Neural correlates of deception: lying about past events ...Neural correlates of deception: lying about past events and personal beliefs Noa Ofen,1 Susan Whitfield-Gabrieli,2 Xiaoqian

Neural correlates of deception lying about past events

and personal beliefsNoa Ofen1 Susan Whitfield-Gabrieli2 Xiaoqian J Chai2

Rebecca F Schwarzlose13 and John D E Gabrieli2

1Department of Psychology Institute of Gerontology Wayne State University Detroit MI 48202 USA 2Brainand Cognitive Sciences Department The McGovern Institute for Brain Research MIT Cambridge MA 02139USA and 3Trends in Cognitive Sciences Cell Press Cambridge MA 02139 USA

Correspondence should be addressed to Noa Ofen Knapp Bldg 257 87 E Ferry St Detroit MI 48202 USA E-mail noaofenwayneedu

Abstract

Although a growing body of literature suggests that cognitive control processes are involved in deception much about theneural correlates of lying remains unknown In this study we tested whether brain activation associated with deception asmeasured by functional magnetic resonance imaging (fMRI) can be detected either in preparation for or during theexecution of a lie and whether they depend on the content of the lie We scanned participants while they lied or told thetruth about either their personal experiences (episodic memories) or personal beliefs Regions in the frontal and parietalcortex showed higher activation when participants lied compared with when they were telling the truth regardless ofwhether they were asked about their past experiences or opinions In contrast lie-related activation in the right temporalpole precuneus and the right amygdala differed by the content of the lie Preparing to lie activated parietal and frontal brainregions that were distinct from those activated while participants executed lies Our findings concur with previous reportson the involvement of frontal and parietal regions in deception but specify brain regions involved in the preparation vs exe-cution of deception and those involved in deceiving about experiences vs opinions

Key words deception fMRI episodic memory beliefs lie

Introduction

For reasons of both security and justice there is considerableinterest in applying neuroimaging methods to detect deceptionso that deception can be identified with increasing reliability(Vrij et al 2006 Sip et al 2008 Abe 2009) Delineating the brainbasis of lying vs truth-telling can also be used to inform a gen-eral understanding of cognitive and neural mechanismsinvolved in deception (Abe 2009 Christ et al 2009 Ganis et al2003 2009 Ganis 2015) Studies using neuroimaging methods todetect deception have underscored the notion that deception isa complex and cognitive demanding task These studies identi-fied several brain regions that show increased activation whenpeople lie compared with when they tell the truth (Langlebenet al 2002 Ganis et al 2003 Spence 2004 Kozel et al 2005Johnson et al 2007 Abe et al 2008 Lisofsky et al 2014)

Deception-related activations were identified in regions that areinvolved in cognitive control processes such as frontal and par-ietal cortex (Christ et al 2009) and regions that are involved inevaluating social context such as the superior temporal cortexand temporal poles (Lisofsky et al 2014) In recent years atten-tion has been directed to generate experimental designs thattake into account the intention of the person and the context inwhich the lie is executed (Sip et al 2008) In this report we sup-plement these efforts by providing evidence for differentiationin the neural correlates of deception based of the type of infor-mation one is asked to lie about We also further investigatewhether preparatory activation that is irrespective of the con-tent of the lie is related to the effectiveness by which a lie isbeing executed

The definition of the verb lsquoliersquo according to the Merriam-Webster dictionary is lsquoto make an untrue statement with intent

Received 21 September 2015 Revised 8 October 2016 Accepted 11 October 2016

VC The Authors (2016) Published by Oxford University Press For Permissions please email journalspermissionsoupcom

