Faces in Infants, 52(3), 804–811.

Csibra, G. (2008). Goal attribution to inanimate agents by 6.5-month-old infants. Cognition, 107(2), 705–17. http://doi.org/10.1016/j.cognition.2007.08.001

Csibra, G., Gergely, G., Bíró, S., Koós, O., & Brockbank, M. (1999). Goal attribution without agency cues: the perception of “pure reason” in infancy. Cognition, 72(3), 237–67. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10519924

Cutting, J. E., Proffitt, D. R., & Kozlowski, L. T. (1978). A biomechanical invariant for gait perception. Journal of Experimental Psychology. Human Perception and Performance, 4(3), 357–72. http://doi.org/10.1037/0096-1523.4.3.357

Dalrymple, K. A., Oruc, I., Duchaine, B., Pancaroglu, R., Fox, C. J., Iaria, G., … Barton, J. J. S. (2011). The anatomic basis of the right face-selective N170 IN acquired prosopagnosia: A combined ERP/fMRI study. Neuropsychologia, 49(9), 2553–2563. http://doi.org/10.1016/j.neuropsychologia.2011.05.003

Dasgupta, S., Tyler, S., Wicks, J., Srinivasan, R., & Grossman, E. D. (2016). Network connectivity of the right STS in three social perception localizers. Journal of Cognitive Neuroscience, 29(2), 221–234. http://doi.org/10.1162/jocn

92

DeBoer, T., Scott, L. S., & Nelson, C. a. (2005). Event-related potentials in developmental populations. Methodological Handbook for Research Using Event-Related Potentials. http://doi.org/10.1016/0042-6989(96)00022-3

Di Giorgio, E., Lunghi, M., Simion, F., & Vallortigara, G. (2016). Visual cues of motion that trigger animacy perception at birth: The case of self-propulsion. Developmental Science, 1–12. http://doi.org/10.1111/desc.12394

Diesendruck, G. (2003). Categories for names or names for categories? The interplay between domain-specific conceptual structure and language. Language and Cognitive Processes, 18(5), 759–787. http://doi.org/10.1080/769813931

Dittrich, W. H., & Lea, S. E. G. (1994). Visual perception of intentional motion, 23(1944), 253–268.

Egeth, H. E., & Yantis, S. (1997). Visual attention: control, representation and time course, Annu. Rev. Psychol., 48, 269–297.

Fonlupt, P. (2003). Perception and judgement of physical causality involve different brain structures. Cognitive Brain Research, 17(2), 248–254. http://doi.org/10.1016/S0926-6410(03)00112-5

Fox, R., & McDaniel, C. (1982). The perception of biological motion by human infants. Science (New York, N.Y.), 218(4571), 486–7. http://doi.org/10.1126/science.7123249

Frankenhuis, W. E., & Barrett, H. C. (2013). Design for Learning : The Case of Chasing, 171–196.

Frankenhuis, W. E., House, B., Barrett, H. C., & Johnson, S. P. (2013a). Author ’ s personal copy Infants ’ perception of chasing Author ’ s personal copy. Cognition, 126, 224–233.

Frankenhuis, W. E., House, B., Barrett, H. C., & Johnson, S. P. (2013b). Infants’ perception of chasing. Cognition, 126(2), 224–33. http://doi.org/10.1016/j.cognition.2012.10.001

Frith, C. D., & Frith, U. (2007). Social Cognition in Humans. Current Biology, 17(16), 724–732. http://doi.org/10.1016/j.cub.2007.05.068

Fugelsang, J. A., Roser, M. E., Corballis, P. M., Gazzaniga, M. S., & Dunbar, K. N. (2005). Brain mechanisms underlying perceptual causality. Cognitive Brain Research, 24(1), 41–47. http://doi.org/10.1016/j.cogbrainres.2004.12.001

Fukuda, H., & Ueda, K. (2013). Can we perceive animacy from artficial agents ? -A brain ERP study using a motion Turing Test-. The 1st Internationa Conference on Human-Agent Interaction, 2013, 2–5.

