Non-human Primates in Neuroscience Research: The Case … · project applications that we have seen. In addition, public information proffered by those that use NHPs in this way amounts

Introduction

Experiments on non-human primates (NHPs) arehighly controversial. Opinion polls consistentlyshow that members of the public are particularlytroubled by the practice (1–4), with many only sup-porting it with important caveats, or indeed oppos-ing it outright. A recent MORI poll in the UK foundthat only 16% of respondents supported the use ofmacaque monkeys for medical research to benefitpeople (5). This is not surprising, since NHP exper-iments in neuroscience, for example, involve manyinvasive and stressful methods, including intracra-nial electrodes, various restraint and trainingtechniques, and water deprivation (6). None -theless, tens of thousands of NHPs continue to beused each year in experiments in the EU and theUSA. According to EU statistics from 2011, themajority of NHPs were used in drug testing, but asignificant number, i.e. 631 (10% of the total), wereused in fundamental biological research, a largeproportion of which would encompass basic neuro-logical investigations, such as those discussed inthis paper (7). Unfortunately, more-exact figuresare not available.

The justification provided by scientists who useNHPs, in neuroscience research as well as in otherareas, centres around a harm–benefit analysis

(now inherent in the EU Directive on the use ofanimals in science, Directive 2010/63/EU; 8), i.e.that harms to the NHPs used are mitigated by ben-efits to humans. However, the benefit to humanhealth of NHP-based neuroscience researchappears to be considered only superficially in theproject applications that we have seen. In addition,public information proffered by those that useNHPs in this way amounts to simple statementsabout the severity of the human disease in ques-tion and the need to use monkeys. The actualextent to which NHP neuroscience research bene-fits human health has not been considered in anydepth or truly independently.

Many publications offer generalisations withregard to the indispensability of NHP research,based on anatomical, physiological, genetic, func-tional and behavioural similarities, which areoften superficial and unreferenced (e.g. 9–11).Claims can be as vague as: “Nonhuman primateshave a unique position in biomedical researchrelated to their close phylogenetic proximity tohumans. This close proximity often serves as thebasis for scientific justification of their use inresearch” (9); and, with regard to neuroscience,“Neuroscience is an area in which research withnon-human primates has played a major part inour understanding of basic neurobiology and thecauses and potential treatments for human disor-

Non-human Primates in Neuroscience Research: The CaseAgainst its Scientific Necessity

Jarrod Bailey and Katy Taylor

Cruelty Free International, London, UK

Summary — Public opposition to non-human primate (NHP) experiments is significant, yet those whodefend them cite minimal harm to NHPs and substantial human benefit. Here we review these claims ofbenefit, specifically in neuroscience, and show that: a) there is a default assumption of their human rele-vance and benefit, rather than robust evidence; b) their human relevance and essential contribution andnecessity are wholly overstated; c) the contribution and capacity of non-animal investigative methods aregreatly understated; and d) confounding issues, such as species differences and the effects of stress andanaesthesia, are usually overlooked. This is the case in NHP research generally, but here we specifically focuson the development and interpretation of functional magnetic resonance imaging (fMRI), deep brain stim-ulation (DBS), the understanding of neural oscillations and memory, and investigation of the neural con-trol of movement and of vision/binocular rivalry. The increasing power of human-specific methods,including advances in fMRI and invasive techniques such as electrocorticography and single-unit record-ings, is discussed. These methods serve to render NHP approaches redundant. We conclude that thedefence of NHP use is groundless, and that neuroscience would be more relevant and successful forhumans, if it were conducted with a direct human focus. We have confidence in opposing NHP neuro-science, both on scientific as well as on ethical grounds.

Key words: brain, electrophysiology, fMRI, monkey, neurology, neuroscience, primate.

Address for correspondence: Jarrod Bailey, Cruelty Free International, 16a Crane Grove, London N78NN, UK.E-mail: [email protected]

ATLA 44, 43–69, 2016 43

ders. This animal model is especially valuablebecause of the many similarities between humanand non-human primates that derive from theircommon ancestry, such as complex cognitive capa-bilities, great social complexity, details of repro-ductive biology, and intricacy of brainorganisation… In neuroscience, non-human pri-mates continue to have important roles in basicand translational research, owing to their behav-ioural and biological similarity to human beings”(12), and the model has a “far-reaching relevancethat is irreplaceable for essential insights into cog-nitive functions, brain disease, and therapy” (13).Many of these general claims are augmented byissues that are not relevant to, and have no placein, the NHP research debate, and which should bedismissed. For example, it is frequently stated thatanimal research as a whole uses a fraction of theanimals that are used and killed by humans, andthat, of those used in research, just a fraction of 1%are NHPs (e.g. 13).

This pro-NHP experimentation ‘canon’ is sup-ported by inquiries into NHP research conductedover the past decade, the findings and conclusionsof which have been influenced by such expressionsof opinion from NHP researchers. For instance, the2006 report commonly known as the WeatherallReport (14), and the consequent Bateson Review(15), both concluded, broadly, in favour of the needfor NHP experimentation. There are, however,important and serious caveats, in addition to well-founded concerns about the review processes them-selves and, accordingly, about their conclusions.These include: their potential lack of objectivity(i.e. they were funded and overseen by organisa-tions and individuals who support NHP research);the lack of genuine harm–benefit analyses and sys-tematic reviews; and questionable processes andcriteria for assessing the value of the research.Therefore, it is particularly notable that they advo-cated a greater focus on the development of alter-natives, improved animal welfare and systematicreview of NHP research. In addition, theyexpressed concerns over the following issues: wel-fare costs; the application and relevance of theresearch to humans; the overstating of medicalbenefits by researchers; that benefits were specu-lative and not commensurate with welfare costs;that, often, little or no evidence of actual medicalbenefit is available; and that NHP work sometimesrepeats previous work and/or ‘confirmed’ priorhuman studies (16). Of particular importance inrelation to neuroscience, Recommendation 8 of theBateson Review stated that, “Highly invasive andlong-term NHP research often carries a high wel-fare cost. In such cases, funders should take par-ticular care only to fund projects with a very highlikelihood of producing scientific, medical or socialbenefit. Wherever possible, funders should takesteps towards encouraging a preferential or com-

plementary use of less invasive techniques such asneuroimaging and transcranial magnetic stimula-tion”. Section 4.2.5 notes that almost half (46.2%)of the reviewed studies were neuroscientific, andthat half of these had “a high welfare impact on theanimals”, while “In most cases, however, littledirect evidence was available of actual medicalbenefit in the form of changes in clinical practice ornew treatments”. Additionally, the “effectivenessof knowledge transfer from basic to appliedresearch” — something frequently claimed to bebeyond doubt and, indeed, “the only way to suc-cessfully cope with devastating disorders andreduce the invasive character of many clinical pro-cedures applied for their diagnosis and treatment”(17) — has been questioned by many (e.g. 18). Thisis acknowledged by Recommendation 4 of theBateson Review, which suggests developing amechanism “…to identify research results withpotential to deliver improvements to healthcare orother significant benefits to society, and to assessthe extent to which the potential benefits areachieved”. Unfortunately, three years later, at thetime of writing, the Medical Research Council(MRC) has indicated that this recommendation “isnot currently being taken forward, as we have pri-oritised work on other recommendations in theBateson Review” (18). Factual information on NHPresearch results and their medical relevance (asopposed to claims of efficacy and human relevanceby NHP researchers) therefore remains unavail-able to the public and the wider medical profes-sion. Overall, few, if any, of the recommendationsfrom the above inquiries and reports have beenupheld and enforced.

This review examines the important question ofthe degree (if any) to which NHP neuroscience ben-efits humans, firstly by reviewing the scope, capac-ity and potential of ‘alternatives’ to the use ofNHPs, such as human research and neuroimaging,asking whether the resultant benefit to humansfrom these approaches alone is not sufficient and,therefore, whether NHP experiments are neededat all; and secondly, by critically assessing salientclaimed human benefits of specific NHP neuro-science studies, asking if there is substance to suchclaims, and whether any such benefits absolutelyrelied on the use of NHPs. Much of the latter is inlarge part in response to statements made recentlyby, for example, the Max Planck Institute forBiological Cybernetics (Germany) and NewcastleUniversity (UK), in support of NHP neuroscience.

Overall, we examine, inter alia: historical claimsof the necessity of NHP use and its contribution tohuman medical progress, both in general termsand with regard to specific examples frequentlyproffered by its advocates; claims of future indis-pensability based on the perceived limitations ofalternative approaches; the nature of inquiries intothe efficacy of NHP research in recent years, and if

44 J. Bailey & K. Taylor

they are fit for purpose; the degree of contributionof other neuroscientific investigative methods tothe field, and how they compare to NHP use; andthe extent to which confounding factors affect NHPdata and their extrapolation to humans. In short,we seek to establish what human neuroscience cando and, as the scientific method demands, we askquestions of NHP neuroscience, testing the hypoth-esis that it is necessary and relevant to humans,rather than seeking to justify it.

The Increasing Power of HumanNeuroscientific InvestigativeMethods

Human-based methods must be the goal, from botha scientific and an ethical perspective. There fol-lows a detailed consideration and description ofthese, in order to gauge how capable they are, andto confront allegations that these methods are infe-rior and less able to tackle important researchquestions that must be answered, in order forprogress to be made in the diagnosis and treat-ment of neurological diseases. These examplesrelate to particular investigative techniques and toareas of neuroscience that are particularly com-mon, such as vision (categorisation, face and colourrecognition and processing), memory function andvision-associated memory function, neuronal con-trol of movement, and analysis of neural oscilla-tions of various frequencies and synchrony. Someof these human-focused methods are also describedin the subsequent section of this review, as part ofindividual cases against specific claims of NHPresearch necessity.

Transcranial magnetic stimulation (TMS)

TMS remains, in its fourth decade of use, anextremely useful and productive technique. Itallows the non-invasive stimulation of discretebrain areas via magnetic pulses through the scalp,activating the cortex over a focused area of arounda few square centimetres.

A 2010 review of TMS outlined its historical andcurrent importance in neuroscience research (19).Originally developed to investigate the propaga-tion of neural signals along the corticospinal tract,spinal roots, and peripheral nerves in humans, itowes its development to human studies. Single-pulse, paired-pulse and repetitive TMS “allowsroutine evaluations of the excitability and conduc-tivity of corticospinal motor pathways” to investi-gate movement physiology in healthy patients andthose with neurological disorders; allowsresearchers to transiently interfere with behaviourin various domains (i.e. create ‘virtual lesions’), toenhance understanding of cortical functions; per-

mits the mapping of motor cortical outputs and thestudy of motor conduction; provides measures ofintracortical facilitation, inhibition and cortico-cor-tical interactions; allows the study of corticalexcitability in neurological diseases, the functionalrelevance of cortical areas in cognitive task per-formance, brain–behaviour relations, and thepathophysiology of various neurological andpsychiatric disorders.

Much information about the mechanisms of TMS(and indeed transcranial electrical stimulation[TES]) — for example, which neurons are specifi-cally activated by it — has been gleaned via humanexperimentation over the past 30 years. Studiesinvolving anaesthetised patients, as well as con-scious volunteers, have elucidated how waves ofneural activity travel down the corticospinal tract,measured by epidural electrodes implanted in thespinal cord (see Di Lazzaro & Ziemann [20]).

fMRI

fMRI is a method of measuring brain activity, inwhich sophisticated scanning machines and com-puters detect changes in blood oxygenation andflow that reflect neural activity. Human fMRIinvestigations are revealing how areas of the braininteract to transform particular sensory informa-tion into specific motor outputs, to achieve goal-directed movements (21). Resting state functionalconnectivity (rs-fc) fMRI has successfully identifiedaltered intrinsic neural networks in many neuro-logical and psychiatric disorders, includingAlzheimer’s disease and schizophrenia, as well asother disorders, such as tinnitus (22). It has alsobeen of use in the identification of neuroimagingbiomarkers of various conditions and disorders.With regard to tinnitus, for example, rs-fc fMRIanalysis has implicated several inherent neuralnetworks. In patients, the neural networksaffected seem to show consistent modifications,including differential connectivity between limbicareas and cortical networks, and brain regionsinvolved in attention and auditory processing.

fMRI and other neuroimaging technologies arealso improving dramatically with time, increasingtheir power and resolution. It is acknowledged that“Advances in the new-generation of ultra-high-res-olution, brain-dedicated PET/MRI systems havebegun to provide many interesting insights intothe molecular dynamics of the brain” (23). Suchultra-high field MRI is, at present, around fivetimes more powerful than some systems that areused routinely. When combined with brain-dedi-cated high-resolution research tomograph (HRRT)positron emission tomography (PET), which is anorder of magnitude more sensitive than a whole-body system, the visualisation of many of the finestructures of the human brain is possible, includ-

Non-human primates in neuroscience research 45

ing the hippocampus, thalamus and the brainstem.Investigation of the neural and functional activityof these regions, such as during memory tasks, isalso possible (see Cho et al. [23]).

Electrocorticography (ECoG)/intracranialelectroencephalography (EEG) and magnetoencephalography (MEG)

ECoG is an invasive technique involving arrays ofelectrodes implanted subdurally on the brain’s sur-face. Clinical monitoring to identify epileptic fociinvolves craniotomy, placement of electrode stripson the surface of the cortex (often beneath thedura), as determined by prior EEG and MRI, thenstimulation and recording over several days, fol-lowed by explantation and therapeutic resection.This procedure often provides an opportunity forresearch allied to therapeutic use (24).

ECoG permits the acquisition of brain signalsthat have “an exceptionally high signal-to-noiseratio, less susceptibility to artefacts than EEG,and a high spatial and temporal resolution (i.e. < 1cm/< 1 millisecond, respectively)” (24). Thisincreased resolution makes it superior to non-inva-sive techniques, in that it helps to explore in detailthe very short-lived dynamics of brain processes.ECoG “carries substantial information about task-related activity, such as motor execution and plan-ning, auditory processing and visual-spatialattention”; “can reveal functional connectivity andresolve finer task-related spatial-temporal dynam-ics, thereby advancing our understanding of large-scale cortical processes”; “has especially provenuseful for advancing brain–computer interfacing(BCI) technology for decoding a user’s intentions toenhance or improve communication and control”;and “yields information that clinicians can subse-quently use to guide the process of functional map-ping by electrical stimulation”; see, for example(24).

Human-based invasive research into visual per-ception and selectivity is commonplace, includingthe area of facial recognition. Intracranial localfield potentials (LFPs) were recorded in 14epilepsy patients by using ECoG, while they werepresented with images of faces and other objects.ECoG is regarded as showing “superior selectivity(and hence spatial resolution) compared withfMRI” (25), and, in this study, it permitted record-ing with up to 187 electrodes for up to 10 days.This revealed exemplar selectivity in the face-related cortex — findings that were related to sim-ilar, previous studies with single-unit electrodes inNHPs as well as fMRI in humans.

ECoG has revealed information about functionalselectivity in the human cortex, in response toaudiovisual stimuli. Twelve neurosurgical patientsimplanted with subdural electrodes on the cortical

surface (590 in total) showed different and uniqueelectrode responses, with selectivity to stimuli. Inother words, the human sensory cortex is arranged“as a mosaic of functionally unique sub-regions inwhich each site manifests its own special responseprofile” (26).

Subdural arrays of electrodes were used toinvestigate the neural processing of partial visualinformation in the recognition of objects (27). Thisbuilt on prior human research involving fMRI,EEG and invasive recordings. LFPs were recordedin 18 epileptic patients, each with an average of 94implanted electrodes, when they were presentedwith visual stimuli. Analysis of neuronal activityalong the visual stream revealed that the humanvisual cortex remains selectively active, even whenpresented with as little as 9% of an object. Thesesignals were delayed, however, indicating thatadditional neural processing, involving spatial inte -gration and extrapolation from prior knowledge,was occurring, to enable the reliable recognition ofpartial objects.

The human cerebellum has scarcely beenexplored in neuroscience, mainly due to the majorfocus of neuroscience on the cortex, but also due toits relative inaccessibility by non-invasive methodssuch as EEG and MEG. Most human informationhas come from lesion studies, as well as fMRI andTMS, though some EEG and MEG reports exist.For instance, correlation between resting-statefMRI activity and beta-band activity in MEG hasbeen shown. In addition, there are some historicalreports of intracranial electrode-based (includingelectrocorticographic) investigations of the humancerebellum, which elucidated neuronal oscillationsand perturbations during task performance (28).Thus, invasive investigation of the human cerebel-lum is not only possible, but has been achieved,and has produced very useful data to augmentnon-invasive studies.