116

Social Cognitive and Affective Neuroscience 2017 116ndash127

doi 101093scannsw151Advance Access Publication Date 19 October 2016Original article

to deceiversquo Implicit in this definition are three facets of the actof lying First an individual must know or determine the truthin order to lie about it Depending on the nature of the lie thisstep could involve remembering specific details of an event orpast experiences or assessing onersquos opinion or emotional stateSecond a deceiver must then prepare an untrue statement andfinally respond by providing the statement that is contrary tohis or her determined truth On a cognitive level these last twosteps may involve cognitive control so as to inhibit the truthfulresponse or manipulate the true information to convert it into alie Indeed a meta-analysis conducted by Christ et al (2009) sug-gest remarkable overlap between region involved in executivecontrol and working memory and this notion is supported byadditional empirical work (Lisofsky et al 2014)

Much of the prior neuroimaging work on deception hasfocused on how participants lie about episodic knowledge orthe detailed memory of past experiences Detecting memory-related activation is critical for approaches to lie detection whenan individual is presumably concealing personally available in-formation about the past Deception regarding past experiencesmay specifically involve brain regions supporting retrieval andmaintenance of episodic memories such as regions in the med-ial temporal lobes and prefrontal cortex (Rissman et al 2010)While this type of deception may be most relevant for lie-detection in a criminal setting it is not the only kind of lie a per-son can tell For instance other work has focused on identifyingthe neural correlates of peoplersquos choices to lie about their per-sonal opinions based on the social pressures and norms of thosearound them (Falkiewicz et al 2015 Volz et al 2015) It is notwell understood whether the same or a different neural circuitis engaged to deceive about memory for an experience vs amoral belief or opinion Regions associated with moral reason-ing include frontal and parietal cortex medial frontal cortex aswell as regions in the superior temporal sulcus and the tem-poral parietal junction (Pascual et al 2013)

Prior neuroimaging work on deception has also focused ondetecting differential brain responses during the execution of alie Although it is likely that the most distinctive signature oflying will occur during its execution it may also be possible todetect differential activation when one prepares to lie comparedto preparing to tell the truth To our knowledge only one pub-lished report has examined the neural correlates of preparing tolie (Ito et al 2012) That study did not find significant differencesin brain activation between preparing to tell the truth and pre-paring to deceive yet it did identify regions that were overallmore active during a preparation to either tell the truth or lie Wehypothesized that there are different neural circuits engaged forthe preparation compared with the execution of deception andthat there would be individual differences in the magnitude oflsquopreparatoryrsquo lie activation that would be related to individualsrsquobehavior when delivering a lie

Thus in this study we aimed to further characterize the neuralmechanisms of deception by examining two fundamental ques-tions about the brain basis of deception First we asked if there aredifferent neural circuits engaged for deception about personal ex-periences (episodic memory) or opinions Second we investigatedif there are different neural circuits engaged for the preparation ofdeception vs the execution of deception and whether individualdifferences in the magnitude of lsquopreparatoryrsquo lie activation wouldbe related to individualsrsquo behavior when delivering a lie

We asked people to either tell the truth or lie about eitherexperiences or opinions Participants were asked yesno ques-tions about two distinct types of personal knowledge episodicexperiences and beliefs or opinions For episodic knowledge

questions probed the recollection of specific but commonplaceevents For beliefs or opinions questions focused on moral orsocietal topics about which most participants would have anopinion Unlike many other studies of deception (Langlebenet al 2002 2005 Davatzikos et al 2005) each question wasunique and did not repeat during the scan session This was im-portant so that response to any of the questions could not berehearsed or directly influenced by prior responses to the samequestion during the scan session We employed a second aspectof the design aimed to contextualize the experimental task in amore ecologically relevant setting Participants were told thatwe would be monitoring their brain activation and that theyshould try to lie in a way that we could not detect by looking atthe images of their brain These instructions were intended toput the participantsrsquo deception in a social context a factor thatis known to affect the process of lying (Abe et al 2007 Sip et al2012 Lisofsky et al 2014)