Gallagher, H. L., & Frith, C. D. (2003). Functional imaging of “theory of mind.” Trends in Cognitive Sciences, 7(2), 77–83. http://doi.org/10.1016/S1364-6613(02)00025-6

Gallagher, H. L., Happé, F., Brunswick, N., Fletcher, P. C., Frith, U., & Frith, C. D. (2000). Reading the mind in cartoons and stories: An fMRI study of “theory of mind” in verbal and nonverbal tasks. Neuropsychologia, 38(1), 11–21. http://doi.org/10.1016/S0028-3932(99)00053-6

Gao, T., McCarthy, G., & Scholl, B. J. (2010). The Wolfpack effect: Perception of animacy irresistibly influences interactive behavior. Psychological Science, 21(12), 1845–1853. http://doi.org/10.1177/0956797610388814

93

Gao, T., Newman, G. E., & Scholl, B. J. (2009). The psychophysics of chasing: A case study in the perception of animacy. Cognitive Psychology, 59(2), 154–79. http://doi.org/10.1016/j.cogpsych.2009.03.001

Gao, T., & Scholl, B. J. (2011). Chasing vs. stalking: Interrupting the perception of animacy. Journal of Experimental Psychology: Human Perception and Performance, 37(3), 669–84. http://doi.org/10.1037/a0020735

Gao, T., Scholl, B. J., & McCarthy, G. (2012). Dissociating the Detection of Intentionality from Animacy in the Right Posterior Superior Temporal Sulcus. The Journal of Neuroscience, 32(41), 14276–14280. http://doi.org/10.1016/j.biotechadv.2011.08.021.Secreted

Gelman, R. (1990). First principles organize attention to and learning about relevant data: Number and the animate-inanimate distinction as examples. Cognitive Science, 14, 79–106.

Gelman, R., Durgin, F., & Kaufman, L. (1996). Distinguishing between animates and inanimates not by motion alone. In Causal cognition: a multidisciplinary discussion. New York. http://doi.org/10.1093/acprof:oso/9780198524021.001.0001

Gemelli, A., & Cappellini, A. (1958). The influence of the subject’s attitude in perception. Acta Psychologica, 14, 12–23. http://doi.org/10.1016/0001-6918(58)90003-9

Georgescu, A. L., Kuzmanovic, B., Santos, N. S., Tepest, R., Bente, G., Tittgemeyer, M., & Vogeley, K. (2014). Perceiving nonverbal behavior: neural correlates of processing movement fluency and contingency in dyadic interactions. Human Brain Mapping, 35(4), 1362–78. http://doi.org/10.1002/hbm.22259

Gergely, G., & Csibra, G. (2003). Teleological reasoning in infancy: The na??ve theory of rational action. Trends in Cognitive Sciences, 7(7), 287–292. http://doi.org/10.1016/S1364-6613(03)00128-1

Gergely, G., Nádasdy, Z., Csibra, G., & Bíró, S. (1995). Taking the intentional stance at 12 months of age. Cognition, 56(2), 165–193. http://doi.org/10.1016/0010-0277(95)00661-H

Gilmore, R. O., & Johnson, M. H. (1995). Working memory in infancy: six-month-olds’ performance on two versions of the oculomotor delayed response task. Journal of Experimental Child Psychology, 59, 397–418.

Gordon, I. E., Day, R. H., & Stecher, E. J. (1990). Perceived causality occurs with stroboscopic movement of one or both stimulus elements. Perception, 19(1), 17–20.

Gredebäck, G., Johnson, S. P., & von Hofsten, C. (2010). Eye tracking in infancy research. Developmental Neuropsychology, 35(1), 1–19. http://doi.org/10.1080/87565640903325758

Gredebäck, G., Kaduk, K., Bakker, M., Gottwald, J., Ekberg, T., Elsner, C., … Kenward, B. (2015). The neuropsychology of infants’ pro-social preferences. Developmental Cognitive Neuroscience, 12(January), 106–113. http://doi.org/10.1016/j.dcn.2015.01.006

Gredebäck, G., Melinder, A., & Daum, M. M. (2010). The development and neural basis of pointing comprehension. Social Neuroscience, 5(5–6), 441–50. http://doi.org/10.1080/17470910903523327

94

Halit, H., De Haan, M., & Johnson, M. H. (2003). Cortical specialisation for face processing: Face-sensitive event-related potential components in 3- and 12-month-old infants. NeuroImage, 19(3), 1180–1193. http://doi.org/10.1016/S1053-8119(03)00076-4

Haxby, J. V., Hoffman, E. A., & Gobbini, M. I. (2000). The distributed human neural system for face perception. Trends in Cognitive Sciences, 4(6), 223–233. http://doi.org/10.1016/S1364-6613(00)01482-0

Heider, F., & Simmel, M. (1944). An experimental study of apparent behavior, 57(2), 243–259.