Such is the power and importance of ECoG, ithas been argued that “ECoG/icEEG [intracranialEEG] informs unresolved questions in the study ofhuman memory and is yielding insights necessaryfor the development of novel interventions to facil-itate memory function in the damaged brain” (29).While imaging such as fMRI reveals the where ofmemory function, and non-invasive (scalp) EEG/MEG the when of memory and, in some instancesand circumstances, also the how of memory, ECoGreveals the “how of human memory across anextended scope of the neurophysiology of memoryin humans” (29). One example is the spatio-tempo-ral functioning of human ‘subsequent memory’.

The importance of EEG and MEG to brainresearch and functional mapping is exemplified byits central role in multi-national collaborative proj-ects delineating neural networks, known as the‘brainnetome’, such as the Human ConnectomeProject in the USA and the CONNECT Project in


Europe (30). This is due to the “outstanding tem-poral resolution” of these methods, and because“they are the primary clinical techniques used tocapture the dynamics of neuronal connections”.The most advanced EEG systems now have sam-pling rates faster than 1kHz, a level of precisionthat means they are able to record “subtle andswift changes in neuronal activity”, such as epilep-tiform spike waves with durations of less than50ms. The power of EEG and MEG has beenincreased substantially by the development ofmulti-channel systems, which are able to record256 channels simultaneously, permitting theanalysis of the functional integration of the brainacross multiple regions. MEG, which records mag-netic fields produced by electrical activity in thebrain, has a comparable resolution to that of EEG,but has higher localisation accuracy. Together, byrelating the results of these approaches to fMRIdata and psychiatric/behavioural/clinical perform-ance, it is believed that neuroscience and clinicalresearch will benefit (30).

Single unit/microelectrode recordings

Similarly to ECoG above, this often takes place inpatients undergoing neurosurgery for a variety ofreasons, including epilepsy, autism, movementdisorders, with the patients’ consent. Singlemicroelectrodes can be inserted into various areasof the brain, for short-term or even longer termperiods, to permit readings to be taken of single-cell electrical activity during a variety of tasks,including vision, memory and navigation (seeFried et al. [31]).

Single-unit spiking activity was recorded in 14epilepsy patients undergoing surgery, to investi-gate grid-like neuronal activity in human spatialnavigation, in both the entorhinal and cingulatedcortex (32). These cells had been identified previ-ously in human fMRI investigations, as well as inrats, bats and monkeys. Notably, the human grid-like cells seemed to have “noisier firing maps thansome grid cells reported in rodents”.

Single-unit recordings are augmenting extensivefMRI investigations of human episodic memory.Research, directed to the hippocampus and associ-ated brain regions via these human fMRI investi-gations, coupled with clinical findings in patientswith amnesia, confirmed the hypothesis that so-called place cells are involved in the encoding andretrieval of episodic memory (33).

Ten epilepsy patients were implanted with chronicdepth electrodes for up to 10 days, to investigate theperception and recognition of faces. This involvedmeasuring the activity of single neurons in themedial temporal lobe (MTL), and the authors con-cluded that the firing of MTL neurons in humans(including various regions, such as the hippocampus,

amygdala, and the entorhinal and parahippocampalcortices) depends on perceptual decisions (i.e. recog-nising faces), rather than on the actual visual fea-tures of the stimuli (34). Single-neuron and LFPactivity in the MTL have been concurrently meas-ured in humans during visual recognition tasks, todetermine the relationship between those measure-ments (35). Building on previous similar investiga-tions in humans, as well as related human fMRI,and scalp and intracranial EEG experiments, thisresearch revealed that single neuron and LFPresponses in the gamma and theta ranges indicatethe conscious processing and recognition of per-ceived visual stimuli, which were phase-locked.Crucially, in humans, there existed a post-stimuluslatency in neuron firing of around 300ms, which ismuch longer than similar latency in monkeys of just100–200ms. This greater latency in humans isthought to reflect the greater processing of stimulifor memory functions.

Prior human-based research into autism impli-cated the amygdala in autism-associated abnormalprocessing of faces. Autistic patients undergoingneurosurgery facilitated the investigation of singleneuron firing in the amygdala, which revealed nor-mal neural electrophysiology, but underlay abnor-mal neural responses to facial features, comparedto control epileptic patients undergoing neuro-surgery (however, the nature of the control group,being epileptic, may be an important caveat; 36).Recordings were made via 56 electrodes implantedin the two autistic patients, compared to 88 neu-rons in the eight controls.

Combinations of invasive single-unit andcell-assembly recordings

A range of applications and successes of invasivehuman neuroscience were reviewed by Engel et al.(37). This 2005 review illustrates the breadth ofhuman invasive neuroscience ten years ago, and,of course, the power and ability of investigativetechniques have greatly improved in the interven-ing decade. This shows that human neurosciencehas indeed been diverse and flourishing, and notlimited to non-invasive imaging techniques, forsome time. Examples of invasive human neuro-science cited by the Engel et al. review include:— identification and characterisation of the rela-

tionship between single-neuron activity in vari-ous structures of the brain and movement ofbody parts and/or sensory stimulation, such asthe somatotropic organisation of the subthala-mic nucleus (STN);

— discovery of the cellular correlates of tremor;— elucidation of the function of the basal ganglia,

including details of oscillations/synchrony/coherence, and the effects of dopamine agonists;


— investigation of pathophysiological changes inepilepsy;

— study of neural coding and representation, par-ticularly with regard to language;

— examination of learning and many types ofmemory (e.g. declarative, episodic, implicit,recognition, verbal);

— general cortical function and associated neuralprocesses;

— confirmation of results obtained non-invasively;and

— analysis of movement-related synchronisationin different frequency ranges.

Many, if not all, of the above have been investi-gated in humans at varying resolutions, rangingfrom single-unit and multi-unit activity to local ormore distributed cell assemblies. This not onlyinforms neural coding, specificity and tuning, butalso neural topography and maps, and the forma-tion, function and spatiotemporal interactions ofneural cell assemblies. Such investigations are notlimited to brief experimental protocols just becausethey are in humans; many investigations takeplace over days or even weeks.

Electrical brain stimulation

Cortico-Cortical Evoked Potentials (CCEP) consti-tute an investigational approach that permits thein vivo human-specific ‘mapping of brain net-works’, i.e. the study of anatomical and functionalconnectivity between motor regions of the cortex.This has previously involved NHP-based invasiveelectrophysiology, though non-invasive human-based methods — such as diffusion tensor imaging(DTI), TMS, transcranial direct-current stimula-tion (tDCS), PET/fMRI, EEG, MEG and post-mortem human brain dissection — are increasinglyemployed (38, 39). These human-based methodsarguably make the NHP work redundant, particu-larly when one considers the associatedadvantages of species specificity.

Connectivity is tracked by applying electricalimpulses to chronically implanted subdural elec-trodes, for example in epilepsy patients undergo-ing pre-surgical evaluation, and then recordingevoked potentials elicited at distant cortical sites.While these non-invasive techniques are valuableand continue to be informative for mapping thehuman brain, CCEP augments them and providesmore detail, by aiding the resolution of functional-ity and directionality of anatomical links — inother words, while non-invasive imaging estab-lishes connections, CCEP can establish whetherthese connections are actually used and in whichdirection. Further, CCEP helps to reveal the actualneural basis of imaging signals specific to humans,

and facilitates the study of perturbations in brainnetwork function during cognitive processing. Todate, many human brain regions have been exam-ined and mapped, including the fronto-parietalnetwork, hippocampus and language networks,and observations have been made on human spa-tial memory, perseverance, motor braking andvisual perception, among others (see Keller et al.[39]).

A review of direct electrical stimulation (DES) ofthe brain comprehensively summarises how thismethod of neuroscience research has contributedto our understanding of human brain function (40).It has informed the organisation of human brainnetworks associated with movement, language andcognition, as well as basic neuroscience conceptssuch as neural transmission, localisation of brainfunctions and arrangement of many sensorimotorareas. Specific examples include impulses to act,face recognition, detection of motion, and produc-tion of language.

Specific Claims of NHP Researchers

Advocates of NHP neuroscience have, over time,cited various specific areas of research in whichthey regard NHP use as crucial. We criticallyexamine some of the recent, commonplace andmost vociferous here.

Single-neuron studies

One of the most common refrains from NHPresearchers is that experiments on the activity ofsingle neurons in the brain — as opposed, forexample, to studying the inputs and outputs ofgroups of neurons, studying neuronal connections,structure, function and so on — are fundamental toneuroscientific investigation, and that they mustbe conducted in NHPs because they cannot easilybe done in humans, and that NHPs are most likelyto provide data of human relevance (e.g. 14, 15, 41,and many more publications, declarations in vari-ous forums, and personal communications). It isinformative to examine these and related claims:

‘Single-neuron studies of the human brain are difficult, and therefore rare’

Superficially, single-neuron studies in humans areostensibly difficult to conduct due to their highlyinvasive nature, and it seems plausible that NHP-based experiments should be easier to sanctionand perform. However, scrutiny of the neuroscien-tific literature reveals that purportedly rarehuman single-neuron studies are not rare at all.Since the first human single-neuron recording in


1955 (42), through experiments “during cognitivemeasures” that began in the 1970s (see Ojemann[43]), a search of scientific literature databasessuggests that hundreds of such human experi-ments have been performed. This is borne out bythe recent publication of a 376-page book on thesubject (31), describing in detail many human sin-gle-neuron studies across many fields of investiga-tion over this period of 60 years. These studiescontinue to be performed, for example, in epilepsypatients undergoing neurosurgical procedures forthe condition (e.g. 44, 45).

That said, human single-neuron studies couldbe, and should be, more routine than they are.They “provide unique insights into the neuralmechanisms of human cognition”, and this insightcould be achieved with little effort. However, thereare many missed opportunities — for example,during surgery for dyskinesias, in which electrodesare used to locate specific brain areas. The addi-tional studies might not be undertaken, despitethe available opportunities, because there could bepressure on neurosurgeons not to do so — thoughit could be considered “a tragedy for human knowl-edge when an effort is not made to utilise theavailable opportunities” (43).

Encouragingly, any possible historical impedi-ments to human investigations, such as limitedclinical scenarios, relatively poor electrodes andprimitive hardware and software, have been over-come. For instance, there are now flexiblemicrowire bundles and tetrode arrays, permittingchronic implantation for long-term analysis. Suchadvances have led to a “mini-explosion in the fieldof human microelectrode recording”, the power ofwhich has been augmented by pre-amplifiers, headstages, noise reduction circuitry, better spike-sort-ing algorithms, etc. (46).

‘Epileptic brains (normally studied in humanresearch) are different from non-epileptic brains’

It is claimed that the study of human single neu-rons is confounded by the fact that, usually, thebrains being studied are those of epileptic patients.However, it has been argued that any differencesbetween epileptic and non-epileptic brains may begross in nature, and therefore that there is no evi-dence that single neuron activity is affected (43),and also that any potential differences are miti-gated by the recording of activity in tissue awayfrom the epileptogenic focus (i.e. with no epilepti-form activity). Further, investigation is notrestricted to areas of the brain around samplingsites, and the use of other sites (such as the senso-rimotor cortex, for example) is generally acceptedif there is informed consent by the patient, as wellas Institutional Review Board approval (43).Indeed, in any surgery, the epileptogenic focus is

not known a priori, and as such requires electrodeimplantation into several sites to determine itslocation, many of which will be unrelated to theseizure network/focus (47). In addition, data areevaluated in the context of data from other studies,such as fMRI of non-epileptic brains, and there areno confounding effects from general anaesthesia,as there tend to be in NHP studies (see below; 47).

‘The scope of human invasive research is limited’

Human single-neuron studies have, in fact, beengreatly informative and powerful in many and var-ied ways, and their scope continues to increase.Examples include:— Declarative memory: This type of memory

enables the rapid transformation of experiences(‘episodes’) into long-term memories that aresubsequently accessed or ‘declared’ by free recallor familiarity. The MTL has been an importantarea of research (hippocampus and surroundingstructures) in this regard, which is central to atype of declarative memory known as episodicmemory (memories of the details of one’s ownpersonal experiences), as well as spatial mem-ory (memories of locations, spatial relationships,navigation, etc.). Notably, the former is thoughtto be almost exclusive to humans, though even ifit were not, its study in non-humans would beclose to impossible, as it chiefly relies on verbalreporting. This is why it has been asserted thatsingle-neuron recordings in humans are“uniquely positioned to contribute toward ourunderstanding of the neural mechanisms ofMTL-dependent memories above and beyondwhat can be learned from animal models” (48).Spatial memory is being investigated in humansby using combined techniques, such as single-neuron, LFP and fMRI approaches. The imprac-ticality of making humans under study navigatemazes (used in the past as a reason to use ani-mals) has been overcome by simply askinghuman subjects to navigate a computer pro-gram. Human studies also involve the use ofwireless electrodes to provide a more completepicture of the neural basis of spatial memoryand navigation, by assessing vestibular and pro-prioceptive inputs (49).

— Sleep: Human studies of slow waves, their syn-chrony and their underlying activity, are con-sidered invaluable for elucidating the linkbetween sleep and cognition, as the simultane-ous recording of activity from multiple brainareas bilaterally, and sampling of activityacross cortical and sub-cortical structures, israrely achieved in animal studies (50).

— Visual cognition: Interest in human-focusedresearch is increasing, due to improved tech-


nologies, such as diffusion tensor imaging,building on historical human contributions tothe field from lesion studies — in, for example,epilepsy patients who have undergone resectivesurgery. Much focus has been on regions of theMTL such as the parahippocampal gyrus, butalso the inferior temporal cortex (ITC) (51).

— Thoughts and deliberations: Much mentalactivity is outside the realm of standard ‘stimu-lus-response’ that is frequently central to neuro-scientific investigation. Thought processes notrelated to direct external input, such as imagery,free recall, deliberations and so on, are difficult,if not impossible, to study in animals, as theyinvolve verbal reporting and/or responses toinstructions. In these cases, human study is notmerely desirable, it is necessary. While EEG,PET, MEG and fMRI have all aided the infer-ence of thoughts from patterns of neural activity,greater clarity has been provided by single-neu-ron studies. Studies involving the use ofintracranial depth electrodes implanted in theMTL have determined correlates of internalvisualisation, and that neurons selective forobjects, animals, etc., are selective for both visu-alisation and actual vision. These studies havealso involved the investigation of recollection(not possible in animals), and have shown thatthe firing rate of specific neurons can indicatethe type of object being visualised/observed (52).The latter has implications for brain–machineinterfaces (BMIs; control of external devices, e.g.robotic arms) via patterns of neural activitydetected in real time.

— Reward processing; investigation of deep brainstructures: Human single-neuron studies havenot been limited to superficial parts of thebrain. Studies of deep structures (such as thenucleus accumbens, caudal anterior cingulate,anterior cingulate, dorsal anterior cingulateand substantia nigra) have also been conducted,enabling human-specific research into the pro -cesses in which these structures are involved.These include reward processing, often investi-gated with NHPs (53).

— Facial processing and recognition: Single-neuron studies of the human amygdala haveinvestigated the perception of faces, moodsand emotions, revealing how human singleneurons respond to different parts of the faceselectively; this is often studied with NHPs(54).

— Language, memory and learning: The output ofjust one laboratory over a period of 24 yearsillustrates what can be achieved with humansingle-neuron studies. Perception, object nam-ing, verbal memory and association, among oth-ers (with many protocols not possible inanimals), were elucidated by using almost 200

patients, more than 250 recording sites, andalmost 500 single neurons (55).

— Reach and grasp, motor prostheses: Single-neu-ron human studies in patients undergoing deepbrain stimulation (DBS) surgery have aided thediscovery and understanding of neurons encod-ing directional movement and intent, and themodulation of gripping force. Work in amy-otrophic lateral sclerosis (ALS) patientsimplanted with neurotrophic electrodes hasenabled patients to control cursors, speech syn-thesisers and robotic fists (56). Recent advancesin BMI technology allow the exclusive study ofhuman motor cortical control and neurophysiol-ogy, with any (arguable) past contribution ofNHP experiments having no bearing on whatcould, and should, be done now and in thefuture.