We predicted that regions previously identified in a meta-analysis of deception activation (Christ et al 2009) likely supportmore common aspects of deception and thus would similar pro-file of activation during deception regardless of the type of ques-tions one is lying about We also hypothesized that otherregions would show differential activation when lying about epi-sodic questions compared to lying about opinions Specificallyregions supporting episodic memory would uniquely supportlying about episodic content whereas regions associated withprocessing of beliefs or moral judgments would specificallysupport lying about beliefs or opinions Finally we hypothesizedthat preparing to lie will engage regions involved in voluntary al-location of attention (Ito et al 2012) and that regions that are re-cruited during the preparation period would have directinfluence on behavioral patterns during the execution of a lie

MethodsParticipants

Eighteen volunteers (nine males mean agefrac14 197 6 10 yearsrange 18ndash21 years) were recruited from the Stanford Universitycommunity All participants were right-handed had normalvisual acuity and were screened for a history of psychiatric ormedical illnesses Participants were paid $20 per hour for theirtime and gave informed consent in accordance with the guide-lines of the Stanford Medical Human Subjects Committee

Materials

One hundred and thirty yesno questions were created for theexperiment Half of the questions were designed to tap episodicknowledge (episodic questions mean lengthfrac14 76 6 15 range4ndash11 words mean propositional idea density calculated withSPIDRVC frac14 04 6 01 range 03ndash07) Examples include Have you everridden a horse Did you watch the last Super Bowl The other half ofthe questions assessed personal beliefs and opinions (beliefquestions mean lengthfrac14 69 6 17 range 4ndash11 words mean prop-ositional idea densityfrac14 04 6 01 range 02ndash07) eg Should samesex marriage be legal Do you believe the death penalty is justifiableThe assignment of questions to Lie or True condition was fullycounterbalanced across participants (see below) Episodic ques-tions were constructed with the intention to elicit a recollectionof a specific episode and a time frame was added such that therewill be roughly an equal distribution of yes and no answersBelief questions were constructed to elicit reflection on a per-sonal belief or opinion such that there would be a roughly equal

N Ofen et al | 117

proportion of agreement and disagreement with the statementsImportantly to minimize the confounding effects of specificresponse-matching patterns and the remapping of practiced re-sponses no items were repeated during the scan session

Procedure

Participants were given 10 practice trials prior to entering theMR scanner Inside the scanner 120 trials were presented in 4blocks of 30 trials each Each trial began with an instruction cueword (lsquoLiersquo or lsquoTruersquo) presented in the center of the screen for 2 sfollowed by a fixation cross that was presented for a variableduration between 3 and 5 s Participants were instructed to usethe cue and the fixation periods to prepare to either tell thetruth or lie in accordance with the cue word A yesno questionwas then presented on the screen for 4 s followed Each trialwas followed by a blank screen presented for a variable durationbetween 9 and 11 s to complete a total of 20 s total between thestart of each trial Participants were instructed to respond to thequestion as accurately and as rapidly as possible Half of the tri-als were cued with lsquoLiersquo and the other half were cued withlsquoTruersquo The specific assignment of each question to lsquoLiersquo orlsquoTruersquo conditions was counter-balanced across participants Thelocation of the lsquoyesrsquo and lsquonorsquo response buttons was indicated atthe bottom of the screen and randomized across trials to pre-vent simple mapping of the response options Sixty episodicand 60 belief questions were presented in a pseudo-randomizedorder Responses (yesno) were recorded and response timeswere used to calculate each individualrsquos deception cost score[deception scorefrac14 100 (mean reaction time to lie meanreaction time to respond truthfully)mean reaction time torespond truthfully] that was used as an indirect measure of theeffort of lying across subjects

