Spatial Deficits in an Amnesic Patient with Hippocampal Damage:Questioning the Multiple Trace Theory

A. Gomez,1* S. Rousset,1 and A. Charnallet1,2

ABSTRACT: Mediotemporal lobe structures are involved in both spa-tial processing and long-term memory. Patient M.R. suffers from amne-sia, due to bilateral hippocampal lesion and temporoparietal atrophyfollowing carbon monoxide poisoning. We compared his performance inimmediate spatial memory tasks with the performance of ten healthymatched participants. Using an immediate reproduction of path, weobserved a dissociation between his performance in three allocentrictasks and in five egocentric-updating tasks. His performance was alwaysimpaired on tasks requiring the use of an egocentric-updating represen-tation but remained preserved on allocentric tasks. These results fit withthe hypothesis that the hippocampus plays a role in spatial memory, butthey also suggest that allocentric deficits previously observed in amnesiamight actually reflect deficits in egocentric-updating processes. Further-more, the cooccurrence of deficits in episodic long-term memory andshort-term egocentric-updating representation without any short-termallocentric deficit suggests a new link between the mnemonic and navi-gational roles of the hippocampus. The Cognitive Map theory, the Multi-ple Trace theory, as well as further models linking spatial and nonspatialfunctions of the hippocampus are discussed. VVC 2011 Wiley-Liss, Inc.

KEY WORDS: navigation; allocentric; spatial memory; episodic memory

INTRODUCTION

Hippocampal lesion is a key feature of the amnesic syndrome(Scoville and Milner, 1957). Delayed recall is predictably impairedamong patients with hippocampal damage [for a review, see Spiers et al.(2001)]. Thus, the hippocampus is assumed to be of critical importancefor long-term memory and especially for episodic memory, this abilityto consciously remember events, and their unique spatiotemporal con-text. Depending on hippocampal lesions, deficits have also been reportedin spatial navigation (e.g., Burgess et al., 2002). Although hippocampal

involvement in spatial tasks is not controversial, itsprecise spatial function as well as its relationship toepisodic memory remains debated (Squire and Zola-Morgan, 1991; Worsley et al., 2001; Burgess, 2002;Lipton and Eichenbaum, 2008). This present articlegets out the results of a patient with hippocampaldamage suffering from spatial deficits questioning the‘‘Cognitive Map’’ theory (O’Keefe and Nadel, 1978).Our results support the hypothesis that self-motioninformation integration might be disrupted in amnesicsyndrome, rather than allocentric representation (i.e.,object-to-object relations).

Multiple Trace theory (MTT, Teyler and DiScenna,1986; Damasio, 1989; Squire and Zola-Morgan,1991; Nadel and Moscovitch, 1998; Teyler and Rudy,2007) assumed a functional relation between twokinds of memory, episodic memory and spatial mem-ory, being under the control of the hippocampal for-mation. The well-known Cognitive Map theory(CMT) has progressively changed into this MTT hy-pothesis (Nadel and Moscovitch, 1998). Followingthis former theory, allocentric spatial map (i.e., a rep-resentation encoding relations between objects, inde-pendent of the observer’s position) is stored in thehippocampus (O’Keefe and Nadel, 1978). This hy-pothesis raised from the discovery of ‘‘place cells’’ inthe hippocampus of freely moving rats (O’Keefe andDostrovsky, 1971). In the MTT, Nadel and Mosco-vitch (1998) extended this idea to episodic memory.They argued that recalling an event requires a contex-tual hippocampal ‘‘trace’’ access. More precisely, thistheory states that hippocampus provides a spatial scaf-fold, binding all the neocortical representations relatedto each episode. This spatial scaffold is presumed toarise from allocentric spatial encoding.

Numerous studies have reported a cooccurrencebetween impairments of topographical and episodicmemory following hippocampal lesions (e.g., Hold-stock et al., 2000; Burgess et al., 2001, 2006; Spierset al., 2001; King et al., 2002; Hort et al., 2007).However, few studies clearly investigated the underly-ing nature (e.g., allocentric or egocentric) of the spa-tial deficits related to episodic memory impairmentsafter hippocampal lesions (Holdstock et al., 2000;King et al., 2002, 2004). Usually, allocentric represen-tations are opposed to static or iconic–egocentric re-presentations, which are considered as inadequate for

1 Laboratoire de Psychologie et NeuroCognition, CNRS–UMR 5105, Uni-versite Pierre Mendes-France, BP 47, 38040 Grenoble Cedex 09, France;2CMRR and Neuropsychologie, Pole de psychiatrie et Neurologie, CHUde Grenoble, BP 217, 38043 Grenoble Cedex 09, FranceAdditional Supporting Information may be found in the online version ofthis article.Contract Grant sponsor: French Grant of the Research and National Edu-cation Department.*Correspondence to: Alice Gomez, Laboratoire de Psychologie et Neuro-Cognition, CNRS-UMR 5105, Universite Pierre Mendes-France, BP 47,38040 Grenoble Cedex 09, France.E-mail: alice.gomez@upmf-grenoble.frAccepted for publication 20 June 2011DOI 10.1002/hipo.20968Published online in Wiley Online Library (wileyonlinelibrary.com).

HIPPOCAMPUS 00:000–000 (2011)

VVC 2011 WILEY-LISS, INC.

long-term memory storage, due to permanent modifications inthe subject’s localization and orientation. In contrast, allocentricrepresentations are independent of subject’s movement. Theyare thus considered as being more stable (Burgess et al., 2001)and, therefore, suitable for long-term memory storage. A fewavailable studies (Holdstock et al., 2000; King et al., 2002,2004) seem to confirm this point of view. These experimentalinvestigations compare patients with hippocampal lesions witha control group (CG) in an iconic–egocentric condition (i.e., avisual pattern matching task) and a shifted-view condition (i.e.,a matching task between two snapshots of the identical envi-ronment but seen from a different perspective). As a whole,these results indicate that amnesic patients have relatively pre-served performance in the first condition and impaired ones inthe latter. This functional dissociation observed in amnesiaraises the hypothesis that episodic memory relies on spatialprocesses involved in shifted-view conditions, rather than iniconic–egocentric ones.

However, spatial memory cannot be limited to either allo-centric or iconic–egocentric processes (Vann et al., 2009; Avraa-mides and Kelly, 2008; Waller et al., 2008). In the previousstudies, the shifted-view condition can be concurrently solvedusing two types of process: an allocentric process or an egocen-tric-updating one. Indeed, participants might have representedspace in an allocentric framework, with no reference to their ini-tial point of view in order to succeed. Nevertheless, they alsocould have mentally simulated an update of their egocentric rep-resentation due to their viewpoint rotation, as the authorsacknowledged (King et al., 2004). Thus, if the aforementionedstudies suggest that iconic–egocentric processing of space is notsuitable for episodic memory, two possible types of spatial proc-essing still remain: (a) an allocentric processing (i.e., code forstatic object-to-object relations) as proposed by the MTT modelor (b) an egocentric processing with self-motion [(i.e., code forself-to-environment relations in a dynamic fashion, Farrell andRobertson (1998)]. The latter has also been suggested to belinked with episodic memory (Gomez et al., 2009).