Hein, G., & Knight, R. T. (2008). Superior temporal sulcus--It’s my area: or is it? Journal of Cognitive Neuroscience, 20(12), 2125–2136. http://doi.org/10.1162/jocn.2008.20148

Hillebrandt, H., Friston, K. J., & Blakemore, S.-J. (2014). Effective connectivity during animacy perception--dynamic causal modelling of Human Connectome Project data. Scientific Reports, 4, 6240. http://doi.org/10.1038/srep06240

Hirai, M., Fukushima, H., & Hiraki, K. (2003). An event-related potentials study of biological motion perception in humans. Neuroscience Letters, 344(1), 41–44. http://doi.org/10.1016/S0304-3940(03)00413-0

Hirai, M., Senju, A., Fukushima, H., & Hiraki, K. (2005). Active processing of biological motion perception: An ERP study. Cognitive Brain Research, 23(2–3), 387–396. http://doi.org/10.1016/j.cogbrainres.2004.11.005

Itier, R. J., & Batty, M. (2009). Neural bases of eye and gaze processing: The core of social cognition. Neuroscience and Biobehavioral Reviews, 33(6), 843–863. http://doi.org/10.1016/j.neubiorev.2009.02.004

Itier, R. J., & Taylor, M. J. (2004). Source analysis of the N170 to faces and objects. Neuroreport, 15(8), 1261–1265. http://doi.org/10.1097/01.wnr.0000127827.73576.d8

Johansson, G. (1973). Visual perception of biological motion and a model for its analysis. Perception & Psychophysics, 14(2), 201–211. http://doi.org/10.3758/BF03212378

Johansson, G., Hofsten, C. von, & Jansson, G. (1980). EVENT PERCEPTION Gunnar Johansson, Claes von Hofsten, and Gunnar Jansson. Perception, 27–63.

Johnson, M. H. (1989). Cortical Maturation and the Development of visual attention in early infancy. Jounral of Cognitive Neuroscience, 2(2), 81–95.

Johnson, M. H. (2001). Infants’ initial “knowledge” of the world: a cognitive neuroscience perspective. In F. Lacerda, C. von Hofsten, & M. Heimann (Eds.), Emerging cognitive abilities in early infancy (pp. 53–72). Mahwah, New Jersey: Lawrence Erlbaum Associates.

Johnson, M. H., Oliver, A., & Shrager, J. (1998). The Paradox of Plasticity: a constrained plasticity approach to the emergence of representations in the neocortex. Cognitive Studies, 5(2), 5–24.

Johnson, S. C. (2003). Detecting agents. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 358(1431), 549–59. http://doi.org/10.1098/rstb.2002.1237

95

Johnson, S. C., Alpha Shimizu, Y., & Ok, S. J. (2007). Actors and actions: The role of agent behavior in infants’ attribution of goals. Cognitive Development, 22(3), 310–322. http://doi.org/10.1016/j.cogdev.2007.01.002

Johnson, S. C., Slaughter, V., & Carey, S. (1998). Whose gaze will infants follow? The elicitation of gaze following in 12-month-olds. Developmental Science, 1(2), 233–238. Retrieved from http://www.scopus.com/inward/record.url?eid=2-s2.0-0000958575&partnerID=tZOtx3y1

Kaduk, K., Elsner, B., & Reid, V. M. (2013). Discrimination of animate and inanimate motion in 9-month-old infants: An ERP study. Developmental Cognitive Neuroscience, 6, 14–22. http://doi.org/10.1016/j.dcn.2013.05.003

Kagitcibasi, C., & Berry, J. W. (1989). Cross cultural psychology: Current research and trends. Annual Review of Psychology, 40, 493–531.