Immediately following the scanning session participantswere presented with all 120 questions and asked to give the trueresponses Inconsistencies between responses in the scannerand the post-scan review were used to define error trials thatwere eliminated from analyses Participants were asked to indi-cate the strength of their belief for each belief question (strongor weak) Participants were asked to indicate whether duringtheir response for each of the episodic questions they recol-lected a specific detail of an episode (remember) or whetherthey lsquosimply knewrsquo the answer but did not recollect any specificdetails of a specific episode when answering the question(know) This procedure follows rememberknow procedure usedin memory studies to gain information about subjective mem-ory phenomenology Participants also filled out a brief question-naire about their performance on the task and completed twopersonality questionnaires (NEO FFI REF and PsychopathyPersonality Inventory PPI Lilienfeld and Andrews 1996) Thesequestionnaires are not discussed further

Image acquisition

Scanning was conducted with a 30T GE Signa scanner (GeneralElectric Milwaukee WI USA) using a custom-built volume headcoil Head movement was minimized using a bite bar formedwith the subjectrsquos dental impression T1-weighted whole-brainanatomical images (256 256 voxels 086 mm in plane reso-lution 12 mm slice thickness) were collected for the purpose ofcreating a subject-specific mask of the functional dataFunctional images were obtained from 24 slices aligned to theanteriorndashposterior commissure covering the entire brain using aspiral inout T2 pulse sequence (Glover and Law 2001) (TRfrac14 2 s

TEfrac14 30 ms 60 flip angle 64 64 voxels 375 mm in plane reso-lution 6mm slice thickness with no slice skip) Three hundredand four volumes were acquired per functional run the firstfour images were discarded to allow signal stabilization

Data quality assessment and preprocessing

Data were visually inspected and reviewed for artifacts and mo-tion using custom software (httpwebmiteduswgsoftwarehtm) Functional data were subjected to artifact detectionimages were defined as artifactual if movement between con-secutive acquisitions exceeded 05 mm in any direction or if theaverage image intensity was more than 3 sd away from themean global intensity of the run Data from one participantwere excluded because of overall large standard deviation inthe global intensity leading to sustained image artifacts SPM2(Welcome Department of Imaging Neuroscience London UKhttpwwwfilionuclacukspmspm2html) was used for allanalyses Images were corrected for motion using sinc interpol-ation and adjusted for the acquisition time (to the middle)functional images were then spatially normalized based on par-ameters determined by normalizing the mean functional imageto the T1 Montreal Neurological Institute template (MNI)Finally images were spatially smoothed with an isotropicGaussian kernel of 6-mm full width at half maximum

Statistical analysis

First-level general linear model-based analyses were conductedin MNI space Models included regressors of interest generatedby convolving task events with a canonical model of the HRF asimplemented in SPM2 The preparation period (5ndash7 s) and thestimuli (4 s) were modeled separately by regressor functionsThis resulted in six regressors of interest two preparationregressors (preparation-lie preparation-true) four stimuliregressors (belief-lie belief-true episodic-lie episodic-true) foreach of the four runs We used only two regressors for the prep-aration period because the specific question type was not avail-able for the participants during this period Cue presentationerror trials motion parameters (three rotation and three transla-tion parameters) and individual artifact images were also mod-eled First-level model estimation was done using an explicitmask created by combining grey and white segments of sub-jectsrsquo high-resolution anatomical brain images Linear combin-ations of regressors were used to define contrasts of interest(i) belief-liegtbelief-true (ii) episodic-liegt episodic-true (iii)preparation-liegtpreparation-true Contrasts constructed at thefirst level were then input into a second-level group analysisusing a random-effects model Group level activation maps werecomputed using a one-sample t-test Regions that showeddeception-related activations (liegt true) for both episodic and beliefquestions were determined by conjunction analysis of groupmaps for (i) belief-liegtbelief-true and (ii) episodic-liegt epi-sodic-true (each map uncorrected voxel-level Plt 0005 clus-tergt 100 contiguous voxels) Regions in which deception effectsdiffered by question type were identified by paired t-test across allparticipants using (i) belief-liegtbelief-true and (ii) episodic-liegt episodic-true (uncorrected voxel-level Plt 0005 clus-tergt 100 contiguous voxels) Regression analysis was used toexamine the relation between activations during lie preparationand deception cost score as continuous variable across subjectsAll reported clusters survived uncorrected voxel-level P thresh-old of 0005 and consisted of 100 or more contiguous voxels