Egocentric-updating representation might be related to ‘‘pathintegration’’ process (Wishaw, Cassel, Jarrard, 1995; Wishaw,Hines, Wallace, 2001) as it also operates using idiothetic cues(i.e., self-motion cues). More precisely, egocentric-updating rep-resentation would primarily rely on idiothetic information (i.e.,self motion cues), even if inputs from allothetic information(i.e., sensorial information extracted from stable stimuli) mightalso contribute to the formed representation. Egocentric-updat-ing representation is best described as a dynamic self-to-envi-ronment relation during navigation, elaborated from continu-ous vestibular, proprioceptive, and visual inputs. Moreover,contrary to allocentric representations, egocentric-updating in-formation is centered on the subject and might be classified asan egocentric reference frame. From a neuroanatomical pointof view, this process could involve areas of the parietal cortex,usually associated with egocentric space processing (Chokron,2003; Nitz, 2009) and the hippocampal formation(McNaughton et al., 1991, 2006). Three studies (Bures et al.,1998; Worsley et al., 2001; Philbeck et al., 2004) have already

reported evidences of pure path integration deficits amongpatients with hippocampal lesions.

Allocentric and egocentric-updating processes could belinked by a hierarchical dependence, which could make theirproper evaluation more complicated. For instance, it has beenproposed that path integration process would be responsiblefor assembling space fragments into a larger map of an entirearea, thus underlying the elaboration of a cognitive map (Mit-telstaedt and Mittelstaedt, 1980; Etienne and Jeffery, 2004).This leads to predict that some deficits in allocentric represen-tations could be due to problems in the preliminary egocen-tric-updating processes, rather than to an allocentric deficitper se. Thus, in order to distinguish allocentric level deficitsfrom egocentric-updating level deficits, allocentric representa-tion evaluations need to be more specific. Pure allocentricrepresentations could be assessed in situations where the repre-sentation can be constructed on the sole basis of allothetic in-formation, with no need to integrate beforehand idiotheticinformation.

By testing M.R, an amnesic patient, on his ability to processallocentric or egocentric-updating spatial information, we aimto confront (1) a possible role of egocentric-updating represen-tation for episodic memory (Gomez et al., 2009) to (2) theMTT hypothesis conferring a specific role to allocentric repre-sentation for episodic memory (Nadel and Moscovitch, 1998;Moscovitch et al., 2005).

For this purpose, we compared M.R. to 10 matched-controlson either motor path reproduction tasks (Experiment 1) or onpath reproduction drawing tasks (Experiment 2). In bothexperiments, each condition was divided in two parts: (1) apath encoding phase, followed by (2) an immediate reproduc-tion phase, with a short total duration to overcome relatedmemory problems. Error sizes in angle production during path(Exp. 1) or drawing reproduction (Exp. 2) were compared. Thegeneral principle of this study was to oppose egocentric-updat-ing to allocentric representation. This was done so by varyingthe reproduction mode as well as the spatial information avail-able during encoding (e.g., with or without auditory and visualcues) and reproduction.

On one hand, conditions relying on idiothetic informationalone assessed egocentric-updating representation. On the otherhand, conditions relying on allothetic information aloneassessed allocentric representation per se. Moreover, to guaran-tee the evaluation of the allocentric representation, two addi-tional methods were used: (1) a reverse way reproduction con-dition and (2) a drawing reproduction assessing directly allo-centric abilities.

We predicted that M.R. would not be able to process andintegrate idiothetic information, neither during its encodingnor during its reproduction. Hence, whenever egocentric-updat-ing information is necessary and compulsory to accurately solvethe task, M.R. would be impaired. We also expected a relativepreservation of his performance in conditions relying on anallocentric representation when it could be built upon allotheticinformation alone, without reliance on preliminary egocentric-updating process.

2 GOMEZ ET AL.

Hippocampus

MATERIALS AND METHODS

Participants

Control group

The control group (CG) of the path motor reproductiontasks was composed of 10 male participants [age range 65–70,mean age 66.7, and standard deviation (SD) 2.0] with a high-educational background, paired with M.R. on age, sex, andsociocultural background.

Patient M.R

The reader is invited to refer to Bastin and colleagues’ report(2004) for an extensive case description detailing life background,postlesion deficits and recovery, extensive neuropsychological,and magnetic resonance imaging (MRI) examination. To summa-rize, M.R. is a 65-year-old right-handed man who was intoxicatedwith carbon monoxide at the age of 48. Structural MRI revealed abilateral atrophy of the hippocampi that extended to the rightmediotemporal lobe. The corrected volume of the hippocampiwas reduced by 57% on the right and 67% on the left relative tothe ten controls’ mean (the uncorrected hippocampal volume losswas 49% on the right and 60% on the left). The corrected volumeof the right mediotemporal lobe (including the right parahippo-campal gyrus) was reduced by 35% (uncorrected volume loss was23%). The volumes for the left and right hemispheres—correctedfor intracranial volumes—were significantly reduced, especially inthe parietal lobes.

In 2000, his full-scale WAIS-R score was 122, whereas hisWechsler Memory Scale-Revised score was 83, resulting in aWAIS-WMS difference of 39 points. On long-term memoryassessment, M.R. is severely impaired in free and cued recall[task adapted from Grober and Buschke (1986)], both on im-mediate and delayed testing. However, Bastin and colleagues(2004) evidenced a spared familiarity-based recognition. M.R.has a normal short-term memory span, normal visuospatialabilities [in mental rotation, Shepard and Metzler (1971), iden-tification of degraded letters, Warrington and James (1991)],visual motor functions (copy of the Rey–Osterriech ComplexFigure Test), and executive functions [Trail making test, Strooptest (Stroop, 1935)].

Experiment 1

Procedure outline and purpose

Each trial was divided in two sequences (see Fig. 1): a pathencoding phase (with a guided return to the original location)and immediately after a path reproduction. Two path-encodingconditions have been compared (see top of Fig. 1): (1) encod-ing through the direct path reproduction (i.e., idiothetic-move-ment encoding); in this experiment, vision is always availablein this phase; (2) encoding, via the observation of the experi-menter producing the path (i.e., allothetic movement encod-ing). In the idiothetic-movement encoding condition (i.e., first-person movement), participants were guided along the correct

path by an experimenter and then performed the path task. Inthe allothetic-movement encoding condition (i.e., third-personmovement), participants watched the experimenter producingthe path, but did not perform it. For the immediate recall task,three path production conditions have been compared (see bot-tom of Fig. 1): (1) a same-way path production, (2) an oppo-site-way path production (i.e., path had to be produced startingfrom the end), and (3) a solely idiothetic path production (i.e.,without visual or auditory inputs). In the same-way path pro-duction condition, participants normally reproduced the pathwith all sensory information available (i.e., allothetic and idiot-hetic). In the opposite-way path production condition, partici-pants reproduced the path with all sensory information avail-able but in the opposite-way (i.e., starting from the end of thepath, and reproducing it in the reverse way). In the solely idiot-hetic path production condition, participants wore a helmetand a visual mask during the path reproduction.

The idiothetic-movement encoding condition provides bothidiothetic (i.e., self-motion) and allothetic information (i.e.,landmark, room configuration). However, in order to succeed,participants mainly needed self-motion encoding (i.e., idiot-hetic information). On the contrary, the allothetic-movementencoding condition only provides information on the stablestimuli configuration (visual and auditory), such as landmark,room configuration, and external movement perception (i.e.,allothetic information). As for the path reproduction, in thesolely idiothetic path production condition, participants couldonly rely on self-motion information (i.e., idiothetic) to solvethe task correctly. Although in the same-way path productionand opposite-way path production conditions, they could relyon both allothetic and idiothetic cues. When in the same-waypath production condition, participants may have relied onboth allocentric and iconic–egocentric snapshots information.However, in the opposite-way path production condition, theyshould rely on an allocentric information because of the modi-fication of the initial position.