— Seizure generation: Invasive and non-invasivehuman research, including electrode-basedmeasurements of LFPs and action potentials ofindividual neurons, have been conducted sincethe 1950s, in patients with movement disordersand pharmaco-resistant focal epilepsy. It hasbeen suggested that there is “plenty of room toaddress interesting new questions by using therecording approaches presently available”,which include subdural arrays of penetratingelectrodes. It is notable that the techniqueshave evolved from using a few electrodes,recording over a few minutes under generalanaesthesia, to arrays of microwires, implantedat depth, recording multiple activity in multiplebrain areas semi-chronically. Such investiga-tions have yielded important insights intohuman epilepsy (57).

fMRI development and basis/interpretationof the BOLD (Blood Oxygen LevelDependent) signal

It is claimed that invasive NHP experiments werecritical to the development of our understanding offMRI imaging, by way of elucidating the nature ofthe neural activity underlying the BOLD fMRIresponse (17, 58, 59), based on what some regardas a seminal paper published in 2001 by MaxPlanck Institute researchers: the Logothetis et al.(63) paper (hereafter referred to as ‘the Logothetispaper’, not to be confused with subsequent reviewsby Logothetis, also referred to here [58, 60]).Indeed, it has been claimed that fMRI images “onlybecame interpretable by doctors at nerve-cell levelthanks to the work of Max Planck researchers”,that work with monkeys was the first to show “thatBOLD fMRI actually does measure changes in theactivity of nerve cells” (61), and that it allowed usto “comprehend the neural processes underlying


such metabolic changes in order to be able to cor-rectly interpret the functional scans used to assessthe condition of patients with various neurologicalor psychiatric diseases” (17). This is of major con-sequence, as fMRI has become one of the core tech-niques of neuroscience, and has revolutionised it.These are very bold and significant claims, so whatfollows here is a detailed and critical assessment ofthem — as is warranted by the associated harms tothe monkeys involved. It is not intended to be, andshould not be inferred as, an overt criticism of theindividual(s) who made the claims and who con-ducted the research. Nonetheless, our examinationleads us to contend that these claims are incorrect,given the history of fMRI development and under-standing described in the literature.

Given the salient consideration of the directionin which neuroscience research should be head-ing, and what techniques current research shouldbe employing, these claims are redundant in anycase. What went before, even if such claims weretrue, is of no consequence. It is of academic inter-est to analyse these claims, to help gauge theveracity and validity of statements asserting theworth of NHP neuroscience. We show here thatthere already existed substantial weight-of-evi-dence from human studies to support the hypoth-esis, and that, despite the argument put forwardby Logothetis, and in view of related human evi-dence, the Logothetis paper, and the monkeyresearch it entailed, simply cannot be consideredseminal.

Human fMRI was first reported in 1991 (adecade prior to publication of the Logothetis paperin 2001), followed by the first BOLD-based humanbrain activation results and other human imagingexperiments in 1992–93. These experiments builtupon developments in PET imaging in the 1970sand 1980s (reviewed in Bandettini [62]). The cruxof the debate is whether NHP experiments wereindispensable to this progress, and in the deci-phering of the nature of the fMRI signals. WouldfMRI have been developed at all, or over a similartime scale, without them? Would we really be lessable to understand the basis of BOLD fMRIwithout monkey experiments?

What Logothetis et al. did was to perform simul-taneous intracortical recordings (LFPs, and single-unit and multi-unit spikes) of neural signals andfMRI responses, in the monkey visual cortex, inorder to investigate and establish the link betweenneural activity, i.e. what we want to measure (andwhat is indirectly measured) in fMRI, and thehaemodynamic response (changes in bloodflow/oxygenation, etc.) in response to neural activ-ity, i.e. what fMRI directly measures. This showedthat increased BOLD contrast did reflect anincrease in neural activity, and, because the great-est correlation was between LFP (synaptic activ-ity) and BOLD signals, that BOLD reflects neural

input to a given area, rather than its ‘spiking’ out-put (63). The main claims are that: a) this was thefirst simultaneous study of intracranial recordingswith fMRI; b) this simultaneous approach wasessential to fully address the nature of the BOLDresponse; c) prior simultaneous recordings of fMRIand EEG or optical imaging suffered from poorspatial resolution (EEG) compared to intracorticalrecordings; d) this research enabled the aforemen-tioned conclusions to be drawn; and e) establishingthe detailed relationship between BOLD signaland neuronal activity will continue to depend onNHP experiments. We argue, however, that thisrepresents an exaggeration of the importance ofthese monkey experiments, and overlooks both thecontributions of, and potential of, human-basedinvestigations at that time. For example:— Prior weight-of-evidence suggested the same

conclusion: While the Logothetis paper mayhave been the first to directly investigate neu-rovascular coupling in the manner it did, priorweight-of-evidence strongly suggested the sameconclusion. The paper itself cited previoushuman research examining concurrent EEGsignals and fMRI images, though the impor-tance of this was downplayed by the authors,based on the ‘poor spatial resolution’ of EEG.One such cited human study, published fiveyears previously and also involving the samearea of the visual cortex (‘V1’ or the ‘striate cor-tex’), set out “to understand how the fMRIresponse relates to neural activity” and reportsthe results of three tests supporting the hypoth-esis that fMRI responses are directly propor-tional to local average neural activity over aperiod of time, with which Logothetis et al.’sdata were “consistent” (64). Another reviewcites human work going back to 1995 that pro-ductively investigated neurovascular coupling:examples include fMRI of patients performingmotor tasks, coupled with electrophysiologicalmapping of the sensory-motor regions of thecortex via cortical stimulation of the motor cor-tex to elicit hand movement and evoked-response recording in the somatosensory cortexduring tactile stimulation of the hand; and com-parison of gamma-band LFP signals in the pre-motor cortex with fMRI activation in patientsperforming tasks (see Mukamel & Fried [65]).Meanwhile, a 2002 review (66) acknowledgesthat Logothetis et al. “pioneered the simultane-ous acquisition of electrical and fMRI data inprimates”, but also noted that, regarding theirconclusion, “This is in agreement with recentdata that show a significant correlationbetween fMRI BOLD responses and evokedpotentials in humans, and the literatureregarding evoked field potentials and cerebralblood flow in animals”. All in all, many exam-ples are provided of human studies that eluci-


date the degree of neurovascular coupling, i.e.the relevance of fMRI imaging to actual neuralactivity, in various areas of the human brainand at different power bands.

— Many questions remain regarding neurovascu-lar coupling in humans, including the validityof the findings of Logothetis et al.: Argumentsover linearity of haemodynamic response, andindeed between BOLD signals and underlyingneural activity, continue to this day: they are“still a matter of debate, even after decades ofresearch” (67), and “…a detailed understandingof the neurovascular coupling process remainselusive” (68). Thus, some 15 years after thestudies reported in the Logothetis paper in2001, they still cannot be considered definitive.Some reports suggest that the response is linearunder certain conditions (e.g. 64), while othersshow non-linear haemodynamic responses (e.g.69); some reports show BOLD fMRI correlateswith underlying LFP rather than spiking activ-ity (as in the Logothetis paper, and also shownin experiments involving cats [70]); othersshow, in contrast, that fMRI does correlate withspiking, while other studies suggest it corre-lates with both (human-based studies in 2005[71]). These findings have been summarised byKim et al. (67). Ekstrom also discusses thisconundrum, accepting that “…the relationbetween the blood oxygen-level dependent sig-nal and underlying neural activity remains anopen and actively researched question”, with“much for us to understand” (72), and alsoasserts that the model supported by Logothetiset al.’s 2001 results (BOLD reflecting peri-synaptic activity, i.e. LFP rather than neuronspiking [63]) is challenged by situations inwhich BOLD, LFP and spiking dissociate. Thisuncertainty is further acknowledged and dis-cussed at length in another review published in2012 (73).

In view of the above, the Logothetis paper —and, therefore, the issue of the basis of theBOLD signal, as reported by it and the otherpapers cited here — is not an ‘open-and-shutcase’, and important questions remain. Theexplanation of neurovascular coupling, it offersmay be only partly correct, and/or may be onlycorrect in certain circumstances. Or it mightactually be incorrect overall.

— Human studies were possible instead, and hadto be conducted anyway: The authors of a 2013review of neurovascular coupling, citing theLogothetis paper, acknowledge that the extrap-olation of those studies to the human brain wasunclear, and that human studies were neces-sary (74). They cite multiple human studiesthat have, since, suggested a similar correlationbetween the BOLD signal and ECoG/LFP sig-nals in the gamma range (30–130Hz), and also

between BOLD and single-neuron firing whenthere is also firing of nearby neurons. Suchhuman studies have been conducted and pub-lished since the late 1990s.

A review contemporary to the Logothetis paper(75) illustrated many examples of simultaneouselectrophysiological and haemodynamic studies inhumans, broadly of the type performed byLogothetis et al. in monkeys. While none usedintracranial deep electrodes, as did Logothetis etal., they at least show that such studies are possi-ble in humans, and include a variety of combina-tions of PET, EEG, fMRI, MEG and invasiverecordings. A 2006 review, citing many humanstudies, described how combined fMRI and func-tional near-infrared spectroscopy (fNIRS) studieshave illuminated the relationship between neuralactivity and the BOLD signal (76). It cites the first‘combined study’ that took place, in humans, in1996, as well as more than a dozen other humanstudies. Further, it concludes that, while moststudies focus on changes in haemoglobin (Hb) con-centrations and correlation with BOLD signal, var-ious other contributory factors have often beenoverlooked, such as blood flow dynamics, blood vol-ume and changes in oxygenation. The authorsremark that “…the details of the translationbetween an ensemble of neurons firing and theensuing increase in focal cerebral blood flowremain controversial. The lack of a detailed under-standing of the underlying physiology did not hin-der an overwhelming success of fMRI; on the otherhand, the more complex the paradigms investi-gated the more mandatory is a thorough under-standing of the imaging signal”. The point is that,undoubtedly, fMRI has been an overwhelming suc-cess without the detailed understanding of theunderlying processes that NHP researchers (whouse monkeys to investigate it) claim they are pro-viding with their research; and, while a deeperunderstanding may or may not be desirable, thiscould be achieved solely via human investigation.

Other salient examples include a human studyby Arthurs et al. in 2000 — a year prior to theLogothetis paper — that measured somatosensoryevoked potentials (at the scalp) in five healthy,unanaesthetised volunteers, alongside fMRIBOLD changes (77). While these two approacheswere not simultaneous, that appears to be a minordetail, given the convincing nature of the resultsand conclusion. It concluded that “…the BOLDresponse correlates with synchronized synapticactivity, which is the major energy consumingprocess of the cortex”. This is the main conclusionof the 2001 Logothetis paper. Further human workby the same (i.e. Arthurs) group was published in2003 (78), which cemented their findings “as hadrecently been shown in primates (Logothetis et al.,2001)”. In another study, combined EEG and fMRI


revealed how different components of EEG signalsare related to positive and negative BOLD signals,aiding their interpretation, and helping to “furtherisolate the neural mechanism underlying bothEEG and fMRI responses” (79). A 2013 review offMRI stated that ‘current’ techniques (which willhave improved further since publication) provideda resolution of 1mm3 spatially, and 1 secondtemporally (80).

While advocates of invasive NHP experimentsfrequently use the spatial resolution provided byinvasive electrodes as a defence of this approach,this resolution is sufficient for it to be central tothe Human Connectome Project, which is mappingthe connectivity of the human brain. Furthermore,this review mentions fMRI machines with evengreater resolutions: 9.4T systems, for example, and7T systems that have detected BOLD responseswith a 0.7mm3 resolution. As previously stated,this is improved when fMRI is used in combinationwith other techniques such as EEG. Finally, a2012 review summarised the contribution of simul-taneous human EEG/fMRI studies to the under-standing of BOLD signals/neural activity, whichelucidated the relationship between BOLD signaland alpha rhythms (81). This review also cited a1998 study, in which the relationship betweenactivity-dependent increases in cerebral blood flowand single-unit activity and LFPs was examined inthe rat cortex (82). This study showed that therewas a strong correlation between activity-depend-ent cerebral blood flow and LFPs. This conclusionis similar to that of the Logothetis paper, in that itlinks the basis of fMRI signal to LFP/synapticactivity, rather than spiking. In other words: whilethis study did not use fMRI as Logothetis et al. did,it provided evidential weight to the hypothesis thatfMRI is reporting synaptic activity/LFPs (in rats,rather than monkeys, and three years before theLogothetis paper). If these human studies wererequired in any case, due to questions over theextrapolation of NHP data, and wereconducted/could have been conducted instead, thenthe fact that Logothetis et al.’s ‘seminal’ experi-ments were performed in NHPs is superfluous:their major conclusion was not dependent on NHPuse.— Human studies alone are more than capable of

addressing the ongoing issue now and in thefuture, and are being used to do so: The 2012Singh review (73) goes on to suggest that neu-rovascular coupling can be (and is being) inves-tigated and resolved by means of non-invasivehuman EEG/fMRI experiments, as it has beensince the mid-1990s (prior to Logothetis’ inva-sive NHP work) as well as “electrode recordingsin implanted human epilepsy patients withBOLD fMRI in healthy human participants”.Regarding the former, non-invasive humanstudies, “At the invasive microscopic level,

these oscillatory signals can be found in LFPrecordings, where they reflect the integratedpost-synaptic potentials of neurons within amillimetre of the recording electrode. However,such signals can also be measured macroscopi-cally at the cortical surface by using either[ECoG, EEG or MEG] — these signals then rep-resent the synchronous activity of many squaremillimetres or centimetres of cortex”. In fact,such studies have revealed which componentsof the electrophysiological signals positivelyand negatively correlate with the BOLDresponse: even simple tasks have shown (viaMEG) reductions in alpha power, increases ingamma power, and evoked sustained DCchanges all closely localised with BOLDresponse (73). The authors conclude that a com-bination of MEG, fMRI and MRS can addressthis complexity in human studies. Other recentexamples include: a) simultaneous icEEG andfMRI in epilepsy patients, providing good qual-ity results as well as more events reported thanwith scalp EEG (83); b) investigation of the linkbetween BOLD response and neuronal activityby using fMRI and high-density ECoG grids inhumans (84, 85); and c) investigation of theassociation between fMRI activation and elec-trical activity, as well as brain connectivity,with simultaneous intracranial electrodes andfMRI in humans (86, 87).

— A detailed understanding of BOLD fMRI maybe of academic interest only, and not necessary:fMRI results generally ‘make sense’, meaningthat knowledge of the mechanistic basis maynot actually be needed for their successfulapplication. One analogy that has been prof-fered is that of a new telescope, offeringastronomers unprecedented clarity in theirview of the heavens. Would it matter, if nobodyfully understood how it worked, particularly ifit had been calibrated with well-known objectsand previous observations (68)?

In conclusion, the use of the Logothetis paper asevidence of the necessity of NHP neurosciencemust be considered specious.

Deep brain stimulation (DBS)

DBS has been, and remains, an effective therapy forthe tremor-related symptoms of PD for tens of thou-sands of patients. It involves the insertion of stimu-lating electrodes into deep brain structures calledthe basal ganglia, which, among other functions,control movement and posture, and whose normalfunction is disrupted in Parkinson’s disease (PD).Advocates of NHP research claim that macaqueexperiments (subsequent to the development of theso-called ‘MPTP macaque model’ of PD in 1983)


have been indispensable to its development, andspecifically to the development of DBS of the sub-thalamic nucleus (STN). Indeed, this argument, inaddition to the others described in this review, isrepeatedly used as a ‘flagship’ to showcase NHPneuroscience as a means of convincing the public, aswell as regulators and legislators, that NHP neuro-science is vital to progress in neurological medicine(e.g. 17, 41, 88–92). However, in common withBOLD fMRI above, this is of historic interest only,and has little or nothing to add to the critical con-sideration of the current and future value of NHPneuroscience. For the same reasons as apply toBOLD fMRI, it is worth addressing here, in order torebut those claims, as well as to illuminate the truenature of such ‘arguments of necessity.’

First, it must be appreciated that such claimshave not been made via a robust, systematicreview of all the available literature. Anecdotalevidence, citing NHP experiments in which STN-DBS has been investigated, is not sufficient; themere use of NHPs in DBS research/investigation ofthe STN is no measure of their crucial nature or oftheir contribution to the field. Indeed, a criticaland comprehensive review of the literature thatincludes all methods of investigation, alongsideother summaries of evidence, provides a com-pelling case against the necessity of NHP researchin the development of STN-DBS, and in support ofhuman observation and neurosurgical investiga-tion alone as the foundation of DBS treatment ofPD (93, 94), and these accounts should be con-sulted for an in-depth argument against it. Thesearguments are summarised in a recent ‘Letter tothe Editor’ published in ATLA, dedicated to theissue (95). Briefly, they demonstrate: — the major role of human studies historically in

uncovering the functional anatomy of the brain,as well as confounding species differences fromparallel animal studies, including of the basalganglia (and the STN);

— that the STN had been linked to movement dis-orders in humans, as long ago as the 1920s, asa result of both clinical and post mortem stud-ies, before similar observations were made inNHPs (e.g. 96); and

— that basal ganglia generally were being oper-ated on in the 1940s, to alleviate movement dis-orders, belying claims that this was down toNHP experiments (see examples in various arti-cles [97–100]).

Further, they describe: — the use of electrostimulation in humans since the

1960s, initially to establish the correct placementof needles for making thalamic lesions duringsurgery (see various articles [101–104]);

— how, during these procedures, it was noted thatthe stimulation of particular brain structures

could suppress the symptoms of movement dis-orders, including PD (see various articles[105–108]);

— how DBS could therefore be a probable alterna-tive to therapeutic lesioning (108, 109); and

— the use of human DBS, in the late 1970s and inthe 1980s, in various parts of the brain, includ-ing the basal ganglia, to control tremor, amongother things (110–113).