118 | Social Cognitive and Affective Neuroscience 2017 Vol 12 No 1

(well within the recommended threshold by Lieberman andCunningham 2009)

In addition to clusters identified from the grouprsquos activationmaps we examined deception effects in a set of regions ofinterest (ROIs) identified in a meta-analysis of deception effects(Christ et al 2009) These literature-based ROIs were createdusing a 6-mm sphere around the peak voxel of clusters reportedin the meta-analysis Mean individual subjectsrsquo contrast valuesfor task conditions were extracted from clusters identified inthese analyses and from literature based ROIs We enteredthese values into a 2 2 repeated measures analysis (deceptionlietrue question type episodicbelief) and performed follow-upt-tests with all reported effects significant at P thresholds of005

To examine effects of participantsrsquo post-scan ratings of thequestions (lsquorememberrsquo or rsquoknowrsquo for episodic questions andlsquostrongrsquo or rsquoweakrsquo for belief questions) we created an additionalfirst-level general linear model with eight stimuli regressors(episodic-remember-lie episodic-know-lie belief-strong-liebelief-weak-lie episodic-remember-true episodic-know-truebelief-strong-true belief-weak-true) With the exception of thenumber of stimuli regressors this model was identical to themodel used in the main analysis described above Mean individ-ual subjectsrsquo contrast values for eight stimuli conditions wereextracted from the clusters and ROIs defined above We entered

these values into two separate 2 2 repeated measures ana-lyses (deception lietrue episodic rememberknow deceptionlietrue belief strongweak)

ResultsBehavior

Accuracy was measured by comparing the answers given in thescanner to the honest answers in the post-scan reviewAccuracy was high overall (830 6 82 mean 6 SD) and did notvary by condition [lie 816 6 98 truth 844 6 86 t(16)frac14 019ns] or by the type of question participants were asked [episodic837 6 90 belief 823 6 93 t(16)frac14 032 ns] Reaction timesacross conditions were compared using a repeated-measuresanalysis (2 2 model deception lietrue question type epi-sodicbelief Figure 1) Participants were slower when lying rela-tive to telling the truth [deception cost effect F(116)frac14 86Plt 001] There was a trend to answer belief questions moreslowly than episodic questions [question type F(116)frac14 43Pfrac14 006] and there was a significant deception by question typeinteraction [F(116)frac14 98 Pfrac14 0007] Follow-up comparisons con-firmed that participants were slower when lying relative to tell-ing the truth about their personal beliefs [t(16)frac14 49 Plt 0001] butnot for episodic memories Furthermore reaction times did not

A

B

Fig 1 Paradigm and behavioral results (A) Example of a trial in each of the four experimental conditions belief-true belief-lie episodic-true and episodic-lie Each trial

began with a LIETRUE cue presented on the screen for 2 s followed by a jittered time for preparation 3ndash5 s followed by the question presented for 4 s and a jittered in-

ter-trial-interval ranging between 9 and 11 s Trial timing information is depicted in seconds (B) Mean reaction times across participants for responses in the four ex-

perimental conditions and in the eight conditions defined by post-scan ratings Bars represent standard error