Kamewari, K., Kato, M., Kanda, T., Ishiguro, H., & Hiraki, K. (2005). Six-and-a-half-month-old children positively attribute goals to human action and to humanoid-robot motion. Cognitive Development, 20(2), 303–320. http://doi.org/10.1016/j.cogdev.2005.04.004

Key, A. P. F., & Stone, W. L. (2012). Processing of novel and familiar faces in infants at average and high risk for autism. Developmental Cognitive Neuroscience, 2(2), 244–255. http://doi.org/10.1016/j.dcn.2011.12.003

Key, A. P. F., Stone, W. L., & Williams, S. M. (2009). What do Infants see in faces? ERP evidence of different roles of eyes and mouth for face perception in 9-month-old infants. Infant and Child Development, 3(December 2007), 3–5. http://doi.org/10.1002/icd

Klin, A. (2000). Attributing Social Meaning to Ambiguous Visual Stimuli in Higher-functioning Autism and Asperger Syndrome: The Social Attribution Task. Journal of Child Psychology and Psychiatry, 41(7), 831–846. http://doi.org/10.1111/1469-7610.00671

Lee, S. M., Gao, T., & McCarthy, G. (2014). Attributing intentions to random motion engages the posterior superior temporal sulcus. Social Cognitive and Affective Neuroscience, 9(1), 81–7. http://doi.org/10.1093/scan/nss110

Leppänen, J. M., Moulson, M. C., Vogel-Farley, V. K., & Nelson, C. A. (2007). An ERP Study of Emotional Face Processing in the Adult and Infant Brain, 78(1), 232–245.

Leslie, A. M. (1996). A theory of agency. In D. Sperber, D. Premack, & A. J. Premack (Eds.), Causal cognition: a multidisciplinary discussion (pp. 1–39). New York. http://doi.org/10.1093/acprof:oso/9780198524021.001.0001

Leslie, A. M., & Keeble, S. (1987). Do 6-month-old infants perceive causality. Cognition, 25(April 1986), 265–288. http://doi.org/10.1016/S0010-0277(87)80006-9

Lobue, V. (2013). What Are We So Afraid of? How Early Attention Shapes Our Most Common Fears. Child Development Perspectives, 7(1), 38–42. http://doi.org/10.1111/cdep.12012

Lobue, V., & Deloache, J. S. (2010). Superior detection of threat-relevant stimuli in infancy. Developmental Science, 13(1), 221–228. http://doi.org/10.1111/j.1467-7687.2009.00872.x

96

Lobue, V., Rakison, D. H., Deloache, J. S., Lobue, V., Rakison, D. H., & Deloache, J. S. (2017). Threat Perception Across the Life Span : Evidence for Multiple Converging Pathways Linked references are available on JSTOR for this article : Threat Perception Across the Life Span : Evidence for Multiple Converging Pathways, 19(6), 375–379.

Luo, Y., & Baillargeon, R. (2005). Can a self-propelled box have a goal? - Psychological reasoning in 5-month-old infants. Psychological Science, 16(8), 601–608. http://doi.org/10.1111/j.1467-9280.2005.01582.x

Luo, Y., Kaufman, L., & Baillargeon, R. (2009). Young infants’ reasoning about physical events involving inert and self-propelled objects. Cognitive Psychology, 58(4), 441–486. http://doi.org/10.1016/j.cogpsych.2008.11.001

Martin, A., & Weisberg, J. (2003). Neural Foundations for Understanding Social and Mechanical Concepts. Cognitive Neuropsychology, 20(3–6), 575–587. http://doi.org/10.1080/02643290342000005

Mascalzoni, E., Regolin, L., & Vallortigara, G. (2010). Innate sensitivity for self-propelled causal agency in newly hatched chicks. Proceedings of the National Academy of Sciences of the United States of America, 107(9), 4483–5. http://doi.org/10.1073/pnas.0908792107

Mascalzoni, E., Regolin, L., Vallortigara, G., & Simion, F. (2013). The cradle of causal reasoning: Newborns’ preference for physical causality. Developmental Science, 16(3), 327–335. http://doi.org/10.1111/desc.12018

Melinder, A. M. D., Konijnenberg, C., Hermansen, T., Daum, M. M., & Gredebäck, G. (2015). The developmental trajectory of pointing perception in the first year of life. Experimental Brain Research, 233(2), 641–647. http://doi.org/10.1007/s00221-014-4143-2

Meyerhoff, H. S., Huff, M., & Schwan, S. (2013). Linking perceptual animacy to attention: evidence from the chasing detection paradigm. Journal of Experimental Psychology. Human Perception and Performance, 39(4), 1003–15. http://doi.org/10.1037/a0030839

Meyerhoff, H. S., Schwan, S., & Huff, M. (2014a). Interobject spacing explains the attentional bias toward interacting objects. Psychonomic Bulletin & Review, 21(2), 412–7. http://doi.org/10.3758/s13423-013-0496-y

Meyerhoff, H. S., Schwan, S., & Huff, M. (2014b). Perceptual animacy: Visual search for chasing objects among distractors. Journal of Experimental Psychology: Human Perception and Performance, 40, 702–717.