As a summary, the participants completed our task under 6conditions (see Fig. 1): four conditions where idiothetic infor-mation was crucial in order to solve the task, assessing egocen-tric-updating representation (idiothetic-movement encoding-same-way path production; idiothetic-movement encoding-op-posite-way path production; idiothetic-movement encoding—solely idiothetic path production; allothetic-movement encod-ing—solely idiothetic path production); two conditions whereallothetic information was compulsory to solve the task, assess-ing allocentric representation (allothetic-movement encoding-same-way path production; allothetic-movement encoding-opposite-way path production).

Methodology

The experimental paradigm took place in a room of �4 36 m with two doors on one side, four windows on anotherside, and a blackboard on a third side (see Supporting Informa-tion Fig. A). No visual features were present on the floor. Ineach of the encoding conditions, participants walked along 15

SPATIAL DEFICITS IN AN AMNESIC CASE WITH HIPPOCAMPAL DAMAGE 3

Hippocampus

different paths. Path included two to four turns (ranging from458 to 3608, mean angle 1278, and SD 728). The amplitudeand number of turns were counterbalanced across conditions.Routes’ length was 16.5 m long on average (10–20 m, mean16.5 m, and SD 3.2). For each trial, encoding phase lastedexactly 30 s and immediate reproduction of paths lasted, forthe CG, 17.26 s on average (SD 3.85, ranging from 9.9 to25.2 s), for M.R., 30.2 s (SD 14.06, ranging from 12 to 70 s).As the starting and ending points were different, participantswere directly guided to the starting point of the trial after eachencoding production. Mean duration was equivalent acrossconditions (F < 1).

To ensure that possible tiring or practice effects were constantacross participants, experimental trials were performed in thesame order for each participant. Trials of the encoding conditionwere presented in block and in the following order: (1) idiotheticmovement encoding and (2) allothetic movement encoding. Tri-als from the production conditions were in the following order:same-way path production, trials 1, 4, 7, 10, and 13; opposite-way path production, trials 2, 5, 8, 11, and 14; solely idiotheticpath production, trials 3, 6, 9, 12, and 15. Participants trainedonce for each condition, before the assessment. An assistant ex-perimenter depicted participants’ motor production on a map ofthe environment containing a control grid (see Supporting Infor-

FIGURE 1. Experiment 1 procedure. Top: Path encoding con-ditions: left, idiothetic movement encoding condition; right, allo-thetic movement encoding condition. P is for participant and forpatient; E is for Experimenter. In the idiothetic movement encod-ing condition, participants simply walked along the path. It pro-vided both idiothetic and allothetic information. In the allotheticmovement-encoding condition, participants did not really walkthe path, but they watched the experimenter producing it. It pro-vided solely allothetic information. Bottom: Path reproductionconditions: Left, same-way production; center, opposite-way pro-

duction; right, solely idiothetic production. In the same-way pathproduction and opposite-way path production conditions, the par-ticipants simply reproduced the path. They could rely on allo-thetic and idiothetic information. In the opposite-way path pro-duction condition, participants started their path from the endand reproduced it in reverse to insure the use of allocentric abil-ities. In the solely idiothetic path production condition, partici-pants reproduced the path blindfolded and with a helmet. Theyrelied solely on idiothetic information. [Color figure can be viewedin the online issue, which is available at wileyonlinelibrary.com.]

4 GOMEZ ET AL.

Hippocampus

mation Fig. A). This allowed a measure of the angles producedand their comparison with the expected angle.

RESULTS AND DISCUSSION

Data analysis

The experimenter recorded participants’ trajectories and angleproduction on each expected turn. From these recorded data, theunsigned (absolute value) angle error size was calculated betweenthe expected angle and the observed angle. If participants did notproduce any turns (i.e., forgot the turn), the maximal unsignedangle error was attributed to the trial (i.e., 3608). In each condi-tion, this happened less than one time upon the 15 trials for theCG (M 5 0.51 and SD 5 0.72) and for M.R. (M 5 0.5, SD 50.83). Angle error size is expressed in degrees.

Control group

For the control group (CG), a repeated measures analysis ofvariance (ANOVA) was conducted on the angle error size ineach encoding (idiothetic-movement encoding and allothetic-movement encoding) and production (same-way path produc-tion, opposite-way path production, and solely idiothetic pathproduction) conditions. Mean and standard error are illustratedin Figure 2. The analysis revealed no effect of the encodingconditions [F < 1], nor of the production condition [F < 1].Furthermore, no interaction between encoding and productionconditions [F(2, 18) 5 1.48, P < 0.25] was observed.

Case M.R.

For each condition, modified t-tests (Crawford, Howell andGarthwaite, 1998) were carried out to compare M.R. and CG’sperformance. Figure 2 (top) shows the mean scores of M.R.and of the control participants in the path motor productiontask. M.R.’s corresponding modified Z-scores are illustrated inFigure 2 (bottom). Classical functional dissociations betweenconditions are calculated following Crawford, Garthwaite, andGray’s recommendations (Crawford et al., 2009)—indicated onFigure 2. These results follow the three following criteria fortwo tasks X and Y, as described by Crawford et al. (2009):

Criterion 1.

Patient’s score on Task X is significantly lower than controlsusing Crawford et al. (1998) modified t-test method [(indicatedby a star on the top of Figures 2 and 4)].

Criterion 2. Patient’s score on Task Y is not significantly lowerthan in the CG (i.e., score difference fails to meet threshold for adeficit and is therefore considered to be within normal limits).

Criterion 3. Patient’s score on Task X is significantly lowerthan patient’s score on Task Y using Crawford et al.’s (1998)test (indicated by a star on the bottom of Figures 2 and 4).

According to the first criterion proposed by Crawford et al.to evidence a functional dissociation (see the top of Fig. 2), wecan conclude that M.R. is impaired in all reproduction condi-tions when encoding relies on idiothetic movement information(all Ps < 0.05). In fact, M.R.’s performance was alwaysimpaired when he learned his path while he walked, whateverthe reproduction task was (i.e., same-way, opposite-way, orsame-way without visual or auditory inputs).

M.R.’s reproduction performance was impaired even if allo-thetic information about the surroundings was still available.We could have expected M.R. to compensate for his impair-ments during the reproduction, in the same-way or in the op-posite-way condition, by using strategies relying on allotheticinformation (e.g., by encoding allothetic information such asvisual snapshot of key landmarks to reproduce the path). Our

FIGURE 2. Experiment 1 results. Top: Overall error size indegrees for M.R. and the control group in each encoding condi-tion (idiothetic movement encoding condition; allothetic move-ment encoding condition) for each reproduction condition (same-way path production; opposite-way path production; solely idiot-hetic production). For the control group, the mean score is shownby the bar, and error bars indicate the standard error. PatientM.R. presents a bilateral medial temporal lobe lesion including thehippocampi (refer to the text for further details). Bottom: Corre-sponding Z-scores derived by comparing M.R. to the controlgroup. The dashed line shows T 5 1.796, P < 0.05. * indicates P< 0.05 deficits, * indicates a functional dissociation with P < 0.05.