All these points illustrate that human studies pre-dated, by decades, the first report of the MPTPmonkey in 1983, as well as its subsequent use toinvestigate and characterise the basal ganglia andSTN in PD. Clearly, claims such as “...the treat-ment of PD by delivering DBS to the STN oweseverything to the research in non-human pri-mates...” (114) cannot be correct. Regarding theapplication of DBS to the STN specifically, in addi-tion to the association of the STN with movementdisorders almost a century ago (as stated above),and as part of myriad human investigations ofthe basal ganglia since then (see 115), it is alsoobvious that the STN would have become moreand more implicated and investigated in humansas a matter of course, without the need for inves-tigations in the ‘new and interesting’ MPTPmonkey.

Though the question of deliberate targeting ofthe STN in humans prior to work with MPTPmonkeys is perhaps debatable, it is clear that theSTN was, in any case, flagged as a potential ther-apeutic target in human investigations, prior tothe availability of MPTP monkey data. The typeof brain lesions performed in these human sur-geries inherently and unavoidably affected anarea within several millimetres (4–6mm, typi-cally) of the tip of the electrode used in those pro-cedures (116, 117). Given that the STN is withina similar several millimetres of various struc-tures that were targeted, it follows that therewill, undoubtedly, have been associated STNlesions. In other words, the consequences of STNlesioning were studied, however unwittingly. Infact, this was overtly discussed by researchers inthe 1960s, such as Andy et al. (116) and Hassleret al. (118).

Over and above human-based investigationunderpinning the development of DBS, it is alsoclear they are central to its further developmentand refinement, as well as to the understandingof PD pathology, with no need for recourse toNHP experiments. Illustrative examples include:— studying the effects of DBS-mediated stimula-

tion of the internal globus pallidus (GPi) onlocal neurons in unanaesthetised PD patientsundergoing surgery for stimulator implanta-tion, revealing details of changes in firing rateand patterns of neurons in the vicinity of theGPi (119);


— extracellular single-unit recordings in the basalganglia revealing the presence of bursting neu-rons and their firing rate; determination of howSTN activity is modified by L-dopa therapy, andhow the pallidal complex functions in terms ofbursting patterns and oscillations (120); and

— neuroimaging studies linking abnormal activa-tion patterns in the insula to PD-related cogni-tive decline, behavioural abnormalities andsomatosensory disturbances (121).

Neural oscillations and memory

It has been claimed that researching how visualstimuli are translated into memories, whichinvolves examining how different areas of thebrain collate information, “can only be examined inapes” (122). This type of research often involvesanalysing neural oscillations, and determininghow their phase and amplitude correlate withspiking, how this underpins ‘phase coding’(whereby variable neuronal spiking at particularphases of oscillations permits the coding of infor-mation such as spatial location), and how gamma-phase oscillations correlate with BOLD signalsfrom fMRI. These are central to memory formationand function, and an important area of research,as they “exhibit specific spatiotemporal patternsthat show active brain regions, indicate the typesof neuronal computations that occur, and revealhow information flows through the brain” (123).We contend that the claim that this is only possi-ble in apes is false. Historically, there are reportsof intracranial electrode-based investigations ofthe human cerebellum, which elucidated neuronaloscillations and perturbations during task per-formance (28). There are many more recentinstances of the human-specific investigation ofneural oscillations at various frequencies, includ-ing the theta band, during processing of (oftenvisual) stimuli, alongside associated memory for-mation and retrieval (e.g. 24, 124–127). Indeed,such human research, even involving intracranialelectrodes, has become quite routine: this permitsgreater spatial and temporal resolution, all in ahuman-specific environment, and because theseelectrodes can be implanted for days or evenweeks, complex human cognitive and task-relatedprocesses can be investigated in consciousindividuals (123).

The use of both surface and depth electrodesexpands the scope of investigation from the cortexto deep brain structures, and often microelectrodesare also implanted to record concurrent actionpotentials of single neurons. This has elucidatedthe neural basis of four complex cognitive domains:working memory, episodic memory, language, andspatial cognition (123). A 2005 review of invasiverecordings from the human brain cites more stud-

ies, including the elucidation of the function of thebasal ganglia, including details of oscillations/syn-chrony/coherence and the effects of dopamineagonists (37).

A 2014 review augments the above, and under-lines the capabilities of human-based research intoneural oscillatory mechanisms (128). This reviewcites numerous human studies that have shownhow these oscillations complement neural firing inthe neural representation of sensory perceptionsand memory, and how they contribute to theencoding and retrieval of memory. For example,transcranial alternating current stimulation(TACS) in humans has implicated oscillatoryphase in sensory processing; human studies haverevealed content-specific LFPs, showing both cate-gory-specific and stimulus-specific neuronal firingand LFP responses at different frequencies in thetemporal lobe; and human EEG studies havehelped to reveal the information contained in oscil-latory power, frequency and phase in facial repre-sentations and in auditory stimuli (128).

Neural control of movement

It is claimed that NHP neuroscience is vital for agreater understanding of the neural control ofmovement, i.e. how the brain, spinal cord and asso-ciated sensory and motor neurons interact to gen-erate synchronised, controlled, movements. Muchof this involves investigating different ‘tracts’ ofthe central nervous system (CNS), such as the cor-ticospinal tract (the major tract of nerves descend-ing from the brain, particularly motor areas of thebrain, through the spinal cord), the reticulospinaltract (the tract of nerves descending from the retic-ular formation of the brainstem to the trunk andlimbs, involved in motor functions such as postureand locomotion), and the dorsal root ganglia of thespine, which relay sensory stimuli to the spinalcord. Both NHP and human investigations are con-ducted, with the major difference between thembeing the methods used to stimulate and recordneural activity.

In common with other areas of research, much ofthe justification for NHP use rests on the claimthat measuring the activity of individual neuronsprovides greater information and resolution thanis possible via non-invasive methods, and that thisis difficult in humans. Indeed, it appears thatmany invasive NHP experiments are replicative ofprevious human or NHP investigations involvingnon-invasive approaches, and/or invasive researchwith other species (such as cats or rodents), per-formed to obtain greater detail and/or species-spe-cific data. For example, the recent research of agroup at Newcastle University that focuses on thisfield, has involved the extensive use of NHPs,many of which have undergone craniotomies and


laminectomies to facilitate the insertion of stimu-latory and recording electrodes into the brain andspinal cord. Much of this work had already beenconducted in cats (129–131) or, in some cases, wasvery similar in nature to work already completedin the same NHP species and/or in humans, butadded very little (or nothing) to existing knowledge(130, 132, 133). Further, some papers reportedimportant species differences between monkeysand humans that confounded their findings (131,132, 134, 135), while others, by referring to priorhuman studies of a very similar nature, suggested(at least indirectly) that further studies could bedone non-invasively in humans (131, 133–136). Forexample, non-invasive, surface-recording, high fre-quency EEGs were shown to accurately reflect thetiming of spikes in single neurons in one of thestudies described below (136), thus validating thisapproach as a non-invasive method, which alsohappens to be frequently used in human studies.For instance, non-invasive monitoring of spikeactivity in the human somatosensory cortex is pos-sible, and effective, when scalp electrodes are usedin combination with improved techniques to min-imise the ‘noise’ that confounded previous studies(137). In addition, tactile — as opposed to electrical— stimulation has been shown to be effective, per-mitting non-invasive investigation of thesomatosensory system in children (138). Combinedwith imaging techniques, such as MRI, fMRI andPET, we argue that these techniques render thistype of invasive single-neuron experimentationwith monkeys redundant.

Vision/binocular rivalry

The Max Planck Institute uses monkeys to investi-gate the neurological basis of binocular rivalry, aphenomenon where, when two different images arepresented to the two eyes simultaneously, theviewer is only conscious of one of the two images ata time (139). However, this is also being investi-gated in humans, such as a study of the modula-tion of responses of visually selective neurons inthe human MTL with alternating percept, withfindings consistent with those in NHP experiments(140). Indeed, reviews of the phenomenon talkabout both monkey and human studies inter-changeably (51).

Confounding Factors AdverselyAffecting NHP Neuroscience

Those who use NHPs in their research often arguethat the results are conclusive or provide crucialevidence, such as in the fMRI case above. Theseclaims are also commonly stated in submissions bychimpanzee researchers to the US Institute of

Medicine’s (IOM) chimpanzee research inquiry(see below; 141), and many others. What they oftenignore are the numerous confounding factors thatmean that the results of NHP research must beviewed with much more scepticism.

Species differences: Genetic

Claiming that similarities in brain structure andfunction are sufficient evidence to support the use ofNHPs in neuroscience is superficial and inadequate.This is axiomatic, because in a complex living sys-tem such as an individual NHP, ostensibly minordifferences can cause significant disparities in bio-logical processes and their outcomes. It is clear fromsome of these biological and phenotypic differencesthat studies of NHP brains can only provide defini-tive information about the species studied, and maybe misleading, if used as analogues for, or to predictresponses in, human brains. This limitation israrely acknowledged by NHP researchers, and itshould be much more fully appreciated and consid-ered. Proponents of NHP neuroscience must be ableto demonstrate a comprehensive correlation with,and predictive nature for, human brain functionthat has resulted in translation to clinical practice,as a result of data that could not have been obtainedin any other way.

The dearth, up to now, of comparative biologyrelating to the suitability of NHPs as a model forhuman neuroscience can be considered surprising,given the extent of their use and the associatedcosts and harms. Encouragingly, however, recentyears have seen increased effort to investigate andunderstand these species differences, though theirapplication to the critical questioning of how theyaffect inter-species extrapolation of experimentalresults remains poor. Some of these fundamentalgenetic and biological differences, as well asobservable physiological and functional dispar -ities, are summarised here.

First, it is useful to appreciate the degree of sig-nificant differences between humans and even our‘least different’ relative, the chimpanzee, whichwas used in biomedical research (at least in theUSA) until recently, when it was deemed to beunnecessary following an in-depth review by theUS IOM (141). A comprehensive review of thesedifferences highlighted disparities in all aspects ofgene expression and protein function, from chro-mosome and chromatin structure to post-transla-tional modifications (142). The collective effects ofthese differences are extensive and widespread,and they revealed the superficial similaritybetween human and chimpanzee geneticsequences to be of little consequence for biomedicalresearch. These differences included some thatwere particular to the brain, and thus are perti-nent to this report:


— A study examining the expression of around12,000 genes in the prefrontal cortex of thebrain found that almost 1,000 were expressedin the human, but not in the chimpanzee, whilethe reverse was the case for 344 genes. In addi-tion, of the genes that were expressed in bothspecies, 20% showed a different expression pro-file (for example, 19 genes linked to Alzheimer’sdisease, PD and Huntington’s disease inhumans were expressed differently in chim-panzees; 143). In the cerebral cortex, at least169 genes are expressed differently (many ofwhich are involved in neuroprotection andsynaptic transport), and 916 genes areexpressed at least two-fold differently in thecerebellum (144). Furthermore, many genesinvolved in oxidative metabolism and mitochon-drial function are expressed to a higher degreein the human brain than in the chimpanzeebrain.

— Of approximately 10,500 genes studied in varioushuman and chimpanzee organs, 34% showed dif-ferential expression in the brain (145).

— The expression levels of 90 transcription factorgenes were significantly different in human andchimpanzee brains. These gene networks areenriched for primate-specific KRAB-ZNF genes,which are central to human and chimpanzeebrains and are associated with genes involvedin the development and maintenance of thisorgan (146).

— In humans, but not in chimpanzees, 61 genesare up-regulated and another 55 are down-regulated by the FOXP2 protein. The genesinvolved are important for brain developmentand function (for example, those involved incraniofacial formation and in establishing theneural circuitry and physical structures neededfor spoken language via cerebellar motor func-tion), and in the formation of cartilage and con-nective tissue (147).

— Many splicing factors are differentially expressedin humans and chimpanzees, including 20 in thebrain. This will result in many protein variants,which may have distinct functions in the brainsof humans and chimpanzees (148).

It follows that these profound differences in genecomplement and expression between human andchimpanzee brains will be even more significantbetween humans and monkeys, and therefore, arelikely to adversely affect the translation of data tohumans to a greater degree. A more recent review(from 2014) illustrates this in detail (149), withmany of the cited differences affecting the brain: — Parallel duplications and losses of the RHOXF2

gene in humans and 16 NHP species, alongsidedifferent patterns of expression, are thought tohave important inter-species biological implica-

tions due to the role of the gene as a transcriptionfactor (modulating the expression of genes underits control) and in developmental processes.Notably, RHOXF2 is expressed differently in thebrains of human newborns and embryos, and itregulates the expression of at least three othergenes involved in the function of the CNS. It istherefore thought to be involved in CNS functionand brain development, with significant implica-tions for inter-species differences (150).

— Humans have a rate of gene turnover 2.5-timesthat of all other mammals, which includes sev-eral gene families, notably genes preferentiallyexpressed in the brain (151).

— One study reported that over 7% (correspondingto 893) and 6% (corresponding to 789) of 12,473genes in the cerebellum showed increased anddecreased expression, respectively, in humanscompared to rhesus macaques (152), whileanother study noted that 91 genes were differen-tially expressed in human brains relative to thoseof rhesus macaques, as well as chimpanzees(144).

— An investigation of micro-RNA (miRNA)expression and regulation in the brain, specifi-cally in the prefrontal cortex and the cerebel-lum of humans, chimpanzees and rhesusmacaques, noted that up to 31% of the 325miRNAs examined “diverged significantly”between humans and rhesus macaques, andthat human-specific miRNAs were associatedwith neurons and with target genes involved inneural functions, supporting the theory thatmiRNAs have contributed to the evolution ofhuman cognitive functions. Of the 413 miRNAsexpressed in the human brain, 11% were notdetected in rhesus macaque brains, and almostone third (31%) of miRNAs common to thehuman and rhesus macaque prefrontal cortexwere differentially expressed in those twospecies. Of those differentially expressed pre-frontal cortex genes, 77% were also differen-tially expressed in the human and the rhesusmacaque cerebellum (153).

— Such is the degree of change of miRNA expres-sion and the repertoire of their target genesacross NHP species, developmentally throughoutNHP lifespan, and developmentally throughoutlifespan across NHP species, that miRNAs arethought to be the basis and major driving forceof the evolution of the human brain. This wasevidenced by a study of the prefrontal cortex andcerebellar cortex transcriptomes of humans,chimpanzees and rhesus macaques of differentages, which revealed significant variationbetween these types, in addition to sequencedivergence in cis-regulatory regions (154).

— One human-specific miRNA, miR-941, is highlyexpressed in the brain and has been implicated


in neurotransmitter signalling via the roles ofsome of its target genes. Of note, the host geneof miR-941 (mi941 is an intronic miRNA),DNAJC5, encodes cysteine-string protein-α(CSPα), which has been linked to neurodegen-erative diseases including Huntington’s andParkinson’s diseases, and adult neuronalneroid-lipofuscinosis; and miR-941 may be asso-ciated with hedgehog and insulin signallingpathways, with associated roles in humanlongevity and in some cancers (155).

— The adenosine-inosine editing rate, and there-fore the resultant changes in gene function andexpression, is higher in humans than in NHPs,including rhesus macaques, due to primate-spe-cific Alu sequences. This appears to particularlyaffect the human brain, via genes associatedwith neuronal functions and neurological dis-eases including bipolar disorder, motor neurondisease, Alzheimer’s and Parkinson’s diseases,schizophrenia, multiple sclerosis, and amy-otrophic lateral sclerosis (156).

— A recent connectome study reported majorspecies differences in the architecture of theinferior parietal cortex, and polar and medialprefrontal cortices (157). These findings aug-mented previous studies demonstrating agreatly expanded, lightly myelinated region ofprefrontal cortex in humans when comparedwith that in rhesus macaques and chimpanzees(158), and a more gyrified prefrontal cortex inhumans compared to other primates, evenallowing for differences in brain size (159).Functional consequences of these differencesmay involve sensory perception, visceral func-tions, higher order cognitive functions, andemotional and reward-related behaviours (see157).

— Comparative studies of human, chimpanzee andrhesus macaque genomes identified, for example,different numbers of long inverted repeats (LIRs)associated with orthologous genes in thesespecies. There were 546 of these in humans, ofwhich 421 (77%) were human-specific, but therewere only 130 in the rhesus macaque, of which107 (82%) were rhesus macaque-specific. Genesassociated with the human-specific LIRs wereinvolved in neural development and function,and in cell communication (160).