Meyerhoff, H. S., Schwan, S., & Huff, M. (2016). Oculomotor capture mediates the attentional bias toward temporarily close objects, 75637, 1–31.

Molina, M. M., Van de Walle, G. a., Condry, K., & Spelke, E. S. (2004). The Animate-Inanimate Distinction in Infancy: Developing Sensitivity to Constraints on Human Actions. Journal of Cognition and Development, 5(4), 399–426. http://doi.org/10.1207/s15327647jcd0504_1

Nairne, J. S., Pandeirada, J. N. S., Gregory, K. J., & Van Arsdall, J. E. (2009). Adaptive memory: Fitness relevance and the hunter-gatherer mind. Psychological Science, 20(6), 740–746. http://doi.org/10.1111/j.1467-9280.2009.02356.x

Natsoulas, T. (1961). Principles of Momentum and Kinetic Energy in the Perception of Causality. The American Journal of Psychology, 74(3), 394–402.

97

New, J., Cosmides, L., & Tooby, J. (2007). Category-specific attention for animals reflects ancestral priorities , not expertise, 104(42).

Nyström, P., Falck-Ytter, T., & Gredebäck, G. (2016). The TimeStudio Project: An open source scientific workflow system for the behavioral and brain sciences. Behavior Research Methods, 48(2), 542–52. http://doi.org/10.3758/s13428-015-0616-x

Oakes, L. M., & Cohen, L. B. (1990). Infant perception of a causal event. Cognitive Development, 5(2), 193–207. http://doi.org/10.1016/0885-2014(90)90026-P

Opfer, J. E. (2002). Identifying living and sentient kinds from dynamic information: The case of goal-directed versus aimless autonomous movement in conceptual change. Cognition, 86(2), 97–122. http://doi.org/10.1016/S0010-0277(02)00171-3

Pelphrey, K. A., Morris, J. P., & McCarthy, G. (2004). Grasping the intentions of others: the perceived intentionality of an action influences activity in the superior temporal sulcus during social perception. Journal of Cognitive Neuroscience, 16(10), 1706–16. http://doi.org/10.1162/0898929042947900

Pinker, S. (2005). So How Does the Mind Work? Mind and Language, 20(1), 1–24. http://doi.org/10.1111/j.0268-1064.2005.00274.x

Poulin-Dubois, D., & Heroux, G. (1994). Movement and children’s attribution of life properties. International Journal of Behavioral Development, 17(2), 329–347.

Poulin-Dubois, D., Lepage, A., & Ferland, D. (1996). Infants’ concept of animacy. Cognitive Development, 11(1), 19–36. http://doi.org/10.1016/S0885-2014(96)90026-X

Prabhakaran, V., Narayanan, K., Zhao, Z., & Gabrieli, J. D. (2000). Integration of diverse information in working memory within the frontal lobe. Nature Neuroscience, 3(1), 85–90. http://doi.org/10.1038/71156

Pratt, J., Radulescu, P. V., Guo, R. M., & Abrams, R. A. (2010). It’s Alive!: Animate Motion Captures Visual Attention. Psychological Science, 21(11), 1724–1730. http://doi.org/10.1177/0956797610387440

Price, C. J. (2012). A review and synthesis of the first 20years of PET and fMRI studies of heard speech, spoken language and reading. NeuroImage, 62(2), 816–847. http://doi.org/10.1016/j.neuroimage.2012.04.062

Puce, A., & Perrett, D. (2003). Electrophysiology and brain imaging of biological motion. Philos Trans R Soc Lond B Biol Sci, 358(1431), 435–45. http://doi.org/10.1098/rstb.2002.1221

Rakison, D. H., & Derringer, J. (2008). Do infants possess an evolved spider-detection mechanism? Cognition, 107(1), 381–393. http://doi.org/10.1016/j.cognition.2007.07.022