SPATIAL DEFICITS IN AN AMNESIC CASE WITH HIPPOCAMPAL DAMAGE 5

Hippocampus

results fail to indicate any improvement of M.R.’s performancein the reproduction condition with available allothetic informa-tion (i.e., same-way and opposite-way) after an idiothetic move-ment encoding (i.e., self-movement). Moreover, during thesolely idiothetic reproduction stage, M.R. was impaired in theallothetic movement encoding condition (P < 0.05). All theseresults point toward a deficit in egocentric-updating representa-tion secondary to an idiothetic information processing deficit.

Regarding the second criterion, M.R.’s performance was pre-served when he could solve the task using allothetic informa-tion (Ps > 0.5, for the same-way and opposite-way productionconditions after an allothetic movement encoding condition).This is particularly striking for the opposite-way reproductioncondition where only allocentric strategy was available. Hence,this result shows that M.R.’s performance is preserved whenencoding relies on allothetic information, suggesting a preserva-tion of allocentric representation.

Last, according to the third criterion defined by Crawfordet al.’s (2009), when encoding was performed in the idiothetic-movement encoding condition, the patient’s score on the solelyidiothetic path production was significantly lower than his scoreon the opposite-way path production (P < 0.05); MR’s scoreon the former (solely idiothetic path production) was alsolower than his score in the same-way path production condi-tion (P < 0.05). Moreover, M.R.’s performance was lower inthe idiothetic-movement encoding than in the allothetic-move-ment encoding condition (Ps < 0.05) for the same-way pathproduction and the opposite-way path production conditions.However, better performance during the allothetic-movementencoding compared to the idiothetic-movement encoding wasnot evidenced if path reproduction was done in the solely idi-othetic condition (P > 0.5).

We predicted that M.R. would be deficient in the four con-ditions assessing egocentric-updating representation and pre-served in the two other conditions (i.e., whenever an allocentricrepresentation is possible). According to this prediction, M.R.’sscore should be lower in conditions assessing egocentric-updat-ing representation than allocentric representation condition. Asa summary, the results from the first and third criterion ofCrawford et al. (2009) indicate that M.R. has a deficit inencoding and producing idiothetic information and is thusimpaired in the use of an egocentric-updating representation.On the other hand, the results from the second criterion indi-cate that M.R. is able to encode a movement based on allo-thetic information and therefore show that his ability to use anallocentric representation is preserved.

This experiment provides a striking result: a preservation ofallocentric spatial processing following hippocampal damage.However, although these results clearly point toward a deficitin idiothetic information processing, the experiment did notdirectly assess the encoding of idiothetic information alone. Wetherefore designed a second experiment in order to test thisspecific ability. In this second experiment, we tried to demon-strate the preservation of M.R.’s ability in cognitive mapping(i.e., allocentric abilities) more directly by testing his ability toreproduce the path with a map drawing,

Experiment 2

Procedure Outline and Purpose

The procedure was very similar to that of the first experi-ment (see Fig. 3). First, it differed from the previous experi-ment on the reproduction modality: participants had to pro-duce a drawing of the path on a map of the environment(instead of a motor reproduction); Secondly, during the encod-ing conditions, one of those was purely idiothetic, and no vis-ual or auditory cues were available. This procedure clearly dis-sociates idiothetic and allothetic information during encoding.

As a consequence, the path encoding was based either onsolely idiothetic information (with no visual or auditory inputs,i.e., solely idiothetic) or, as in the previous experiment, via theobservation of the experimenter producing the path (i.e., allo-thetic movement, see Fig. 3). During the encoding phase of thesolely idiothetic condition, participants wore visual and audi-tory masks and were guided by the experimenter on a path.During the encoding phase of the solely allothetic condition,participants watched the experimenter producing the path, andthey did not perform the movement. They only acquired infor-mation from the stable stimuli configuration (visual and audi-tory)—such as landmark, room configuration and externalmovement perception (i.e., allothetic information). In this con-dition, the crucial information allowing to solve the task (i.e.,the path or movement information in itself ) relied solely onallothetic cues. During the path-drawing phase, participantswere asked to draw a map of their path as accurately as possi-ble. This procedure was chosen in order to assess mapping abil-ities as directly as possible (i.e., allocentric abilities). Resultswere analyzed with the same procedure as in Experiment 1.

Methodology

Four trajectories were used with three to four turns (rangingfrom 458 to 1358, mean angle 998, and SD 298). Mean lengthof routes was 18.75 m long (15–21 m, SD 2.6). Each triallasted less than 40 s for the CG and for M.R.

RESULTS AND DISCUSSION

Control group

For the control group (CG), a repeated measures ANOVAwas conducted on the angle error size with encoding conditions[(1) idiothetic-movement; (2) allothetic movement] as awithin-participant factor. Mean and standard error are plottedin Figure 4. The analysis did reveal an advantage of the 2-allo-thetic encoding condition over the 1-solely idiothetic encodingcondition [F(1, 9) 5 7.83, MSe 5 1,372.9, P < 0.05].

Case M.R.

Mean scores of M.R. and corresponding Z-score of thedrawing reproduction task are displayed in Figure 4. Results

6 GOMEZ ET AL.

Hippocampus

indicate that M.R. is impaired in the solely idiothetic encodingcondition compared to the CG (P < 0.05), but not in the allo-thetic encoding condition. Moreover, these results indicate aclear dissociation between these tasks (P < 0.05).

These results provide additional data consistent with previousresults obtained with the self-movement encoding condition.They clearly indicate that the results observed in the 1-idiot-hetic movement encoding condition of experiment 1 are due toan idiothetic information processing deficit. More importantly,the replication of preserved performance in the allothetic condi-tion during path drawing reproduction clearly points toward apreservation of allocentric representation.

GENERAL DISCUSSION

The present study aims was (1) to provide further insight onthe spatial deficits observed after a bilateral hippocampal lesioninvolving temporo-parietal structures and (2) to contrast an hy-

pothesis allocating a specific role to egocentric-updating in epi-sodic memory (Gomez et al., 2009) with concurrent theoriesallocating a specific role to allocentric representations, like theMTT (Nadel and Moscovitch, 1998; Moscovitch et al., 2005).

To investigate these questions, patient M.R. (who suffered fromamnesia due to bilateral hippocampal damage following carbonmonoxide poisoning) was compared to a CG on several spatialtasks. The experiments were designed to assess egocentric-updatingwith self motion representation and allocentric representation.Each condition included a path encoding followed by immediatereproduction. Information available during encoding and repro-duction, as well as the reproduction modality, differed across con-ditions. On one hand, egocentric-updating tasks relied mainly orsolely on idiothetic information (i.e., self-motion cues), whereasallocentric tasks relied mainly on allothetic information. Moreover,motor reproduction (i.e., same-way and opposite-way) and pathsketch map were the two different reproduction modalities chosento insure the use of allocentric representations, with an increasingdemand on allocentric process. Each participant’s performance wasassessed through error size in angle production.