— The study of the MCPH1 gene (one of at leastseven key genes known to be involved in the reg-ulation of brain size during development) illus-trates how specific mutations can result infunctional changes, leading to altered regulatoryeffects in downstream genes, and ultimately tosignificant species-specific phenotypes and evolu-tion (161). The regulatory effects of human andrhesus macaque MCPH1 were different in threeout of eight downstream genes tested, and the

human-specific mutations altered the regulatoryeffects on the downstream genes.

— The profound effects of genetic differences canbe illustrated by the SRGAP2A gene, whichproduces a truncated protein in humans, butnot in great apes. This appears to underlie dif-ferential morphology and density of dendrites,linked to different behaviour and cognition.

— Several human-accelerated enhancers havebeen discovered. These are non-coding DNAsequences involved in enhancing the expressionof one or more genes, which have evolved inhumans to be significantly different to their cor-responding sequences in other species.Recently, one has been directly linked to differ-ential brain development. This enhancer(HARE5) of the Fzd8 gene has features uniqueto humans in terms of sequence, temporal andspatial expression, and transcript abundance,notably between humans and macaques, whichaffect “…the cell-cycle dynamics of a criticalpopulation of stem cells during corticogenesisand may underlie some distinctive anatomicalfeatures of the human brain”, including braindevelopment and size (162).

Genetic differences such as these clearly have bio-logical sequelae. Indeed, it has been postulatedthat “…the entire topology of a complex brain net-work can be reprogrammed by subtle adjustmentsof many genes that act additively to produce agiven phenotype” (163). Some work has uncoveredthe functional consequences of differential genecomplement and expression in the brain, while thegenetic basis of other, empirical, biological differ-ences has yet to be established.

Species differences: Physiological and functional

Primate brains are known to “differ in aspects ofstructural detail, as well as in overall size” (164).For example, the human neocortex differs from thatof other great apes in several ways, including hav-ing an altered cell cycle, prolonged corticogenesis,and increased size (see Boyd et al. [162]). It is nowaccepted that various primate brains are far fromsimply scaled-up or scaled-down versions of eachother. Over and above this general observation,detailed knowledge of important genetic, structuraland functional differences is beginning to emerge asthe question of inter-species brain differences isaddressed. While some features of cortical organisa-tion are common to various mammalian species, itis clear that “phyletic variation in cortical organiza-tion is far more extensive than has generally beenappreciated or acknowledged” (165). Examplesinclude differences in cortical neuron genetics andbiochemistry, as well as their connectivity, organ-


isation and function in rats, and visual system dif-ferences in monkeys (165). Varied techniques suchas fMRI, PET imaging and diffusion-weighted imag-ing (DWI) have revealed human brain specialisa-tions (compared to other primates) with regard todevelopment, cortical organisation, connectivity,ageing, and visual and auditory pathways (166).The insular cortex in humans is involved in varioussomatosensory and visceral sensorimotor functions,emotions, music, language, and other aspects ofawareness and perception, and it shows extrememorphological variability between species. Thesedifferences include not only gross morphology, butalso “laminar organization, cellular specialization,and structural association” (167). Due to this, and toits connectivity with important and well-researchedbrain areas, such as the anterior cingulate cortex,the frontal pole and the dorsolateral prefrontal cor-tex, the parietal and temporal lobes, the entorhinalcortex, and the amygdala (167), associated func-tional differences between species will confoundtranslational research.

A review in 2013, outlining the biological basis ofcortical evolution, outlined many human-specificand species differences in the cerebral cortex (168).For example: — “The human cerebral cortex has expanded sig-

nificantly relative to other hominids, includingintroduction of new regions in the frontal andparieto-temporal lobes in humans.”

— “…although the basic principles of brain devel-opment in all mammals may be conserved, themodifications of developmental events duringevolution produce not only quantitative butqualitative changes as well.”

— Differences in brain size (such as betweenhumans and monkeys) reflect not only differ-ences in cell number, but also in the arrange-ments and connectivity of those cells.

— Much of cerebral expansion and evolution is dueto the action of genes involved in the control ofcell division/the cell cycle, in addition to thevery different durations of cortical neurogenesis(humans 100 days; macaques 60 days; mice sixdays).

— Cerebral evolution/expansion is governed andaffected by many genes, and in turn by smallmodifications in those genes and their regula-tory elements. This is evidenced by mutationsin these genes causing intellectual disability inhumans.

— Human-specific gene networks (mainly involvedin neuronal morphology and synaptic function)have been linked to differences in the cerebralcortex, most specifically the frontal lobe, forexample.

— Human neuropil (in effect, grey matter) is sig-nificantly expanded compared to other pri-

mates, especially the prefrontal cortex, andprocesses such as dendritic and synaptic matu-ration and synaptic elimination are prolongedin humans compared to other primates.

A comparative analysis of the macro-scale connec-tivity of the human and macaque brains has beenconducted by Goulas et al. (169). While it was con-cluded that, “on the whole” they are “similarlywired”, there were also instances of “divergingwiring patterns” and “novel evolutionary aspects” inparticular areas, leading to concerns that the suit-ability of macaques for human neuroscience may bechallenged by unique human features, includingconnectivity reconfigurations. It was discoveredthat just over half (45/82 regions, 55%) of the braincan be considered as significantly similar — mean-ing, of course, that almost half (45%) cannot.Overall, the authors concluded there were differ-ences in macro-scale connectivity in the prefrontal,parietal and cingulated regions of the cortex: allregions extensively studied in macaques. Such dif-ferences are thought to underlie cognitive processesunique to humans. Further, there are “pronouncedchanges” in the arcuate and inferior fronto-occipitalfasciculi; there are differences in the functional andconnectional architecture of some regions of theparietal cortex, notably in the medial region; the dif-ferent connectivity in, and divergent functionalroles of, the anterior cingulate cortex may cause dif-ferences in decision-making, cognitive, motivationaland motor processes, while that of the posterior cin-gulated may differentially influence social cogni-tion. Notably, these differences are underpinned byboth genetic, epigenetic and environmental factors,and further, the functions of various brain regionsdepend on other factors in addition to connectivity,such as the laminar patterns of meso-scale connec-tions (169). Consequently, it is acknowledged thatmacro-scale connective similarities do not guaran-tee functional similarity; functional divergence isknown in conserved networks. It is therefore notenough for advocates of macaque neuroscience touse any degree of connective similarity betweenmacaques and humans to support their claims ofthe validity and human relevance of their NHPmodel.

The impact of genetic, epigenetic and environ-mental factors on the composition and function ofthe cortex has been thoroughly reviewed (170). Asexpected, intrinsic cortex genes are directlyinvolved in the development and specification ofcortical fields, which are also affected by extrinsicgene products that are associated with sensoryreceptor type, location and function, as well as epi-genetic factors that depend on environment andstimuli. Examples include numerous genes thataffect the expression of transcription factors andother regulatory genes, which regulate patterningin the developing cortex and cortical field size and


location. In summary, “…it is clear that genes actin a sequential and combinatorial fashion, andthat an alteration in the spatial and temporal pat-tern of expression at any stage could result in dra-matic changes in the resulting cortex” (170).

A recent study, also acknowledging the paucityof work comparing the connectomes of primatebrains, compared inter-regional brain connectionsacross humans, chimpanzees and rhesus macaques(157). In common with the Goulas et al. studydescribed above, this revealed a largely conservedstructural architecture in the three species, butalso revealed “major differences” in the inferiorparietal, polar and medial prefrontal cortices,including hubs present in these areas of the NHPbrains that were absent in humans. Due to thefunctional roles of these areas, these differencesmay affect high-order cognition, emotional andreward-related behaviours, visceral functions, andsensory processing. Notably, the human prefrontalcortex is one of the most enlarged brain regionscompared to that in NHPs. It is more gyrified,shows major differences in functional organisation,perhaps especially in visual pathways, and struc-tural differences are supported by fMRI data. The“pronounced changes” in the arcuate fasciculusmentioned by Goulas et al. were investigated byRilling et al. (171), revealing its substantial expan-sion in the human brain compared to the brains ofmacaques and chimpanzees, compatible with itsrole in language. Other temporal-frontal pathwayshave expanded, too. Indeed, the temporal cortexseems to have “undergone a substantial reorgani-sation since the last common human–macaqueancestor some 29 million years ago”, and the func-tional connectivity between higher-order auditoryareas and the medial and lateral frontal cortex dif-fers between humans and macaques. There is aparietal-frontal network in humans that “cannotbe matched to any macaque network” (172).

The impact of genetics on CNS function has beenelegantly illustrated by studies of neuroplasticity,which is defined as “a multifaceted and dynamicprocess involving gene–environment interactionsthat result in both short- and long-term changes ingene expression, cellular function, circuit forma-tion, neuronal morphology, and behaviour” (173).Both genetic and epigenetic changes mediate “var-ious aspects of experience-dependent plasticity,such as learning and memory, stress responsivity,and cognition”, and regulate normal brain func-tion, including memory (173). The intricate rela-tionship between genetic/epigenetic factors andbrain function is illustrated by the link betweenmutations in genes that encode chromatin bind-ing/modifying enzymes and many different neuro-logical disorders; as well as the link betweenenvironmentally-induced chromatin alterations inthe absence of mutations that have been shown tobe critical for neuronal functions including synap-

tic activity and cognition (173). Histone modifica-tions may influence gene expression so heavilythat they have been linked to various neurodegen-erative diseases of the CNS (Friedreich’s ataxia,and Huntington’s and Alzheimer’s diseases; 174).

It has been concluded that primate brains arequalitatively, as well as quantitatively, different,which explains species differences in cognitiveabilities (164). Some differences in the neocortex —particularly the prefrontal and temporal areas —have already been discussed above. Notably, thefrontal lobes of humans, in absolute terms, ‘dwarf’those of NHPs, which is thought to be of great rel-evance in explaining inter-species cognitive differ-ences. However, the frontal lobes are, at the sametime, much smaller than would be expected for aprimate of our brain size. The primary and premo-tor cortices occupy a much smaller proportion ofthe cortex in humans than in NHPs, and thebranching complexity of layer 3 pyramidal-cellbasal dendrites is markedly higher in the humanprefrontal cortex than in those of macaques ormarmosets, reflecting increased cortical connectiv-ity. Major differences have also been noted in thecerebellum, which is extensively connected to thecerebral cortex, and which is involved in movementas well as cognition. It has areas that are unique tohumans and apes, and is larger in humans, evenaccounting for body weight — although, relative tocortex size, the human cerebellum is smaller thanthat of NHPs.

The impact of such differences as those sum-marised here is acknowledged in some papers. Forexample, Boynton (71) discusses the growing dis-crepancy between monkey electrophysiologicaldata and human fMRI results, for which one of thepossible explanations is species differences. Theauthor suggests that, due to such differences, themonkey model will undoubtedly “break down” atsome point, as science pushes toward “higher-levelprocesses such as consciousness, learning and deci-sion making”, based on, for example, observationsthat the firing rate of neurons in the primaryvisual cortex of macaques (V1) is less affected thanthat in the human V1 by attention and by saccadicand binocular suppression, as evidenced by strongmodulation of the BOLD signal.

Other issues with NHP experimentation,affecting its human relevance

Experimental

Various significant and confounding species differ-ences have been described:— In vision research: Discrepancies between

humans and NHPs have been noted in experi-ments concerning the neurological basis of the


control of eye movements and gaze. Specifically,“with regard to the role of SMC [supplementarymotor complex] in response cancellation, non-human primate and human studies haveyielded somewhat different conclusions, poten-tially due to differences in species, responseeffector, and/or methodology” (175). It remainsunclear what the human homologue of the mon-key ITC actually is, and while data from LFPinvestigations in humans are “coarsely compa-rable” to those from NHPs, it is cautioned that“functional similarities should not be inter-preted to imply direct homology at the anatom-ical, cellular, or connectivity levels betweenhuman and monkey structures, species whosecommon ancestor lived about 30 million yearsago.” One “intriguing aspect of MTL responses”is that latency times differ greatly betweenspecies — human responses generally seemlonger than those of monkeys by two-fold orthree-fold. In the ITC, human latency times arealso “considerably longer” than those in themacaque. It is not known why, though it hasbeen postulated that major inter-species struc-tural differences mean the human ITC does notproject directly to the MTL, and/or includesmore synapses (51).

— In spatial navigation, it is acknowledged,“Whether results from rodent studies can bedirectly mapped onto humans is unclear, sincesubtle anatomical differences in MTL circuitrybetween species do exist” (49).

— There are notable (and important) differencesbetween results from the study of oscillations inhumans and non-humans (see Jacobs &Kahana [123] for specific citations). In workingmemory, cortical oscillatory-phase synchronyoccurs during memory retention in the betarange in humans, in contrast to the gammarange in non-humans. Hippocampal activityrelated to episodic memory differs in humansand non-humans. Human memory formation isassociated with decreased hippocampal activityat many frequencies, while in non-humans hip-pocampal theta oscillations increase in ampli-tude during memory encoding. Finally, “thetiming of neural responses and oscillations dif-fers between humans and monkeys in general”(176).

— In sleep research: Human microelectrode stud-ies of sleep ‘slow waves’ revealed “exciting find-ings” that were not expected based onnon-invasive studies and the animal literature.Slow oscillations were “remarkably synchro-nous” in animals under anaesthesia, but humanstudies of natural sleep concerning multiplebrain areas revealed that slow waves, andunderlying active and inactive neuronal states,occur locally, i.e. some regions can be active,

while others are silent. Also, wake-like andsleep-like activities have different durations indifferent cortical areas in different species (50).

— The confounding issue of species difference hasalso been acknowledged in BOLD fMRI:“because the BOLD signal is dependent on manyphysiological and biophysical parameters, whichcould vary between different species, these rela-tionships [between BOLD and neural activity]can be considered as semi-quantitative” (con-founding results regarding the correlation ofLFPs and spikes with BOLD), and “the haemo-dynamic response can vary widely across corti-cal areas and between species. Different aspectsof the haemodynamic response might change ondifferent timescales, and might have differentneural determinants and different consequencesfor the BOLD signal” (66).

— In seizure generation: While animal researchconcomitant to human studies has been on alarge scale, confounding species differencesexist. For example, it appears that there is “amore distributed epileptogenesis in humanepileptogenic cortex compared to a more focaland concentrated epileptogenic neuronal aggre-gate in experimental [animal] models” (57).

— Finally, and more generally, it is acknowledgedthat, relative to other animals used in experi-ments, NHPs “…possess substantial outbredgenetic variation, reducing statistical powerand potentially confounding interpretation ofresults in research studies” (177).

Anaesthesia

One major, and often overlooked, inherent problemwith animal models in neuroscience is the use ofanaesthesia. This is used to mitigate suffering inthe animals used, such as that associated with theplacement and use of electrodes, and also to preventmovement during data collection (178). Thoughsome protocols on animals do not involve anaesthe-sia (though crucially these tend to be limited toresearch into cortical structures only), many do.This is a significant concern, because anaesthesiaadversely affects the translation of findings fromanimals to humans, over and above species differ-ences. For instance, anaesthetic agents disrupt neu-rovascular coupling in several ways, such asaltering baseline haemodynamic parameters, anddifferent agents differentially affect the systemiceffects of experimental stimuli, with consequent dif-ferential BOLD fMRI responses (178). Anaestheticagents appear to delay haemodynamic responsefunctions, and, further, they disrupt neurovascularcoupling in different ways, posing particular prob-lems for pharmacological and neuroimaging studies(for a review see 179).


Welfare issues and stress

While not as commonplace as it once was, the cap-ture and confinement of wild monkeys continue forthe purposes of breeding and supply to laboratoriesworldwide. For example, almost 4,200 monkeyswere exported from Mauritius to the USA in 2014,including hundreds which were caught in the wild(180). This unavoidably causes severe stress anddistress, with lifelong consequences. However,once laboratory-housed (whether wild-caught orpurpose-bred), the handling, routine laboratoryprocedures, experimental protocols and so on, areall part of life for the monkeys, and all of these fac-tors cause unavoidable stress.

Neurological and vision experiments often causesignificant suffering, classed as ‘severe’. For exam-ple, electrophysiology experiments often requirehead restraint, with experiments typically involv-ing such procedures as: the removal of an area ofskull to expose the brain; posts being cementedonto the skull for restraining the monkey by thehead during recording and stimulating sessions;and, in some cases, scleral search coils beingimplanted in the eye to monitor eye movements.While it must be noted that some refinements havebeen reported in this area, such as the use of min-imally invasive ‘halos’, face masks and head capsin place of surgically implanted posts (181), andinfra-red eye tracking systems in place ofimplanted scleral eye coils (182), our investigationsshow that the former, more invasive approachesstill seem to be widely used. Either way, the asso-ciated restraint causes significant welfareproblems.