Reid, V., Hoehl, S., & Striano, T. (2006). The perception of biological motion by infants: An event-related potential study. Neuroscience Letters, 395(3), 211–214. http://doi.org/10.1016/j.neulet.2005.10.080

Richards, D. D., & Siegler, R. S. (1986). Children’s understandings of the attributes of life. Journal of Experimental Child Psychology, 42(1), 1–22. http://doi.org/10.1016/0022-0965(86)90013-5

98

Richards, J. E. (2003). Attention affects the recognition of briefly presented visual stimuli in infants: An ERP study. Developmental Science, 6(3), 312–328. http://doi.org/10.1111/1467-7687.00287

Richards, J. E., Reynolds, G. D., & Courage, M. L. (2010). The neural bases of infant attention. Current Directions in Psychological Science, 19(1), 41–46. http://doi.org/10.1177/0963721409360003

Rips, L. J. (2011). Causation From Perception. Perspectives on Psychological Science, 6(1), 77–97. http://doi.org/10.1177/1745691610393525

Rochat, P., Morgan, R., & Carpenter, M. (1997). Young infants’ sensitivity to movement information specifying social causality. Cognitive Development, 12(4), 537–561. http://doi.org/10.1016/S0885-2014(97)90022-8

Rochat, P., Striano, T., & Morgan, R. (2004). Who is doing what to whom? Young infants’ developing sense of social causality in animated displays. Perception, 33(3), 355–369. http://doi.org/10.1068/p3389

Rolfs, M., Dambacher, M., & Cavanagh, P. (2013). Visual adaptiation of the perception of causality. Current Biology, 23(3), 250–254. http://doi.org/10.1016/j.cub.2012.12.017

Rosander, K., Nyström, P., Gredebäck, G., & Hofsten, C. von. (2007). Cortical processing of visual motion in young infants. Vision Research, 47(12), 1614–1623. http://doi.org/10.1016/j.visres.2007.03.004

Rutherford, M. D., Pennington, B. F., & Rogers, S. J. (2006). The perception of animacy in young children with autism. Journal of Autism and Developmental Disorders, 36(8), 983–992. http://doi.org/10.1007/s10803-006-0136-8

Santos, N. S., David, N., Bente, G., & Vogeley, K. (2008). Parametric induction of animacy experience. Consciousness and Cognition, 17(2), 425–437. http://doi.org/10.1016/j.concog.2008.03.012

Santos, N. S., Kuzmanovic, B., David, N., Rotarska-Jagiela, a, Eickhoff, S. B., Shah, J. N., … Vogeley, K. (2010). Animated brain: a functional neuroimaging study on animacy experience. NeuroImage, 53(1), 291–302. http://doi.org/10.1016/j.neuroimage.2010.05.080

Saxe, R. (2006). Uniquely human social cognition. Current Opinion in Neurobiology, 16(2), 235–239. http://doi.org/10.1016/j.conb.2006.03.001

Schlaggar, B., & O’Leary, D. (1991). Potential of visual cortex to develop an array of functional units unique to somatosensory cortex. Science, 252(5012), 1556–1560.

Schlottmann, A., & Ray, E. (2010). Goal attribution to schematic animals: Do 6-month-olds perceive biological motion as animate? Developmental Science, 13(1), 1–10. http://doi.org/10.1111/j.1467-7687.2009.00854.x

Schlottmann, A., Ray, E. D., Mitchell, A., & Demetriou, N. (2006). Perceived physical and social causality in animated motions: Spontaneous reports and ratings. Acta Psychologica, 123(1–2), 112–143. http://doi.org/10.1016/j.actpsy.2006.05.006

Schlottmann, A., & Surian, L. (1999). Do 9-month-olds perceive causation-at-a-distance? Perception, 28(9), 1105–1113. http://doi.org/10.1068/p2767

Scholl, B. J., & Gao, T. (2013). Perceiving Animacy and Intentionality : Visual Processing or Higher-Level Judgment ? Social Perception: Detection and Interpretation of Animacy, Agency, and Intention, 4629, 197–230.

99

Scholl, B. J., & Leslie, A. M. (1999). Explaining the Infant’s Object Concept : Beyond the Perception/Cognition Dichotomy. What Is Cognitive Science?, 49.