FIGURE 3. Experiment 2 procedure. Top: Path encoding con-ditions: Left, solely idiothetic; Right, solely allothetic. P is for par-ticipant and for patient; E is for Experimenter. In the solely idiot-hetic conditions, participants produced the path blindfolded andwith a helmet. It provided solely idiothetic information. In thesolely allothetic condition, participants did not produce the path,

but they watched the experimenter producing it. It provided solelyallothetic information. Bottom: The path drawing reproductionassesses directly allocentric representation through mapping abil-ities. [Color figure can be viewed in the online issue, which isavailable at wileyonlinelibrary.com.]

SPATIAL DEFICITS IN AN AMNESIC CASE WITH HIPPOCAMPAL DAMAGE 7

Hippocampus

Evidencing an Idiothetic Information ProcessingDeficit: Path Integration, Egocentric-Updating,and the Hippocampo-Parietal Network

This study indicates that M.R. was always impaired when hehad to learn the path on the basis of idiothetic information,whatever the restitution mode was. He was also impaired inthe allothetic encoding condition, but only if he had to rely onidiothetic information during the reproduction phase.

These deficits are congruent with numerous evidence sug-gesting that medial temporal lobe structures are important forspatial cognition, including animal electrophysiology or lesion(O’Keefe and Dostrovsky, 1971; Morris et al., 1982), humanpatients investigations, or neuroimaging studies of spatial navi-gation (Maguire et al., 1996a,b; Teng and Squire, 1999; Hold-stock et al., 2000; Mellet et al., 2000; King et al., 2004).

However, despite this wealth of literature on spatial memoryand medio-temporal structures, the precise nature of spatialprocess remains unclear. In humans, most lesion studies involv-ing medio-temporal lobe have focused on topographical impair-ment, such as topographical agnosia (Mendez and Cherrier,2003) or anterograde topographical disorientation (Maguire

et al., 1996a). Most studies assessing spatial memory of patientswith hippocampal lesions have been taken as evidence of thegenerally admitted allocentric function of this structure (Hold-stock et al., 2000; Burgess et al., 2001; Spiers et al., 2001;King et al., 2002; Burgess, 2006; Hort et al., 2007). Forinstance, taking advantage of the numerous paradigms devel-oped to assess rodents’ spatial memory, some studies have inves-tigated the effect of hippocampus damage on navigation inhumans. Using an adapted version of the famous Morris Watermaze task, it has been shown that patients with hippocampallesions (surgical resections or brain trauma) were severelyimpaired in spatial navigation (Astur et al., 2002; Skeltonet al., 2006; Livingstone and Skelton, 2007; Goodrich-Hun-saker et al., 2010). However, these studies did not help todetermine whether the observed deficit was due to an inabilityto use path integration or a deficit in allocentric process. Vir-tual reality tasks, in which subjects did not actually move butonly watched the surrounding changes on a computer screen,might have led to an approximation of human spatial memoryin the real world (Lavenex and Lavenex, 2010; Lavenex et al.,2011). In fact, the construction of a cognitive map usuallyrelies on idiothetic information and path integration processing[responsible for space fragments assembling, either from a realor a mental exploration, Etienne and Jeffery (2004) and Mittel-staedt and Mittelstaedt (1980)]. Although some studies haveused real environment to assess hippocampal and parahippo-campal involvement in spatial memory, their tasks did notdirectly assess path integration processing (Bohbot et al., 1998,2000). As acknowledged by several authors (e.g., Astur et al.,2002), deficits observed in such human patients studies couldarise either from a path integration deficit (i.e., or more widelyan egocentric-updating processes deficit) or from a real cogni-tive mapping deficit (i.e., allocentric representation). The pur-pose of our study was to disentangle both interpretations. Toour knowledge, this is the first study to ecologically assess pathintegration deficits in navigation, after medio-temporal lesions,by manipulating visual and auditory cues available (i.e., allo-tethic cues).

The deficits evidenced in M.R. investigations support the hy-pothesis that he cannot process idiothetic information, neitherduring the encoding phase nor during the reproduction phase.This is congruent with previous studies reporting path integra-tion deficits following hippocampal damage both in humans(Worsley et al., 2001; Philbeck et al., 2004) and in the animals(Whishaw et al., 2001). Moreover, our results clearly show thatthe spatial deficit is massive and restricted to idiothetic infor-mation processing. This result is of major importance as noprevious studies disentangled allocentric from egocentric-updat-ing deficit.

No Deficit in Allothetic Information Processing:Bringing the MTT Into Question

In experiments 1 and 2, M.R.’s performance was unimpairedin several tasks where encoding relied on an external movementobservation, whatever the reproduction mode (same way, oppo-

FIGURE 4. Experiment 2 results. Top: Overall error size indegrees for M.R. and the control group in each encoding condi-tion (A: solely idiothetic; B: solely allothetic) with a drawingreproduction task. For the control group, the mean score is shownby the bar, and error bars indicate the standard error. Bottom:Corresponding Z-scores derived by comparing M.R. to the controlgroup. The dashed line shows T 5 1.796, P < 0.05. * indicates P< 0.05 deficits, * indicates a functional dissociation with P < 0.05.

8 GOMEZ ET AL.

Hippocampus

site way, drawing). This suggests a preservation of allocentricabilities. This hypothesis is further supported by the fact thatMR performs significantly better in all conditions assessingallocentric representation (i.e., when encoding relied on exter-nal movement observation, whatever the reproduction mode)compared to other conditions. This result is crucial, as it ques-tions the usual interpretation of spatial memory deficitsobserved in hippocampal lesions studies: M.R. is able to buildan allocentric representation of the environment.

At first sight, this lack of deficit in allothetic informationprocessing may appear inconsistent with previous interpreta-tions of hippocampal implication in spatial memory. As previ-ously mentioned, human lesion studies have reported allocen-tric deficits following hippocampal damage (Holdstock et al.,2000; Astur et al., 2002; King et al., 2002; Goodrich-Hunsakeret al., 2010) or parahippocampal ones (Bohbot et al., 1998,2000). However, the use of an allocentric representation wasgenerally assessed through object-location memory tasks, aftereither a perspective switch or a change in start location. Asacknowledged by the authors (King et al., 2004), failure inthese tasks can also be due to a deficit in egocentric-updatingby self motion cues. In contrast, in the current study, the ex-perimental procedure assesses allocentric representation process-ing without requiring idiothetic information processing or pathintegration and egocentric-updating utilization. To our mind,these results are in agreement with previous findings, and thealternative hypothesis of egocentric-updating deficits in amnesicpatients, proposed by the authors, seems to be a more suitableinterpretation of their results.

However, one might argue that the preservation (or not) ofallocentric representation could be just a matter of hippocam-pal atrophy (or lesion) extent. If so, despite the fact that M.R.suffers from a large atrophy (more than 60% bilaterally), theremaining neurons, if not dysfunctional, could allow him tobuild an efficient allocentric representation. In other words, anallocentric disruption would require a complete removal of hip-pocampal cells, whereas even an incomplete hippocampal lesionwould be likely to lead to an egocentric-updating disruption.