In addition, animals are sometimes deprived offood or water for many hours prior to the experi-ments, to motivate them to perform visual tasks.During recording or stimulating sessions, whichcan last for several hours each day, NHPs are usu-ally conscious and restrained in chairs by themetal fixtures cemented to the skull. To avoidother NHPs tampering with the implants, in somelaboratories the animals are singly housed for theduration of experiments, which may last formonths or even years. Due to the investment madein ‘preparing’ these animals, such NHPs are oftenused and re-used in similar experiments for verylong periods of time, and so inevitably suffer long-term stress when subjected to neuroscientificinvestigations. All of this has attendant conse-quences for their welfare, as well as for the scien-tific data derived from such experiments. Waterdeprivation — often core to some behaviouralresearch involving task training — is accepted as astressor, as the UK refinement working groupacknowledged: “…restricting access to food or fluidcan elicit behavioural and physiological responsesthat compromise animal health and welfare andmay affect the scientific data being collected” (183).

This also applies to restraint, the associated stressof which “must be carefully taken into account asthis is likely to have a range of general physiologi-cal and neurophysiological effects” (178).

Stress-related elevations of heart rate, bloodpressure and a variety of hormone levels (includingcortisol) are well known to affect scientific dataobtained from animals in laboratories (184–187),particularly those involving the nervous system.Indeed, warnings have been issued about the con-sequences of disregarding the effects of stress dueto laboratory routines (185–187), yet this remainsunder-reported in scientific studies, or notreported at all (188). Thus, stress associated withneuroscience experiments could, inter alia:— lead to alterations of blood pressure and flow

that could impact on fMRI research, whichrelies on blood flow to the brain for its meas-urement;

— increase the time taken to train macaques forcertain procedures, such as eye tracking tasks,as stressed animals take longer to train;

— affect the length of sessions that macaques willtolerate, causing research to take longer thannecessary; and

— affect vision in macaques, with increased likeli-hood of eyestrain impacting on the length ofsessions or the quality of data through issuessuch as poor accommodation (focusing onimages) and increased likelihood of involuntaryeye movements.

Human-specific attributes

One other consideration is that animal models —even if the above species differences and confound-ing factors were overcome to any significant degree— can never inform human cognitive processesthat simply cannot be studied well, or at all, in ani-mals, such as “language, imagery, episodic mem-ory, volition, and even consciousness” (65), as wellas dreams, imagined future scenarios, etc. (189).Further, it is contended that non-humans may noteven be sharing the same perceptions as humanswhen presented with visual stimuli, and thereforeit is desirable to perform relevant electrophysiol-ogy experiments in awake humans who can reporttheir perceptions (see 190). The repertoire anddepth of emotions that can be studied in animalsare also limited, rendering human research essen-tial (189).

Concluding Remarks

NHP experiments — in neuroscience, as well as inmany other areas of biomedical research — areconceived, funded and conducted on the basis of a


general assumption of human relevance and even-tual benefit, rather than firm scientific evidence oftheir value. This default position is maintainedsuperficially, on the basis of opinions of those whopractise NHP research, and anecdotal claims ofworth that fail to withstand scrutiny, or which, atthe very least, are controversial. Instead, anyrationale for NHP use should be based firmly onsystematic, robust, critical and independentevaluation.

In defending their practices and in condemningany criticism (or even questioning) of them, manyNHP researchers overstate the human relevance ofneuroscience research involving NHPs, its contri-bution to human neuroscience in the past, its cur-rent necessity, and its likely future contribution,with little or no substantiation. At the same time,there is a gross understatement of the contributionof human-based research to neuroscience, the sig-nificance of what this has achieved, the powerfuland ever improving performance of non-invasivemethods, the scope of what can be done in humans(both non-invasively and invasively), and the sig-nificance of species differences between monkeysand humans.

This defence of NHP neuroscience, based on aninflated portrayal of its importance alongsideundervalued and denigrated alternatives to it, isconsequently poor and misleading. This reviewaims to reset the balance of the argument, byshowing that: humane human-relevant neuro -cognitive (and associated) research is much morecapable, widespread, important and powerful thanthose who use NHPs accept; claims by NHPresearchers of the exclusive capabilities of NHPneuroscience are incorrect; and NHP experimentsare unjustifiable, due to both the lack of scientificnecessity, and the existence of significant inter-species differences that confound any resultsderived from them.

We have shown in this review that this appliesboth to NHP neuroscience generally, as well as toa number of salient, specific, claims, such as theuse of NHPs in the development of BOLD fMRI,DBS, etc. Overall, the value of current NHPresearch cannot be supported, either by anecdotalevidence or by any claimed historical successes.Any successes, even if they could be proven, mustbe weighed against failures. If there are a few ‘suc-cess stories’ in historical NHP neuroscience whereNHP experiments have contributed significantlywith data that could not have been derived byanother means, the many thousands of researchprogrammes that did not translate to human ben-efit must be taken into account. A few successes donot validate a model, if there are also orders ofmagnitude more failures. Secondly, in any case,this has little bearing on the need for NHP neuro-science now and in the future. New and improvingnon-NHP technologies, not available to science in

the past, provide ethical, novel and unprecedentedmeans of human-specific investigation.

Overall, the case that neuroscience would bemuch improved, and more relevant to and ulti-mately more successful for humans, should it beconducted with a solely human focus, is supportedby comprehensive and robust evidence. Given thegreat ethical and financial cost of NHP neuro-science, the inherent suffering and crueltyinvolved (as revealed by several recent investiga-tions [see crueltyfreeinternational.org]), as well asthe human ethical aspect in terms of the urgentneed for greater understanding of human neurol-ogy and neurological disorders, the onus must beon those who use NHPs in neuroscience to make anevidence-based case for what they insist they mustdo. Given the content of this review, we contendthat such a case cannot be made, and have confi-dence in our position in opposing NHP neuro-science scientifically and ethically.

Acknowledgement

The authors acknowledge and thank Cruelty FreeInternational Trust for funding this work, whichhas not been presented anywhere else in printprior to this publication. There are no financialconflicts of interest.

Received 27.05.15; received in final form 27.08.15;accepted for publication 17.09.15.

References

1. Humane Research Council (2005). US PublicOpinion of Chimpanzee Research, Support for a Ban,and Related Issues. Available at: https://faunalytics.org / feature-art ic le /u-s -publ ic -opinion-of -chimpanzee-research-support-for-a-ban-and-related-issues/ (Accessed 25.01.16).

2. Animal Aid (2003). Public Says ‘No’ to PrimateResearch. Available at: http://www.animalaid.org.uk/h/n/NEWS/news_experiments/ALL/136// (Acc -essed 25.01.16).

3. Aldhous, P., Coghlan, A. & Copley, J. (1999). Animalexperiments: Where do you draw the line? NewScientist 162, 26–31.

4. TNS Opinion & Social (2010). Special Eurobar -ometer 340 “Science and Technology” (Wave 73.1).Available at: https://open-data.europa.eu/data/dataset/S806_73_1_EBS340 (Accessed 25.01.16).

5. Leaman, J., Latter, J. & Clemence, M. (2014).Attitudes to Animal Research in 2014, 53pp. London,UK: Ipsos MORI.

6. APC (2013). Review of the Assessment of CumulativeSeverity and Lifetime Experience in Non-humanPrimates Used in Neuroscience Research, 145pp.London, UK: Animal Procedures Committee.

7. European Commission (2013). Seventh Report on theStatistics on the Number of Animals used forExperimental and other Scientific Purposes in the


Member States of the European Union, 14pp.Brussels, Belgium: European Parliament.

8. Anon. (2010). Directive 2010/63/EU of the Euro -pean Parliament and of the Council on the protec-tion of animals used for scientific purposes. OfficialJournal of the European Union L276, 20.10.2010,33–79.

9. Tardif, S.D., Coleman, K., Hobbs, T.R. & Lutz, C.(2013). IACUC review of nonhuman primateresearch. ILAR Journal 54, 234–245.

10. Vallender, E.J. & Miller, G.M. (2013). Nonhumanprimate models in the genomic era: A paradigmshift. ILAR Journal 54, 154–165.

11. Chan, A.W. (2013). Progress and prospects forgenetic modification of nonhuman primate modelsin biomedical research. ILAR Journal 54, 211–223.

12. Capitanio, J.P. & Emborg, M.E. (2008). Contri -butions of non-human primates to neuroscienceresearch. Lancet 371, 1126–1135.

13. Roelfsema, P.R. & Treue, S. (2014). Basic neuro-science research with nonhuman primates: Asmall but indispensable component of biomedicalresearch. Neuron 82, 1200–1204.

14. Weatherall, D. (2006). The Use of Non-HumanPrimates in Research, 147pp. London, UK: TheAcademy of Medical Sciences.

15. Bateson, P. (2011). Review of Research Using Non-Human Primates, 51pp. Available: http://www.bbsrc.ac.uk/web/FILES/Reviews/review-research-using-nhps.pdf (Accessed 25.02.15).

16. Moore, E.J. (2008). Non-human primate research:Whither now? Journal of the Royal Society ofMedicine 101, 165–167.

17. Logothetis, N.K. (2014). MPI-Research: BUAV-SOKO Claims & Reality. Available at: http://hirnforschung.kyb.mpg.de/en/press/the-greatest-enemy-of-science-is-ignorance/soko-tierschutz-and-buav.html (Accessed 04.12.15).

18. Moore, E. (2014). Medical relevance of UK-fundednon-human primate research published fromJanuary 1997 to July 2012. Journal of the RoyalSociety of Medicine 107, 264–270.

19. Rossini, P.M., Rossini, L. & Ferreri, F. (2010). Brain-behavior relations: Transcranial magnetic stimula-tion: A review. IEEE Engineering in Medicine &Biology Magazine 29, 84–95.

20. Di Lazzaro, V. & Ziemann, U. (2013). The contribu-tion of transcranial magnetic stimulation in thefunctional evaluation of microcircuits in humanmotor cortex. Front Neural Circuits 7, 18.

21. Barany, D.A., Della-Maggiore, V., Viswanathan, S.,Cieslak, M. & Grafton, S.T. (2014). Feature inter -actions enable decoding of sensorimotor transforma-tions for goal-directed movement. Journal ofNeuroscience 34, 6860–6873.

22. Husain, F.T. & Schmidt, S.A. (2014). Using restingstate functional connectivity to unravel networks oftinnitus. Hearing Research 307, 153–162.

23. Cho, Z.H., Son, Y.D., Choi, E.J., Kim, H.K., Kim,J.H., Lee, S.Y., Ogawa, S. & Kim, Y.B. (2013). In-vivo human brain molecular imaging with a brain-dedicated PET/MRI system. MAGMA 26, 71–79.

24. Hill, N.J., Gupta, D., Brunner, P., Gunduz, A.,Adamo, M.A., Ritaccio, A. & Schalk, G. (2012).Recording human electrocorticographic (ECoG) sig-nals for neuroscientific research and real-time func-tional cortical mapping. Journal of VisualizedExperiments 64, e3993.

25. Davidesco, I., Zion-Golumbic, E., Bickel, S., Harel,

M., Groppe, D.M., Keller, C.J., Schevon, C.A.,McKhann, G.M., Goodman, R.R., Goelman, G.,Schroeder, C.E., Mehta, A.D. & Malach, R. (2014).Exemplar selectivity reflects perceptual similaritiesin the human fusiform cortex. Cerebral Cortex 24,1879–1893.

26. Meshulam, M., Ramot, M., Harel, M., Kipervasser,S., Andelman, F., Neufeld, M.Y., Kramer, U., Fried,I. & Malach, R. (2013). Selectivity of audiovisualECoG responses revealed under naturalistic stimuliin the human cortex. Journal of Neurophysiology109, 2272–2281.

27. Tang, H., Buia, C., Madhavan, R., Crone, N.E.,Madsen, J.R., Anderson, W.S. & Kreiman, G. (2014).Spatiotemporal dynamics underlying object comple-tion in human ventral visual cortex. Neuron 83,736–748.

28. Dalal, S.S., Osipova, D., Bertrand, O. & Jerbi, K.(2013). Oscillatory activity of the human cerebellum:The intracranial electrocerebellogram revisited.Neuroscience & Biobehavioral Reviews 37, 585–593.

29. Johnson, E.L. & Knight, R.T. (2014). Intracranialrecordings and human memory. Current Opinion inNeurobiology 31C, 18–25.

30. Zhang, X., Lei, X., Wu, T. & Jiang, T. (2014). Areview of EEG and MEG for brainnetome research.Cognitive Neurodynamics 8, 87–98.

31. Fried, I., Rutishauser, U., Cerf, M. & Kreiman, G.(ed.) (2014). Single Neuron Studies of the HumanBrain: Probing Cognition, 408pp. Cambridge, MA,USA: The MIT Press.

32. Jacobs, J., Weidemann, C.T., Miller, J.F., Solway,A., Burke, J.F., Wei, X.X., Suthana, N., Sperling,M.R., Sharan, A.D., Fried, I. & Kahana, M.J. (2013).Direct recordings of grid-like neuronal activity inhuman spatial navigation. Nature Neuroscience 16,1188–1190.

33. Niediek, J. & Bain, J. (2014). Human single-unitrecordings reveal a link between place-cells andepisodic memory. Frontiers in Systems Neuroscience8, 158.

34. Quian Quiroga, R., Kraskov, A., Mormann, F., Fried,I. & Koch, C. (2014). Single-cell responses to faceadaptation in the human medial temporal lobe.Neuron 84, 363–369.

35. Rey, H.G., Fried, I. & Quian Quiroga, R. (2014).Timing of single-neuron and local field potentialresponses in the human medial temporal lobe.Current Biology 24, 299–304.

36. Rutishauser, U., Tudusciuc, O., Wang, S., Mamelak,A.N., Ross, I.B. & Adolphs, R. (2013). Single-neuroncorrelates of atypical face processing in autism.Neuron 80, 887–899.

37. Engel, A.K., Moll, C.K., Fried, I. & Ojemann, G.A.(2005). Invasive recordings from the human brain:Clinical insights and beyond. Nature ReviewsNeuroscience 6, 35–47.

38. Matsumoto, R., Nair, D.R., LaPresto, E., Bingaman,W., Shibasaki, H. & Luders, H.O. (2007). Functionalconnectivity in human cortical motor system: A cor-tico-cortical evoked potential study. Brain 130,181–197.

39. Keller, C.J., Honey, C.J., Megevand, P., Entz, L.,Ulbert, I. & Mehta, A.D. (2014). Mapping humanbrain networks with cortico-cortical evoked poten-tials. Philosophical Transactions of the Royal Societyof London Series B: Biological Sciences 369,20130528.

40. Desmurget, M., Song, Z., Mottolese, C. & Sirigu, A.


(2013). Re-establishing the merits of electrical brainstimulation. Trends in Cognitive Sciences 17,442–449.

41. SCHER (2009). The Need for Non-human Primatesin Biomedical Research, Production and Testing ofProducts and Devices, 38pp. Brussels, Belgium:Scientific Committee on Health and EnvironmentalRisks.

42. Ward, A.A. & Thomas, L.B. (1955). The electricalactivity of single units in the cerebral cortex of man.Electroencephalography & Clinical Neurophysiology7, 135–136.

43. Ojemann, G. (2014). Fifty-plus years of single neu-ron recordings: A personal perspective. In SingleNeuron Studies of the Human Brain: ProbingCognition (ed. I. Fried, U. Rutishauser, M. Cerf & G.Kreiman), Chapter 2, pp. 7–15. Cambridge, MA,USA: The MIT Press.

44. Quiroga, R.Q., Reddy, L., Kreiman, G., Koch, C. &Fried, I. (2005). Invariant visual representation bysingle neurons in the human brain. Nature, London435, 1102–1107.

45. Vannucci, M., Pezer, N., Helmstaedter, C., Schaller,K., Viggiano, M.P., Elger, C.E. & Grunwald, T.(2008). Hippocampal response to visual objects isrelated to visual memory functioning. Neuroreport19, 965–968.

46. Mamelak, A.N. (2014). Ethical and practical consid-erations for human microelectrode recording stud-ies. In Single Neuron Studies of the Human Brain:Probing Cognition (ed. I. Fried, U. Rutishauser, M.Cerf & G. Kreiman), Chapter 4, pp. 27–42. Cam -bridge, MA, USA: The MIT Press.

47. Fried, I. (2014). The neurosurgical theater of themind. In Single Neuron Studies of the HumanBrain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 3, pp.19–26. Cambridge, MA, USA: The MIT Press.

48. Rutishauser, U., Schuman, E.M. & Mamelak, A.N.(2014). Single neuron correlates of declarative mem-ory formation and retrieval in the human medialtemporal lobe. In Single Neuron Studies of theHuman Brain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 7, pp.101–120. Cambridge, MA, USA: The MIT Press.