Scholl, B. J., & Tremoulet, P. D. (2000). Perceptual causality and animacy. Trends in Cognitive Sciences, 4(8), 299–309. http://doi.org/10.1016/S1364-6613(00)01506-0

Schultz, J., Friston, K. J., O’Doherty, J., Wolpert, D. M., & Frith, C. D. (2005). Activation in Posterior Superior Temporal Sulcus parallels parameter inducing the percept of animacy. Neuron, 45, 625–635. http://doi.org/10.1016/j.neuron.2004.12.052

Schultz, J., Imamizu, H., Kawato, M., & Frith, C. D. (2004). Activation of the human superior temporal gyrus during observation of goal attribution by intentional objects. Journal of Cognitive Neuroscience, 16(10), 1695–1705. http://doi.org/10.1162/0898929042947874

Scott, L. S., Shannon, R. W., & Nelson, C. A. (2006). Neural correlates of human and monkey face processing in 9-month-old infants. Infancy, 10(February), 171–186. http://doi.org/10.1207/s15327078in1002_4

Senju, A., Johnson, M. H., & Csibra, G. (2006). The development and neural basis of referential gaze perception. Social Neuroscience, 1(3–4), 220–34. http://doi.org/10.1080/17470910600989797

Senju, A., Southgate, V., Snape, C., Leonard, M., & Csibra, G. (2011). Do 18-Month-Olds Really Attribute Mental States to Others?: A Critical Test. Psychological Science, (June). http://doi.org/10.1177/0956797611411584

Sherry, D. F., Jacobs, L., & Gaulin, S. (1992). Spatial memory and adaptative specialization of the hippocampus. Trends in Neuroscience, 15(8), 298–303.

Shimizu, Y. A., & Johnson, S. C. (2004). Infants’ attribution of a goal to a morphologically unfamiliar agent. Developmental Science, 7(4), 425–430. http://doi.org/10.1111/j.1467-7687.2004.00362.x

Shultz, S., Lee, S. M., Pelphrey, K., & Mccarthy, G. (2011). The posterior superior temporal sulcus is sensitive to the outcome of human and non-human goal-directed actions. Social Cognitive and Affective Neuroscience, 6(5), 602–611. http://doi.org/10.1093/scan/nsq087

Shultz, S., & McCarthy, G. (2012). Goal-directed actions activate the face-sensitive posterior superior temporal sulcus and fusiform gyrus in the absence of human-like perceptual cues. Cerebral Cortex, 22(5), 1098–1106. http://doi.org/10.1093/cercor/bhr180

Simion, F., Regolin, L., & Bulf, H. (2008). A predisposition for biological motion in the newborn baby. Proceedings of the National Academy of Sciences, 105(2), 809–813. http://doi.org/10.1073/pnas.0707021105

Sodian, B., Schoeppner, B., & Metz, U. (2004). Do infants apply the principle of rational action to human agents? Infant Behavior and Development, 27(1), 31–41. http://doi.org/10.1016/j.infbeh.2003.05.006

Spelke, E. S., Breinlinger, K., Macomber, J., & Jacobson, K. (1992). Origin of Knowledge. Psychological Review. http://doi.org/10.1037/0033-295X.99.4.605

Spelke, E. S., & Kinzler, K. D. (2007). Core knowledge. Developmental Science, 10(1), 89–96. http://doi.org/10.1111/j.1467-7687.2007.00569.x

100

Spelke, E. S., Philips, A., & Woodward, A. L. (1995). Infants’ knowledge of object motion and human action. In D. Sperber, D. Premack, & A. J. Premack (Eds.), Causal cognition: a multidisciplinary discussion (pp. 1–48). Oxford University Press. http://doi.org/10.1093/acprof:oso/9780198524021.001.0001

Sperber, D. (2005). Modularity and Relevance. In P. Carruthers, S. Laurence, & S. Stich (Eds.), The innate mind: Structure and Contents (pp. 54–80). Oxford Scholarship Online. http://doi.org/10.1093/acprof

Stets, M., & Reid, V. M. (2011). Infant ERP amplitudes change over the course of an experimental session: Implications for cognitive processes and methodology. Brain and Development, 33(7), 558–568. http://doi.org/10.1016/j.braindev.2010.10.008

Stets, M., Stahl, D., & Reid, V. . (2012). Improving tools for the analysis of brain based measures of infant information processing abstract. Developmental Neuropsychology, 37, 226–252.