Moreover, as often argued (Teng and Squire, 1999), the defi-cit observed might also arise from associated atrophy, and thuscould be the result from a dysfunction of a larger neuronal net-work including parietal areas and the right medio-temporallobe (i.e., parahippocampal areas). Parietal atrophy is frequently[35% of patients, Caine and Watson (2000)] encountered aftercarbon monoxide poisoning or after epileptical surgery as forpatient H.M (considered as a reference in amnesia). In fact, avolumetric analysis of HM’s whole-brain performed 50 yearsafter surgery also revealed a parietal atrophy (Salat et al.,2006). In both cases (HM and MR), this parietal atrophy wasobserved after several years and could reflect a specific func-tional link between hippocampal and parietal structures. Suchhypotheses however would require a longitudinal volumetricanalysis at the time of hippocampal lesion onset. Unfortunately,due to the novelty of the MRI volumetric method, these dataare not available for the patient M.R. From available results, itis however clear that the human parietal lobe is extensively

involved in spatial processing, especially in egocentric referenc-ing [for a review, Nitz (2009)] and egocentric navigation (e.g.,Ciaramelli et al., 2010). In rodents, the posterior parietal cortex(PPC) is thought to play a role in mnemonic processing of idi-othetic spatial representations (based primarily on head direc-tion information); but no study has demonstrated a soleinvolvement of the PPC for purely idiothetic process inhumans (Kesner, 2009). Moreover, the retrosplenial cortex hasbeen involved in head-direction processing (Cho and Sharp,2001; Maguire, 2001; Epstein et al., 2007; Vann et al., 2009)as well as idiotethic information and path integration process-ing (Redish and Touretzky, 1997; Wiener and Taube, 2005;Wolbers et al., 2007). It could also be involved in the transfor-mation of egocentric-parietal representation in allocentric onesand vice versa (Byrne et al., 2007). Thus, if the parietal areasthinnings contribute to M.R.’s deficits, we suggest that it mightbe through a disruption of the cooperation with atrophicmedio-temporal structure, even if a spatial role of the parietalareas alone (especially the retrosplenial areas) cannot be clearlyruled out. Therefore, considering spatial memory theories, theseprevious interpretations must be kept in mind, and clear-cutconclusions must await further similar experimentation on purebi-hippocampal amnesia. Concerning the right medio-temporallobe volume reduction, previous studies have clearly pointed itsinvolvement in spatial memory (Aguirre et al., 1998; Bohbotet al., 1998, 2000; Bohbot and Corkin, 2007), and the deficitscould arise from this reduction. However, previous resultsrather suggested that this region could be particularly involvedin processing surrounding scenes geometric features and rele-vant landmarks (Maguire et al., 1998; Epstein et al., 1999;Kohler et al., 2002; Janzen and Van Turennout, 2004). There-fore, even if its involvement in the egocentric-updating deficitcannot be excluded, one would have rather expected this reduc-tion to impair allotethic information processing and allocentrictasks (which are preserved in patient M.R.).

To summarize the spatial data in M.R., we clearly showedthat this patient performs as well as control participants in tasksrequiring allocentric processing, but is impaired in egocentric-updating processing. Considering spatial memory theories, fur-ther experimentations will be necessary to provide better under-standing of the relationship between egocentric-updating deficitand the hippocampo-parietal network. We will now turn to-ward the direct relevance of these results for functional theoriesof long-term episodic memory allocating a specific role to spa-tial processing.

Toward Revised Theories of Spatial ProcessingUnderlying Episodic Memory

Looking at M.R.’s long-term memory performance, as previ-ously reported (Bastin et al., 2004), he suffers from heavy defi-cits in recollection processes and episodic memory, contrastingwith relatively preserved recognition performance. If the above-suggested interpretation of M.R.’s pattern of impairment/pres-ervation remains consistent with the spatial CMT (O’Keefeand Nadel, 1978) by assuming that the functional dissociation

SPATIAL DEFICITS IN AN AMNESIC CASE WITH HIPPOCAMPAL DAMAGE 9

Hippocampus

observed is due to the atrophy extent, it clearly challenges thefirst instance of the MTT (Nadel and Moscovitch, 1998). Thislatter suggests a functional relationship between nonspatial (epi-sodic) and spatial information, based on their common neuro-nal origin: the hippocampus. This theory predicts that the spa-tial deficit likely to disrupt episodic memory is allocentric innature and that it relies on the hippocampal structure. Contro-versially, we observed that the spatial deficit associated to an ep-isodic memory deficit is not allocentric but egocentric-updat-ing, and it relies on idiothetic information.

The new instantiation of the MTT (Nadel et al., 2000; Mos-covitch et al., 2005) also fails to account for our results. Thistheory assumes a functional relationship between nonspatialand spatial information, but it further distinguishes betweenrecently and remotely acquired information. Recently, acquiredinformation is assumed to be linked with detailed spatial mem-ory, whereas remote memories and semantic information wouldrely on allocentric-schematic spatial memory.

However, this theory does not clearly state which brain areamight underlie recently acquired allocentric information, astested in our study. It only states that the hippocampal struc-ture is involved in recently acquired, detailed spatial informa-tion; it is not involved in remotely acquired, schematic-allocen-tric spatial information. Hence, this model does not allowclear-cut predictions for recently acquired allocentric spatial in-formation following hippocampal lesions. Based on the recent-remote distinction, the model would probably predict a deficitin all tasks involving recently acquired spatial memory, with apreservation of remotely acquired spatial information amongpatients with bi-hippocampal lesions (Teng and Squire, 1999).In this study, we did not show that M.R. is impaired in alltasks involving recently acquired spatial memory. We observedinstead a functional dissociation within the category of recentlyacquired information, which is not predicted by the newinstantiation of the MTT. Furthermore, it would be interestingto test if this dissociation could also be found within the cate-gory of remotely acquired information—as Teng and Squireemphasized a preservation of allocentric representation, butthey did not assess the egocentric-updating representation.

Yet, our results can be better understood in the frameworkof Burgess and colleague’s model of episodic memory (Burgesset al., 2001; Byrne et al., 2007). They have hypothesized thatthe retrosplenial structure is crucial to provide the spatial trans-formation of an event from multiple egocentric–iconic repre-sentations into a single allocentric representation for long-termmemory storage. Hence, this proposal assumes that the retro-splenial and posterior parietal structure could be relevant toperform a transformation from the parietal egocentric–iconicrepresentation to the medio-temporal and hippocampal allocen-tric representation. As the idiothetic information and the ego-centric-updating representation are at the root of this referentialtranslation system, we assume that the core of episodic memo-ries is grounded in this translation system (Gomez et al.,2009). In fact, during retrieval, the feeling of reliving a specificevent would rely on the capacity of the system to experience afluency in the translational procedure, leading to an iconic–ego-

centric representation, or in other words, in a specific perspec-tive. The recollection process would end on a specific perspec-tive and would provide a recollection feeling when this proce-dural translation fluency is experienced. This autonoeticconsciousness would distinguish real events (events whichoccurred in space and time) from imagined events (events thatmight be built from an allocentric representation). This inter-pretation predicts the pattern of deficit observed in patientM.R. It clearly accounts for a deficit in egocentric-updatingwith a preservation of the allocentric representation when it isbuilt directly on allothetic information. To summarize, we sug-gest that M.R.’s results can be easily accounted by the Burgess’and colleagues theory by allocating some memory properties tothe translational system.