49. Suthana, N. & Fried, I. (2014). Navigating our envi-ronment: Insights from single neuron recordings inthe human brain. In Single Neuron Studies of theHuman Brain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 9, pp.153–164. Cambridge, MA, USA: The MIT Press.

50. Nir, Y., Le Van Quyen, M., Tononi, G. & Staba, R.J.(2014). Microelectrode studies of human sleep. InSingle Neuron Studies of the Human Brain: ProbingCognition (ed. I. Fried, U. Rutishauser, M. Cerf & G.Kreiman), Chapter 10, pp. 165–188. Cambridge,MA, USA: The MIT Press.

51. Mormann, F., Ison, M.J., Quiroga, R.Q., Koch, C.,Fried, I. & Kreiman, G. (2014). Visual cognitiveadventures of single neurons in the human medialtemporal lobe. In Single Neuron Studies of theHuman Brain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 8, pp.121–150. Cambridge, MA, USA: The MIT Press.

52. Cerf, M., Gelbard-Sagiv, H. & Fried, I. (2014).Studying thoughts and deliberations using singleneuron recordings in humans. In Single NeuronStudies of the Human Brain: Probing Cognition (ed.I. Fried, U. Rutishauser, M. Cerf & G. Kreiman),

Chapter 11, pp. 189–204. Cambridge, MA, USA: TheMIT Press.

53. Patel, S.R., Sierra-Mercado, D., Martinez-Rubio, C.& Eskandar, E.N. (2014). Human single neuronreward processing in the basal ganglia and anteriorcingulate. In Single Neuron Studies of the HumanBrain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 12,pp. 205–228. Cambridge, MA, USA: The MIT Press.

54. Adolphs, R., Kawasaki, H., Tudusciuc, O., Howard,III, M., Heller, C., Sutherling, W., Philpott, L., Ross,I.B., Mamelak, A.N. & Rutishauser, U. (2014).Electrophysiological responses to faces in thehuman amygdala. In Single Neuron Studies of theHuman Brain: Probing Cognition (ed. I. Fried, U.Rutishauser, M. Cerf & G. Kreiman), Chapter 13,pp. 229–246. Cambridge, MA, USA: The MIT Press.

55. Ojemann, G. (2014). Human lateral temporal corti-cal single neuron activity during language, recentmemory, and learning. In Single Neuron Studies ofthe Human Brain: Probing Cognition (ed. I. Fried,U. Rutishauser, M. Cerf & G. Kreiman), Chapter 14,pp. 247–272. Cambridge, MA, USA: The MIT Press.

56. Bansal, A.K. (2014). Human single unit activity forreach and grasp motor prostheses. In Single NeuronStudies of the Human Brain: Probing Cognition (ed.I. Fried, U. Rutishauser, M. Cerf & G. Kreiman),Chapter 17, pp. 305–326. Cambridge, MA, USA: TheMIT Press.

57. Schulze-Bonhage, A. & Köling, R. (2014). Humansingle neuron recording as an approach to under-stand the neurophysiology of seizure generation. InSingle Neuron Studies of the Human Brain: ProbingCognition (ed. I. Fried, U. Rutishauser, M. Cerf & G.Kreiman), Chapter 18, pp. 327–344. Cambridge,MA, USA: The MIT Press.

58. Logothetis, N.K. (2008). What we can do and whatwe cannot do with fMRI. Nature, London 453,869–878.

59. KU Leuven (2014). Symposium and Debate. Non-human Primates in Biomedical Research: Science vs.Public Opinion? Leuven, The Neth er lands: KULeuven. Available at: http://alum.kuleuven. be/eng/agenda/symposium-and-debate-non-human-primates-in-beiomedical-research-science-vs-public-opinion (Accessed 25.01.16).

60. Logothetis, N.K. (2002). The neural basis of theblood-oxygen-level-dependent functional magneticresonance imaging signal. Philosophical Trans -actions of the Royal Society of London Series B:Biological Sciences 357, 1003–1037.

61. Max Planck Society (2014). Max Planck SocietyStatement: Statement dated 18 September 2014.Munich, Germany: Max Planck Society. Availableat: http://www.mpg.de/8412572/statement (Accessed25.01.16).

62. Bandettini, P.A. (2012). Twenty years of functionalMRI: The science and the stories. Neuroimage 62,575–588.

63. Logothetis, N.K., Pauls, J., Augath, M., Trinath, T.& Oeltermann, A. (2001). Neurophysiological inves-tigation of the basis of the fMRI signal. Nature,London 412, 150–157.

64. Boynton, G.M., Engel, S.A., Glover, G.H. & Heeger,D.J. (1996). Linear systems analysis of functionalmagnetic resonance imaging in human V1. Journalof Neuroscience 16, 4207–4221.

65. Mukamel, R. & Fried, I. (2012). Human intracranialrecordings and cognitive neuroscience. Annual


Review of Psychology 63, 511–537.66. Arthurs, O.J. & Boniface, S. (2002). How well do we

understand the neural origins of the fMRI BOLDsignal? Trends in Neurosciences 25, 27–31.

67. Kim, S.G. & Ogawa, S. (2012). Biophysical and phys-iological origins of blood oxygenation level-depend-ent fMRI signals. Journal of Cerebral Blood Flow &Metabolism 32, 1188–1206.

68. Boynton, G.M. (2011). Spikes, BOLD, attention, andawareness: A comparison of electrophysiological andfMRI signals in V1. Journal of Vision 11, 12.

69. Vazquez, A.L. & Noll, D.C. (1998). Nonlinear aspectsof the BOLD response in functional MRI. Neuro -image 7, 108–118.

70. Viswanathan, A. & Freeman, R.D. (2007).Neurometabolic coupling in cerebral cortex reflectssynaptic more than spiking activity. NatureNeuroscience 10, 1308–1312.

71. Mukamel, R., Gelbard, H., Arieli, A., Hasson, U.,Fried, I. & Malach, R. (2005). Coupling betweenneur onal firing, field potentials, and FMRI inhuman auditory cortex. Science, New York 309,951–954.

72. Ekstrom, A. (2010). How and when the fMRI BOLDsignal relates to underlying neural activity: Thedanger in dissociation. Brain Research Reviews 62,233–244.

73. Singh, K.D. (2012). Which “neural activity” do youmean? fMRI, MEG, oscillations and neurotransmit-ters. Neuroimage 62, 1121–1130.

74. Ojemann, G.A., Ojemann, J. & Ramsey, N.F. (2013).Relation between functional magnetic resonanceimaging (fMRI) and single neuron, local field poten-tial (LFP) and electrocorticography (ECoG) activityin human cortex. Frontiers in Human Neuroscience7, 34.

75. Bonmassar, G., Schwartz, D.P., Liu, A.K., Kwong,K.K., Dale, A.M. & Belliveau, J.W. (2001). Spatio -temporal brain imaging of visual-evoked activityusing interleaved EEG and fMRI recordings.Neuroimage 13, 1035–1043.

76. Steinbrink, J., Villringer, A., Kempf, F., Haux, D.,Boden, S. & Obrig, H. (2006). Illuminating theBOLD signal: Combined fMRI-fNIRS studies. Mag -netic Resonance Imaging 24, 495–505.

77. Arthurs, O.J., Williams, E.J., Carpenter, T.A.,Pickard, J.D. & Boniface, S.J. (2000). Linear cou-pling between functional magnetic resonance imag-ing and evoked potential amplitude in humansomatosensory cortex. Neuroscience 101, 803–806.

78. Arthurs, O.J. & Boniface, S.J. (2003). What aspect ofthe fMRI BOLD signal best reflects the underlyingelectrophysiology in human somatosensory cortex?Clinical Neurophysiology 114, 1203–1209.

79. Whittingstall, K., Wilson, D., Schmidt, M. &Stroink, G. (2008). Correspondence of visual evokedpotentials with FMRI signals in human visual cor-tex. Brain Topography 21, 86–92.

80. Buxton, R.B. (2013). The physics of functional mag-netic resonance imaging (fMRI). Reports on Progressin Physics 76, 096601.

81. Villringer, A. (2012). The intravascular susceptibil-ity effect and the underlying physiology of fMRI.Neuroimage 62, 995–999.

82. Mathiesen, C., Caesar, K., Akgoren, N. & Laurit -zen, M. (1998). Modification of activity-dependentincreases of cerebral blood flow by excitatory synap-tic activity and spikes in rat cerebellar cortex.Journal of Physiology 512, 555–566.

83. Carmichael, D.W., Vulliemoz, S., Rodionov, R.,Thornton, J.S., McEvoy, A.W. & Lemieux, L. (2012).Simultaneous intracranial EEG-fMRI in humans:Protocol considerations and data quality. Neuro -image 63, 301–309.

84. Siero, J.C., Hermes, D., Hoogduin, H., Luijten, P.R.,Petridou, N. & Ramsey, N.F. (2013). BOLD consis-tently matches electrophysiology in human sensori-motor cortex at increasing movement rates: Acombined 7T fMRI and ECoG study on neurovascu-lar coupling. Journal of Cerebral Blood Flow &Metabolism 33, 1448–1456.

85. Siero, J.C., Hermes, D., Hoogduin, H., Luijten, P.R.,Ramsey, N.F. & Petridou, N. (2014). BOLD matchesneuronal activity at the mm scale: a combined 7TfMRI and ECoG study in human sensorimotor cor-tex. Neuroimage 101, 177–184.

86. Jones, S.E., Zhang, M., Avitsian, R., Bhattacharyya,P., Bulacio, J., Cendes, F., Enatsu, R., Lowe, M.,Najm, I., Nair, D., Phillips, M. & Gonzalez-Martinez, J. (2014). Functional magnetic resonanceimaging networks induced by intracranial stimula-tion may help defining the epileptogenic zone. BrainConnectivity 4, 286–298.

87. Jones, S.E., Beall, E.B., Najm, I., Sakaie, K.E.,Phillips, M.D., Zhang, M. & Gonzalez-Martinez, J.A.(2014). Low consistency of four brain connectivitymeasures derived from intracranial electrode meas-urements. Frontiers in Neurology 5, 272.

88. Pereira, E.A., Green, A.L., Nandi, D. & Aziz, T.Z.(2008). Stereotactic neurosurgery in the UnitedKingdom: The hundred years from Horsley to Hariz.Neurosurgery 63, 594–606; discussion 606.

89. Moshe, B. & Understanding Animal Research(2014). Primates in Medical Research, 73pp. Avail -able at: https://itunes.apple.com/gb/book/primates-in-medical-research/id676974662?mt=11 (Accessed25.01.16).

90. Phillips, K.A., Bales, K.L., Capitanio, J.P., Conley,A., Czoty, P.W., ‘t Hart, B.A., Hopkins, W.D., Hu,S.L., Miller, L.A., Nader, M.A., Nathanielsz, P.W.,Rogers, J., Shively, C.A. & Voytko, M.L. (2014). Whyprimate models matter. American Journal ofPrimatology 76, 801–827.

91. Nuffield Council on Bioethics (2005). The Ethics ofResearch Involving Animals, 335pp. London, UK:Nuffield Council on Bioethics.

92. Verdier, J.M., Acquatella, I., Lautier, C., Devau, G.,Trouche, S., Lasbleiz, C. & Mestre-Francés, N.(2015). Lessons from the analysis of nonhuman pri-mates for understanding human aging and neurode-generative diseases. Frontiers in Neuroscience 9, 64.

93. Maxwell, M. (2009). Lies, Damned Lies and MonkeyScience. [Open letter to VERO (Voice for EthicalResearch at Oxford)]. Available at: http://www.vero.org.uk/Openletter.pdf (Accessed 25.01.16).

94. Greek, R. & Hansen, L.A. (2012). The developmentof deep brain stimulation for movement disorders.Journal of Clinical Research & Bioethics 3, 137.

95. Bailey, J. (2015). Letter to the Editor. ATLA 43,428–431.

96. Spatz, H. (1927). Physiologie und Pathologie derStammganglien. In Handbuch der normalen undpathologischen Physiologie (ed. A. Bethe, G. vBergman, G. Embden & A. Ellinger), pp. 318–412.Berlin, Germany: Springer.

97. Meyers, R. (1955). Parkinsonism, athetosis and bal-lism; a report of progress in surgical therapy.Postgraduate Medicine 17, 369–381.


98. Browder, J. (1947). Parkinsonism, is it a surgicalproblem? New York State Journal of Medicine 47,2589–2592.

99. Browder, J. (1948). Section of the fibers of the ante-rior limb of the internal capsule in parkinsonism.American Journal of Surgery 75, 264–268.

100. Gabriel, E.M. & Nashold, B.S. (1998). Evolution ofneuroablative surgery for involuntary movementdisorders: An historical review. Neurosurgery 42,575–590.

101. Albe-Fessard, D., Arfel, G., Guiot, G., Derome, P.,Dela, H., Korn, H., Hertzog, E., Vourch, G. &Aleonard, P. (1963). [Characteristic electric activit -ies of some cerebral structures in man]. Annales deChirurgie 17, 1185–1214.

102. Albe-Fessard, D., Arfel, G., Guiot, G., Derome, P.,Hertzog, E., Vourc’h, G., Brown, H., Aleonard, P., Dela Herran, J. & Trigo, J.C. (1966). Electro physiol -ogical studies of some deep cerebral structures inman. Journal of the Neurological Sciences 3, 37–51.

103. Albe-Fessard, D., Arfel, G., Guiot, G., Derome, P. &Guilbaud, G. (1967). Thalamic unit activity in man.Electroencephalography & Clinical NeurophysiologySuppl. 25, 132.

104. Gross, C.E., Boraud, T., Guehl, D., Bioulac, B. &Bezard, E. (1999). From experimentation to the sur-gical treatment of Parkinson’s disease: Prelude orsuite in basal ganglia research. Progress inNeurobiology 59, 509–532.

105. Rosenow, J., Das, K., Rovit, R.L. & Couldwell, W.T.(2002). Irving S. Cooper and his role in intracranialstimulation for movement disorders and epilepsy.Stereotactic & Functional Neurosurgery 78, 95–112.

106. Cooper, I.S. (1973). Effect of chronic stimulation ofanterior cerebellum on neurological disease. Lancet1, 206.

107. Cooper, I.S. (1973). Effect of stimulation of posteriorcerebellum on neurological disease. Lancet 1, 1321.

108. Medtronic (2015). Deep Brain Stimulation forMovement Disorders. Dublin, Ireland: Medtronic.Available at: http://professional.medtronic.com/pt/neuro/dbs-md/edu/about/#.VmGH3YSzWoc (Acc -essed 29.02.16).

109. Sarem-Aslani, A. & Mullett, K. (2011). Industrialperspective on deep brain stimulation: History, cur-rent state, and future developments. Frontiers inIntegrative Neuroscience 5, 46.

110. Cooper, I.S., Upton, A.R. & Amin, I. (1982). Chroniccerebellar stimulation (CCS) and deep brain stimu-lation (DBS) in involuntary movement disorders.Applied Neurophysiology 45, 209–217.

111. Brice, J. & McLellan, L. (1980). Suppression ofintention tremor by contingent deep-brain stimula-tion. Lancet 1, 1221–1222.

112. Siegfried, J. (1986). Effect of stimulation of the sen-sory nucleus of the thalamus on dyskinesia andspasticity. Revue Neurologique 142, 380–383.

113. Siegfried, J. (1987). Sensory thalamic neurostimula-tion for chronic pain. Pacing & ClinicalElectrophysiology 10, 209–212.

114. Benabid, A., Delong, M.R. & Hariz, M.I. (2015).Letter to the Editor. ATLA 43, 427–428.

115. Bailey, J. (2015). Letter to the Editor. ATLA 43,205–207.

116. Andy, O.J., Jurko, M.F. & Sias, F.R.J. (1963). Sub -thalamotomy in treatment of parkinsonian tremor.Journal of Neurosurgery 20, 860–870.

117. Alberts, W.W., Wright, E.W., Jr., Feinstein, B. & vonBonin, G. (1966). Experimental radiofrequency

brain lesion size as a function of physical parame-ters. Journal of Neurosurgery 25, 421–423.

118. Hassler, R., Mundinger, F. & Riechert, T. (1965).Correlations between clinical and autoptic findingsin stereotaxic operations of parkinsonism. Stereo -tactic & Functional Neurosurgery 26, 282–290.

119. Cleary, D.R., Raslan, A.M., Rubin, J.E., Bahgat, D.,Viswanathan, A., Heinricher, M.M. & Burchiel, K.J.(2013). Deep brain stimulation entrains local neu-ronal firing in human globus pallidus internus.Journal of Neurophysiology 109, 978–987.

120. Boraud, T., Bezard, E., Bioulac, B. & Gross, C.E.(2002). From single extracellular unit recording inexperimental and human Parkinsonism to thedevelopment of a functional concept of the roleplayed by the basal ganglia in motor control.Progress in Neurobiology 66, 265–283.