Steward, J. A. (1982). Perception of animacy. University of Pennsylvania. Theeuwes, J., Kramer, A. F., Hahn, S., & Irwin, D. E. (1998). Our Eyes do Not

Always Go Where we Want Them to Go: Capture of the Eyes by New Objects. Psychological Science, 9(5), 379–385. http://doi.org/10.1111/1467-9280.00071

Tooby, J., Cosmides, L., & Barrett, H. C. (2003). The second law of thermodynamics is the first law of psychology: evolutionary developmental psychology and the theory of tandem, coordinated inheritances: comment on Lickliter and Honeycutt (2003). Psychological Bulletin, 129(6), 858–865. http://doi.org/10.1037/0033-2909.129.6.858

Tremoulet, P. D., & Feldman, J. (2000). Perception of animacy from the motion of a single object. Perception, 29(8), 943–951. http://doi.org/10.1068/p3101

Vogeley, K., Bussfeld, P., Newen, A., Herrmann, S., Happé, F., Falkai, P., … Zilles, K. (2001). Mind Reading: Neural Mechanisms of Theory of Mind and Self-Perspective. NeuroImage, 14, 170–181. http://doi.org/10.1006/nimg.2001.0789

Von Hofsten, C., & Rosander, K. (1996). The development of gaze control and predictive tracking in young infants. Vision Research, 36(1), 81–96. http://doi.org/10.1016/0042-6989(95)00054-4

Von Hofsten, C., & Rosander, K. (1997). Development of smooth pursuit tracking in young infants. Vision Research, 37(13), 1799–1810. http://doi.org/10.1016/S0042-6989(96)00332-X

Vouloumanos, A., & Werker, J. F. (2007). Listening to language at birth: Evidence for a bias for speech in neonates. Developmental Science, 10(2), 159–164. http://doi.org/10.1111/j.1467-7687.2007.00549.x

Wagemans, J., Elder, J., Kubovy, M., Palmer, S., Peterson, M., Singh, M., & von der Heydt, R. (2012). A Century of Gestalt Psychology in Visual Perception I. Perceptual Grouping and Figure-Ground Organization. Psychological Bulletin, 138(6), 1172–1217. http://doi.org/10.1037/a0029333.A

Wertz, A. E., & Wynn, K. (2014a). Selective social learning of plant edibility in 6- and 18-month-old infants. Psychological Science, 25(January), 874–82. http://doi.org/10.1177/0956797613516145

101

Wertz, A. E., & Wynn, K. (2014b). Thyme to touch: Infants possess strategies that protect them from dangers posed by plants. Cognition, 130(1), 44–49. http://doi.org/10.1016/j.cognition.2013.09.002

White, A. L., Rolfs, M., & Carrasco, M. (2013). Adaptive deployment of spatial and feature-based attention before saccades. Vision Research, 85, 26–35. http://doi.org/10.1016/j.visres.2012.10.017

White, P. A., & Milne, A. (1999). Impressions of enforced disintegration and bursting in the visual perception of collision events. Journal of Experimental Psychology: General, 128(4), 499–516. http://doi.org/10.1037/0096-3445.128.4.499

Zilbovicius, M., Meresse, I., Chabane, N., Brunelle, F., Samson, Y., & Boddaert, N. (2006). Autism, the superior temporal sulcus and social perception. Trends in Neurosciences, 29(7), 359–366. http://doi.org/10.1016/j.tins.2006.06.004

Zipster, D., & Andersen, R. A. (1988). A back-propagation programmed network that simulates response properties of a subset of posterior parietal neurons. Nature, 331, 679–684. http://doi.org/10.1038/332141a0

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Social Sciences 144

Editor: The Dean of the Faculty of Social Sciences

A doctoral dissertation from the Faculty of Social Sciences,Uppsala University, is usually a summary of a number ofpapers. A few copies of the complete dissertation are keptat major Swedish research libraries, while the summaryalone is distributed internationally through the series DigitalComprehensive Summaries of Uppsala Dissertations from theFaculty of Social Sciences. (Prior to January, 2005, the serieswas published under the title “Comprehensive Summaries ofUppsala Dissertations from the Faculty of Social Sciences”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-321904

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017