To conclude, results of this case study are consistent withthose of previous studies reporting spatial memory deficits fol-lowing hippocampal amnesia. However, the performance’s dis-sociation between allocentric tasks and egocentric-updatingtasks does not agree with previous interpretations of those defi-cits. Actually, it suggests that previous reports of cognitive map-ping deficits might reflect a prior deficit in egocentric-updating.Allocentric processing is preserved in M.R. as long as it is onlydriven by allothetic information and does not require the inte-gration of idiothetic information. Most importantly, accordingto episodic memory theories, this study indicates that the spa-tial deficit associated with a pure episodic memory deficit couldbe linked to an egocentric-updating deficit rather than to ageneral allocentric deficit. Although further studies are neededto replicate this result, this case study strikingly questions theMTT (Moscovitch and Nadel, 1998; Nadel and Moscovitch,1998; Nadel et al., 2000; Moscovitch et al., 2005), but theresults fit easily in an adaptation of the Burgess and colleagues’model.

Acknowledgments

We thank all participants, in particular Mr. M.R, for theirtime and effort. The authors also thank Yaelle Spirli and Ma-rine Rigal for their precious experimental and neuropsychologi-cal contribution to this project. We are very grateful to Dr.Olivier Moreaud for allowing access to the NeuropsychologyUnit facilities at the University of Grenoble, to Dr. ChristineBastin for providing M.R.’s MRI scan. We thank Ami Sugiartoand Olivier Casey for their comments on previous versions ofthe manuscript.

REFERENCES

Aguirre GK, Zarahn E, D’Esposito M. 1998. Neural components oftopographical representation. Proc Natl Acad Sci USA 95:839–846.

Astur RS, Taylor LB, Mamelak AN, Philpott L, Sutherland RJ. 2002.Humans with hippocampus damage display severe spatial memoryimpairments in a virtual Morris water task. Behav Brain Res132:77–84.

10 GOMEZ ET AL.

Hippocampus

Avraamides MN, Kelly JW. 2008. Multiple systems of spatial memoryand action. Cognitive Process 9:93–106.

Bastin C, Linden M, Charnallet A, Denby C, Montaldi D, RobertsN, Andrew M. 2004. Dissociation between recall and recognitionmemory performance in an amnesic patient with hippocampaldamage following carbon monoxide poisoning. Neurocase 10:330–344.

Bohbot VD, Corkin S. 2007. Posterior parahippocampal place learn-ing in H.M. Hippocampus 17:863–872.

Bohbot VD, Kalina M, Stepankova K, Spackova N, Petrides M, NadelL. 1998. Spatial memory deficits in patients with lesions to theright hippocampus and to the right parahippocampal cortex. Neu-ropsychologia 36:1217–1238.

Bohbot VD, Allen JJ, Nadel L. 2000. Memory deficits characterizedby patterns of lesions to the hippocampus and parahippocampalcortex. Ann NY Acad Sci 911:355–368.

Bures J, Fenton AA, Kaminsky Y, Wesierska M, Zahalka A. 1998.Rodent navigation after dissociation of the allocentric and idiot-hetic representations of space. Neuropharmacology 37:689–699.

Burgess N. 2002. The hippocampus, space, and viewpoints in episodicmemory. Quart J Exp Psychol 55:1057–1080.

Burgess N, Becker S, King JA, O’Keefe J. 2001. Memory for eventsand their spatial context: Models and experiments. Philos TranscrR Soc 356:1493–1503.

Burgess N, Maguire EA, O’Keefe J. 2002. The human hippocampusand spatial and episodic memory. Neuron 35:625–641.

Burgess N, Trinkler I, King J, Kennedy A, Cipolotti L. 2006.Impaired allocentric spatial memory underlying topographical dis-orientation. Review of Neuroscience 17:239–251.

Byrne P, Becker S, Burgess N. 2007. Remembering the past and imag-ining the future: A neural model of spatial memory and imagery.Psychol Rev 114:340.

Caine D, Watson JDG. 2000. Neuropsychological and neuropatholog-ical sequelae of cerebral anoxia: A critical review. J Int Neuropsy-chol Soc 6:86–99.

Cho J, Sharp PE. 2001. Head direction, place, and movement corre-lates for cells in the rat retrosplenial cortex. Behav Neurosci 115:3.

Chokron S. 2003. Right parietal lesions, unilateral spatial neglect, andthe egocentric frame of reference. Neuroimage 20:75–81.

Ciaramelli E, Rosenbaum RS, Solcz S, Levine B, Moscovitch M.2010. Mental space travel: Damage to posterior parietal cortex pre-vents egocentric navigation and reexperiencing of remote spatialmemories. J Exp Psychol: Learn Mem Cogn 36:619–634.

Crawford JR, Howell DC, Garthwaite PH. 1998. Payne and Jonesrevisited: Estimating the abnormality of test score differences usinga modified paired samples t test. J Clin Exp Neuropsychol 20:898–905.

Crawford JR, Garthwaite PH, Howell DC. 2009. On comparing asingle case with a control sample: An alternative perspective. Neu-ropsychologia 2009; 47:2690–2695.

Damasio AR. 1989. Time-locked multiregional retroactivation: A sys-tems-level proposal for the neural substrates of recall and recogni-tion. Cognition 33:25–62.

Epstein R, Harris A, Stanley D, Kanwisher N. 1999. The parahippo-campal place area: Recognition, navigation, or encoding? Neuron23:115–125.

Epstein RA, Parker WE, Feiler AM. 2007. Where am I now? Distinctroles for parahippocampal and retrosplenial cortices in place recog-nition. J Neurosci 27:6141–6149.

Etienne AS, Jeffery KJ. 2004. Path integration in mammals. Hippo-campus 14:180–192.

Farrell MJ, Robertson IH. 1998. Mental rotation and the automaticupdating of body-centered spatial relationships. J Exp Psychol:Learn Mem Cogn 24:227–233.

Gomez A, Rousset S, Baciu M. 2009. Egocentric-updating during nav-igation facilitates episodic memory retrieval. Acta Psychol (Amst)132:221–227.

Goodrich-Hunsaker NJ, Livingstone SA, Skelton RW, Hopkins RO.2010. Spatial deficits in a virtual water maze in amnesic partici-pants with hippocampal damage. Hippocampus 20:481–491.

Grober E, Buschke H. 1986. Genuine memory deficits in dementia.Dev Neuropsychol 3:13–36.

Holdstock JS, Mayes AR, Cezayirli E, Isaac CL, Aggleton JP, RobertsN. 2000. A comparison of egocentric and allocentric spatial mem-ory in a patient with selective hippocampal damage. Neuropsycho-logia 38:410–425.

Hort J, Laczo J, Vyhnalek M, Bojar M, Bures J, Vlcek K. 2007. Spa-tial navigation deficit in amnestic mild cognitive impairment. Pro-ceedings of the National Academy of Sciences of the United Statesof America 104:4042–4047.

Janzen G, Van Turennout M. 2004. Selective neural representation ofobjects relevant for navigation. Nat Neurosci 7:673–677.

Kesner RP. 2009. The posterior parietal cortex and long-term memoryrepresentation of spatial information. Neurobiol Learn Mem91:197–206.

King JA, Burgess N, Hartley T, Vargha Khadem F, O’Keefe J. 2002.Human hippocampus and viewpoint dependence in spatial mem-ory. Hippocampus 12:811–820.

King JA, Trinkler I, Hartley T, Vargha-Khadem F, Burgess N. 2004.The hippocampal role in spatial memory and the familiarity–recol-lection distinction: A case study. Neuropsychology 18:405–417.

Kohler S, Crane J, Milner B. 2002. Differential contributions of theparahippocampal place area and the anterior hippocampus tohuman memory for scenes. Hippocampus 12:718–723.