121. Christopher, L., Koshimori, Y., Lang, A.E., Criaud,M. & Strafella, A.P. (2014). Uncovering the role ofthe insula in non-motor symptoms of Parkinson’sdisease. Brain 137, 2143–2154.

122. Max Planck Society (2014). SynchronousOscillations in the Short-term Memory. Munich,Germany: Max Planck Society. Available at: http://www.mpg.de/8427680/Synchronous-oscillations_short-term-memory?filter_order=L&research_topic(Accessed 25.01.16).

123. Jacobs, J. & Kahana, M.J. (2010). Direct brainrecordings fuel advances in cognitive electrophysiol-ogy. Trends in Cognitive Sciences 14, 162–171.

124. Hanslmayr, S., Staudigl, T., Aslan, A. & Bauml,K.H. (2010). Theta oscillations predict the detrimen-tal effects of memory retrieval. Cognitive, Affective &Behavioral Neuroscience 10, 329–338.

125. Lega, B.C., Jacobs, J. & Kahana, M. (2012). Humanhippocampal theta oscillations and the formation ofepisodic memories. Hippocampus 22, 748–761.

126. Itthipuripat, S., Wessel, J.R. & Aron, A.R. (2013).Frontal theta is a signature of successful work-ing memory manipulation. Experimental BrainResearch 224, 255–262.

127. Jensen, O. & Tesche, C.D. (2002). Frontal thetaactivity in humans increases with memory load in aworking memory task. European Journal ofNeuroscience 15, 1395–1399.

128. Watrous, A.J., Fell, J., Ekstrom, A.D. & Axmacher,N. (2014). More than spikes: Common oscillatorymechanisms for content specific neural representa-tions during perception and memory. CurrentOpinion in Neurobiology 31C, 33–39.

129. Soteropoulos, D.S., Edgley, S.A. & Baker, S.N.(2013). Spinal commissural connections to motoneu-rons controlling the primate hand and wrist.Journal of Neuroscience 33, 9614–9625.

130. Soteropoulos, D.S., Williams, E.R. & Baker, S.N.(2012). Cells in the monkey ponto-medullary reticu-lar formation modulate their activity with slow fin-ger movements. Journal of Physiology 590,4011–4027.

131. Witham, C.L. & Baker, S.N. (2011). Modulation andtransmission of peripheral inputs in monkeycuneate and external cuneate nuclei. Journal ofNeurophysiology 106, 2764–2775.

132. Witham, C.L. & Baker, S.N. (2012). Coding of digitdisplacement by cell spiking and network oscilla-tions in the monkey sensorimotor cortex. Journal ofNeurophysiology 108, 3342–3352.

133. Fisher, K.M., Zaaimi, B. & Baker, S.N. (2012).Reticular formation responses to magnetic brain


stimulation of primary motor cortex. Journal ofPhysiology 590, 4045–4060.

134. Zaaimi, B., Edgley, S.A., Soteropoulos, D.S. &Baker, S.N. (2012). Changes in descending motorpathway connectivity after corticospinal tract lesionin macaque monkey. Brain 135, 2277–2289.

135. Soteropoulos, D.S., Edgley, S.A. & Baker, S.N.(2011). Lack of evidence for direct corticospinal con-tributions to control of the ipsilateral forelimb inmonkey. Journal of Neuroscience 31, 11,208–11,219.

136. Telenczuk, B., Baker, S.N., Herz, A.V. & Curio, G.(2011). High-frequency EEG covaries with spikeburst patterns detected in cortical neurons. Journalof Neurophysiology 105, 2951–2959.

137. Fedele, T., Scheer, H.J., Waterstraat, G., Telenczuk,B., Burghoff, M. & Curio, G. (2012). Towards non-invasive multi-unit spike recordings: Mapping 1kHzEEG signals over human somatosensory cortex.Clinical Neurophysiology 123, 2370–2376.

138. Broser, P.J. & Braun, C. (2012). Hydraulic drivenfast and precise nonmagnetic tactile stimulator forneurophysiological and MEG measurements. IEEETransactions on Bio-medical Engineering 59,2852–2858.

139. Blake, R. & Logothetis, N. (2002). Visual competi-tion. Nature Reviews Neuroscience 3, 13–21.

140. Kreiman, G., Fried, I. & Koch, C. (2002). Single-neu-ron correlates of subjective vision in the humanmedial temporal lobe. Proceedings of the NationalAcademy of Sciences of the USA 99, 8378–8383.

141. Institute of Medicine (2011). Chimpanzees inBiomedical and Behavioral Research: Assessing theNecessity, 190pp. Washington, DC, USA: NationalAcademies Press.

142. Bailey, J. (2011). Lessons from chimpanzee-basedresearch on human disease: The implications ofgenetic differences. ATLA 39, 527–540.

143. Marvanova, M., Menager, J., Bezard, E., Bontrop,R.E., Pradier, L. & Wong, G. (2003). Microarrayanalysis of nonhuman primates: Validation of exper-imental models in neurological disorders. FASEBJournal 17, 929–931.

144. Caceres, M., Lachuer, J., Zapala, M.A., Redmond,J.C., Kudo, L., Geschwind, D.H., Lockhart, D.J.,Preuss, T.M. & Barlow, C. (2003). Elevated geneexpression levels distinguish human from non-human primate brains. Proceedings of the NationalAcademy of Sciences of the USA 100, 13,030–13,035.

145. Arora, G., Polavarapu, N. & McDonald, J.F. (2009).Did natural selection for increased cognitive abilityin humans lead to an elevated risk of cancer?Medical Hypotheses 73, 453–456.

146. Nowick, K., Gernat, T., Almaas, E. & Stubbs, L.(2009). Differences in human and chimpanzee geneexpression patterns define an evolving network oftranscription factors in brain. Proceedings of theNational Academy of Sciences of the USA 106,22,358–22,363.

147. Konopka, G., Bomar, J.M., Winden, K., Coppola, G.,Jonsson, Z.O., Gao, F., Peng, S., Preuss, T.M.,Wohlschlegel, J.A. & Geschwind, D.H. (2009).Human-specific transcriptional regulation of CNSdevelopment genes by FOXP2. Nature, London 462,213–217.

148. Grosso, A.R., Gomes, A.Q., Barbosa-Morais, N.L.,Caldeira, S., Thorne, N.P., Grech, G., von Lindern,M. & Carmo-Fonseca, M. (2008). Tissue-specificsplicing factor gene expression signatures. NucleicAcids Research 36, 4823–4832.

149. Bailey, J. (2014). Monkey-based research on humandisease: The implications of genetic differences.ATLA 42, 287–317.

150. Niu, A.L., Wang, Y.Q., Zhang, H., Liao, C.H., Wang,J.K., Zhang, R., Che, J. & Su, B. (2011). Rapid evo-lution and copy number variation of primateRHOXF2, an X-linked homeobox gene involved inmale reproduction and possibly brain function. BMCEvolutionary Biology 11, 298.

151. Hahn, M.W., Demuth, J.P. & Han, S-G. (2007).Accelerated rate of gene gain and loss in primates.Genetics 177, 1941–1949.

152. Lin, L., Liu, S., Brockway, H., Seok, J., Jiang, P.,Wong, W.H. & Xing, Y. (2009). Using high-densityexon arrays to profile gene expression in closelyrelated species. Nucleic Acids Research 37, e90.

153. Hu, H.Y., Guo, S., Xi, J., Yan, Z., Fu, N., Zhang, X.,Menzel, C., Liang, H., Yang, H., Zhao, M., Zeng, R.,Chen, W., Pääbo, S. & Khaitovich, P. (2011).MicroRNA expression and regulation in human,chimpanzee, and macaque brains. PLoS Genetics 7,e1002327.

154. Somel, M., Liu, X., Tang, L., Yan, Z., Hu, H., Guo, S.,Jiang, X., Zhang, X., Xu, G., Xie, G., Li, N., Hu, Y.,Chen, W., Pääbo, S. & Khaitovich, P. (2011).MicroRNA-driven developmental remodeling in thebrain distinguishes humans from other primates.PLoS Biology 9, e1001214.

155. Hu, H.Y., He, L., Fominykh, K., Yan, Z., Guo, S.,Zhang, X., Taylor, M.S., Tang, L., Li, J., Liu, J.,Wang, W., Yu, H. & Khaitovich, P. (2012). Evolutionof the human-specific microRNA miR-941. NatureCommunications 3, 1145.

156. Paz-Yaacov, N., Levanon, E.Y., Nevo, E., Kinar, Y.,Harmelin, A., Jacob-Hirsch, J., Amariglio, N.,Eisenberg, E. & Rechavi, G. (2010). Adenosine-to-inosine RNA editing shapes transcriptome diversityin primates. Proceedings of the National Academy ofSciences of the USA 107, 12,174–12,179.

157. Li, L., Hu, X., Preuss, T.M., Glasser, M.F., Damen,F.W., Qiu, Y. & Rilling, J. (2013). Mapping putativehubs in human, chimpanzee and rhesus macaqueconnectomes via diffusion tractography. Neuro -image 80, 462–474.

158. Glasser, M.F., Goyal, M.S., Preuss, T.M., Raichle,M.E. & Van Essen, D.C. (2014). Trends and proper-ties of human cerebral cortex: Correlations with cor-tical myelin content. Neuroimage 93, Pt 2, 165–175.

159. Rilling, J.K. & Insel, T.R. (1999). The primate neo-cortex in comparative perspective using magneticresonance imaging. Journal of Human Evolution 37,191–223.

160. Wang, Y. & Leung, F.C.C. (2009). A study ongenomic distribution and sequence features ofhuman long inverted repeats reveals species-specificintronic inverted repeats. FEBS Journal 276,1986–1998.

161. Shi, L., Li, M., Lin, Q., Qi, X. & Su, B. (2013).Functional divergence of the brain-size regulatinggene MCPH1 during primate evolution and the ori-gin of humans. BMC Biology 11, 62.

162. Boyd, J.L., Skove, S.L., Rouanet, J.P., Pilaz, L.J.,Bepler, T., Gordan, R., Wray, G.A. & Silver, D.L.(2015). Human–chimpanzee differences in a FZD8enhancer alter cell-cycle dynamics in the developingneocortex. Current Biology 25, 722–729.

163. Wittkopp, P.J. (2007). Variable gene expression ineukaryotes: A network perspective. Journal ofExperimental Biology 210, 1567–1575.


164. Rilling, J.K. (2006). Human and nonhuman primatebrains: Are they allometrically scaled versions of thesame design? Evolutionary Anthropology 15, 65–77.

165. Preuss, T.M. (2000). Taking the measure of diver-sity: Comparative alternatives to the model-animalparadigm in cortical neuroscience. Brain, Behavior& Evolution 55, 287–299.

166. Rilling, J.K. (2014). Comparative primate neu-roimaging: Insights into human brain evolution.Trends in Cognitive Sciences 18, 46–55.

167. Butti, C. & Hof, P.R. (2010). The insular cortex: Acomparative perspective. Brain Structure &Function 214, 477–493.

168. Geschwind, D.H. & Rakic, P. (2013). Cortical evolu-tion: Judge the brain by its cover. Neuron 80,633–647.

169. Goulas, A., Bastiani, M., Bezgin, G., Uylings, H.B.,Roebroeck, A. & Stiers, P. (2014). Comparativeanalysis of the macroscale structural connectivity inthe macaque and human brain. PLoS Computa -tional Biology 10, e1003529.

170. Larsen, D.D. & Krubitzer, L. (2008). Genetic andepigenetic contributions to the cortical phenotype inmammals. Brain Research Bulletin 75, 391–397.

171. Rilling, J.K., Glasser, M.F., Preuss, T.M., Ma, X.,Zhao, T., Hu, X. & Behrens, T.E. (2008). The evolu-tion of the arcuate fasciculus revealed with compar-ative DTI. Nature Neuroscience 11, 426–428.

172. Mars, R.B., Neubert, F.X., Verhagen, L., Sallet, J.,Miller, K.L., Dunbar, R.I. & Barton, R.A. (2014).Primate comparative neuroscience using magneticresonance imaging: Promises and challenges.Frontiers in Neuroscience 8, 298.

173. Maze, I., Noh, K.M. & Allis, C.D. (2013). Histoneregulation in the CNS: Basic principles of epigeneticplasticity. Neuropsychopharmacology 38, 3–22.

174. Miranda-Dominguez, O., Mills, B.D., Grayson, D.,Woodall, A., Grant, K.A., Kroenke, C.D. & Fair, D.A.(2014). Bridging the gap between the human andmacaque connectome: A quantitative comparison ofglobal interspecies structure–function relationshipsand network topology. Journal of Neuroscience 34,5552–5563.

175. Thakkar, K.N., van den Heiligenberg, F.M.Z., Kahn,R.S. & Neggers, S.F.W. (2014). Frontal-subcorticalcircuits involved in reactive control and monitoringof gaze. Journal of Neuroscience 34, 8918–8929.

176. Ng, B.S., Logothetis, N.K. & Kayser, C. (2013). EEGphase patterns reflect the selectivity of neural firing.Cerebral Cortex 23, 389–398.

177. Vallender, E.J. (2011). Expanding whole exomeresequencing into non-human primates. GenomeBiology 12, R87.

178. Martin, C. (2014). Contributions and complexitiesfrom the use of in vivo animal models to improveunderstanding of human neuroimaging signals.Frontiers in Neuroscience 8, 211.

179. Masamoto, K. & Kanno, I. (2012). Anesthesia and

the quantitative evaluation of neurovascular cou-pling. Journal of Cerebral Blood Flow & Metabolism32, 1233–1247.

180. Engebretson, M. (2015). Thousands of Monkeys areBeing Bred for the Testing Industry in Mauritius.Available at: http://www.onegreenplanet.org/animalsandnature/monkeys-are-being-bred-for-the-testing-industry-in-mauritius/ (Accessed 25.01.16).

181. Pigarev, I.N., Saalmann, Y.B. & Vidyasagar, T.R.(2009). A minimally invasive and reversible systemfor chronic recordings from multiple brain sitesin macaque monkeys. Journal of NeuroscienceMethods 181, 151–158.

182. De Luna, P., Mohamed Mustafar, M.F.B. & Rainer,G. (2014). A MATLAB-based eye tracking controlsystem using non-invasive helmet head restraint inthe macaque. Journal of Neuroscience Methods 235,41–50.

183. Jennings, M., Prescott, M.J., Buchanan-Smith,H.M., Gamble, M.R., Gore, M., Hawkins, P.,Hubrecht, R., Hudson, S., Jennings, M., Keeley,J.R., Morris, K., Morton, D.B., Owen, S., Pearce,P.C., Prescott, M.J., Robb, D., Rumble, R.J., Wolfen -sohn, S. & Buist, D. (2009). Refinements in hus-bandry, care and common procedures for non-human primates: Ninth report of the BVAAWF/FRAME/RSPCA/UFAW Joint Working Group onRefinement. Laboratory Animals 43, Suppl. 1, 1–47.

184. Anon. (2007). NC3Rs Blood Sampling MicrositeLaunched. Laboratory Animals 41, 407.

185. Mason, J.W., Wool, M.S., Wherry, F.E., Pennington,L.L., Brady, J.V. & Beer, B. (1968). Plasma growthhormone response to avoidance sessions in the mon-key. Psychosomatic Medicine 30, Suppl., 760–773.

186. Roberts, R.A., Soames, A.R., James, N.H., Gill, J.H.& Wheeldon, E.B. (1995). Dosing-induced stresscauses hepatocyte apoptosis in rats primed by therodent nongenotoxic hepatocarcinogen cyproteroneacetate. Toxicology & Applied Pharmacology 135,192–199.

187. Brenner, G.J., Cohen, N., Ader, R. & Moynihan, J.A.(1990). Increased pulmonary metastases and natu-ral killer cell activity in mice following handling.Life Sciences 47, 1813–1819.

188. Reinhardt, V. & Reinhardt, A. (2000). Blood collec-tion procedure of laboratory primates: A neglectedvariable in biomedical research. Journal of AppliedAnimal Welfare Science 3, 321–333.

189. Rutishauser, U., Fried, I., Cerf, M. & Kreiman, G.(2014). The next ten years and beyond. In SingleNeuron Studies of the Human Brain: ProbingCognition (ed. I. Fried, U. Rutishauser, M. Cerf & G.Kreiman), Chapter 19, pp. 347–358. Cambridge,MA, USA: The MIT Press.

190. Murray, M.M. & Herrmann, C.S. (2013). Illusorycontours: A window onto the neurophysiology of con-structing perception. Trends in Cognitive Sciences17, 471–481.


Documents

Non-human Primates in Neuroscience Research: The Case … · project applications that we have seen. In addition, public information proffered by those that use NHPs in this way amounts