Lavenex PB, Lavenex P. 2010. Spatial relational learning and memoryabilities do not differ between men and women in a real-world,open-field environment. Behav Brain Res 207:125–137.

Lavenex PB, Lecci S, Pretre V, Brandner C, Mazza C, Pasquier J, LavenexP. 2011. As the world turns: Short-term human spatial memory inegocentric and allocentric coordinates. Behav Brain Res 219:132–141.

Lipton PA, Eichenbaum H. 2008. Complementary roles of hippocam-pus and medial entorhinal cortex in episodic memory. Neural Plast2008:258467.

Livingstone SA, Skelton RW. 2007. Virtual environment navigationtasks and the assessment of cognitive deficits in individuals withbrain injury. Behav Brain Res 185:21–31.

Maguire EA. 2001. The retrosplenial contribution to human naviga-tion: A review of lesion and neuroimaging findings. Scand J Psy-chol 42:225–238.

Maguire EA, Burke T, Phillips J, Staunton H. 1996a. Topographicaldisorientation following unilateral temporal lobe lesions in humans.Neuropsychologia 34:993–1001.

Maguire EA, Frackowiak RS, Frith CD. 1996b. Learning to find yourway: A role for the human hippocampal formation. Proc Biol Sci263:1745–1750.

Maguire EA, Frith CD, Burgess N, Donnett JG, O’Keefe J. 1998. Know-ing where things are: Parahippocampal involvement in encoding objectlocations in virtual large-scale space. J Cogn Neurosci 10:61–76.

McNaughton BL, Chen LL, Markus EJ. 1991. Dead reckoning land-mark learning, and the sense of direction: A neurophysiologicaland computational hypothesis. J Cogn Neurosci 3:190–202.

McNaughton BL, Battaglia FP, Jensen O, Moser EI, Moser MB. 2006.Path integration and the neural basis of the ‘cognitive map’. NatRev Neurosci 7:663–678.

Mellet E, Briscogne S, Tzourio-Mazoyer N, Ghaem O, Petit L, ZagoL, Etard O, Berthoz A, Mazoyer B, Denis M. 2000. Neural corre-lates of topographic mental exploration: The impact of route versussurvey perspective learning. Neuroimage 12:588–600.

Mendez MF, Cherrier MM. 2003. Agnosia for scenes in topographag-nosia. Neuropsychologia 41:1387–1395.

Mittelstaedt ML, Mittelstaedt H. 1980. Homing by path integrationin a mammal. Naturwissenschaften 67:566–567.

Morris RGM, Garrud P, Rawlins JNP, O’Keefe J. 1982. Place naviga-tion in rats with hippocampal lesions. Nature 297:681–683.

SPATIAL DEFICITS IN AN AMNESIC CASE WITH HIPPOCAMPAL DAMAGE 11

Hippocampus

Moscovitch M, Nadel L. 1998. Consolidation and the hippocampalcomplex revisited: In defense of the multiple-trace model. CurrOpin Neurobiol 8:297–300.

Moscovitch M, Rosenbaum RS, Gilboa A, Addis DR, Westmacott R,Grady C, McAndrews MP, Levine B, Black S, Winocur G, et al.2005. Functional neuroanatomy of remote episodic, semantic andspatial memory: A unified account based on multiple trace theory.J Anat 207:35–66.

Nadel L, Moscovitch M. 1998. Hippocampal contributions to corticalplasticity. Neuropharmacology 37:431–439.

Nadel L, Samsonovich A, Ryan L, Moscovitch M. 2000. Multipletrace theory of human memory: Computational, neuroimaging,and neuropsychological results. Hippocampus 10:352–368.

Nitz D. 2009. Parietal cortex, navigation, and the construction of arbi-trary reference frames for spatial information. Neurobiol LearnMem 91:179–185.

O’Keefe J, Dostrovsky J. 1971. The hippocampus as a spatial map.Preliminary evidence from unit activity in the freely-moving rat.Brain Res 34:171–175.

O’Keefe J, Nadel L. 1978. The Hippocampus as a Cognitive Map.London: Oxford University Press.

Philbeck JW, Behrmann M, Levy L, Potolicchio SJ, Caputy AJ. 2004.Path integration deficits during linear locomotion after humanmedial temporal lobectomy. J Cogn Neurosci 16:510–520.

Redish AD, Touretzky DS. 1997. Cognitive maps beyond the hippo-campus. Hippocampus 7:15–35.

Salat DH, van der Kouwe AJW, Tuch DS, Quinn BT, Fischl B, DaleAM, Corkin S. 2006. Neuroimaging HM: A 10-year follow-up ex-amination. Hippocampus 16:936–945.

Scoville WB, Milner B. 1957. Loss of recent memory after bilateralhippocampal lesions. J Neurol Neurosurg Psychiatry 20:11–21.

Shepard RN, Metzler J. 1971. Mental rotation of three-dimensionalobjects. Science 171:701–703.

Skelton RW, Ross SP, Nerad L, Livingstone SA. 2006. Human spatialnavigation deficits after traumatic brain injury shown in the arenamaze, a virtual Morris water maze. Brain Injury 20:189–203.

Spiers HJ, Maguire EA, Burgess N. 2001. Hippocampal amnesia.Neurocase 7:357–382.

Squire LR, Zola-Morgan S. 1991. The medial temporal lobe memorysystem. Science 253:1380–1386.

Stroop JR. 1935. Studies of interference in serial verbal reactions. JExp Psychol 6:661.

Teng E, Squire LR. 1999. Memory for places learned long ago isintact after hippocampal damage. Nature 400:675–677.

Teyler TJ, DiScenna P. 1986. The hippocampal memory indexingtheory. Behav Neurosci 100:147–154.

Teyler TJ, Rudy JW. 2007. The hippocampal indexing theory and epi-sodic memory: Updating the index. Hippocampus 17:1158–1169.

Vann SD, Aggleton JP, Maguire EA. 2009. What does the retrosplenialcortex do? Nat Rev Neurosci 10:792–802.

Waller D, Lippa Y, Richardson A. 2008. Isolating observer-based refer-ence directions in human spatial memory: head, body, and the self-to-array axis. Cognition 106:157–183.

Warrington EK, James M. 1991. Visual Object and Space PerceptionBattery. England: Thames Valley Test Company.

Whishaw IQ, Hines DJ, Wallace DG. 2001. Dead reckoning (pathintegration) requires the hippocampal formation: Evidence fromspontaneous exploration and spatial learning tasks in light (allo-thetic) and dark (idiothetic) tests. Behav Brain Res 127:49–69.

Wiener SI, Taube JS. 2005. Head Direction Cells and the NeuralMechanisms of Spatial Orientation. Cambridge, MA: MIT Press.

Wishaw IQ, Cassel JC, Jarrard LE. 1995. Rats with fimbria-fornixlesions display a place response in a swimming pool: A dissociationbetween getting there and knowing where. J Neurosci 15:5779–5788.

Wolbers T, Wiener JM, Mallot HA, Buchel C. 2007. Differentialrecruitment of the hippocampus, medial prefrontal cortex, and thehuman motion complex during path integration in humans. J Neu-rosci 27:9408–9416.

Worsley CL, Recce M, Spiers HJ, Marley J, Polkey CE, Morris RG.2001. Path integration following temporal lobectomy in humans.Neuropsychologia 39:452–464.

12 GOMEZ ET AL.

Hippocampus