N Ofen et al | 119

differ between episodic and belief questions in the truthconditions

In post-scan ratings participants indicated that 486 6 90of the episodic questions evoked a specific memory (lsquorememberrsquoas opposed to lsquoknowrsquo) and 657 6 62 of the belief questionstapped a personal belief that participants rated as lsquostrongrsquo (asopposed to lsquoweakrsquo) Two participants had fewer than 10 re-sponses in one of the post-scan rating categories and were thusomitted from further analyses that included the post-scan rat-ings A 2 2 repeated-measures analysis was conducted separ-ately for episodic (deception lietrue memory type rememberknow) and belief (deception lietrue belief strength strongweak) questions For episodic questions there was no deceptioncost but there was a trend for a main effect of memory type[F(114)frac14 43 Pfrac14 006] and an interaction of deception cost bymemory type [F(114)frac14 46 Plt 005] indicating that participantswere slower when lying relative to telling the truth only for epi-sodic questions later rated as lsquorememberrsquo For belief questionsthere were main effects of deception cost [F(114)frac14 190Plt 0001] as well as a main effect of belief strength [F(114)frac14 116Pfrac14 0004] and no interaction of deception cost by beliefstrength indicating that participants were overall slower whenlying relative to telling the truth and overall slower in respond-ing about personal beliefs rated as lsquoweakrsquo compared to thoserated as lsquostrongrsquo

Imaging

Common regions for lying about episodic and belief questionsRegions that showed deception-related activations (liegt true)for both episodic and belief questions were determined by con-junction analysis (group maps for belief-liegtbelief-true andepisodic-liegt episodic-true each map uncorrected voxel-levelPlt 0005 clustergt 100 contiguous voxels) Bilateral lateral par-ietal and bilateral medial superior frontal regions were activatedmore when responding with a lie compared to giving an honestresponse for both episodic and belief questions as determinedby a conjunction analysis (Figure 2 and Table 1A) In all clusters

identified by this conjunction analysis there was a main effectof deception [F(116)gt 131 Plt 0003] with no interaction of de-ception by question type [F(116)lt 29 Pgt 011] We examined ac-tivation in these regions based on the participantsrsquo post-scanratings of questions by either the episodic memory type itevoked (lsquorememberrsquo vs lsquoknowrsquo) or the strength of the personalbelief it addressed (lsquostrongrsquo vs lsquoweakrsquo) Both analyses revealed amain effect of deception without interactions between eithermemory type [Frsquos(116)lt 27 Prsquosgt 012] or belief strength[Frsquos(114)lt 31 Prsquosgt 01] In sum bilateral lateral parietal andmedial superior frontal regions were consistently activatedwhen a participant lied (compared with truth-telling) irrespect-ive of the type of information they lied about

Distinct regions for lying about episodic or belief questions Weconducted a second analysis to identify regions in which decep-tion effects differed by question type (paired t-test belief-liegt be-lief-true and episodic-liegt episodic-true uncorrected voxel-levelPlt 0005 clustergt 100 contiguous voxels) We identified regionsin the right temporal pole precuneus right amygdala and rightprecentral gyrus (Figure 3 and Table 1B) These regions were usedas functional ROIs to further investigate the nature of the activa-tion modulation by question type Activation in these regionswas extracted across conditions and a significant interaction ofdeception by question type was confirmed [Frsquos(116)gt 153Prsquoslt 0002] Below we present the findings across conditions andquestion types separately in each of these regions

Activation in the right temporal pole seemed to reflect lyingabout episodic content however it is possible that differentialeffect for lying about personal beliefs was masked by this regionbeing actively engaged in all conditions relating to personal be-liefs This was evident by a main effect of deception (liegt true)for episodic questions [F(116)frac14 76 Pfrac14 001] that did not differ bythe memory type remember vs know [F(116)frac14 30 Pfrac14 010] Incontrast activation in this region was only marginally related tolying vs responding truthfully about personal beliefs [F(114)frac14 37Pfrac14 008] Prior studies have implicated the temporal pole in rea-soning about belief-laden material (Goel and Dolan 2003

440

58

-2

-15

-1

-05

0

05

1

episodic belief -2

-15

-1

-05

0

05

1

episodic belief

0

05

1

15

2

25

3

35

episodic belief

A B

C

A B

C

Fig 2 Deception effects common to episodic and belief questions Activation maps (conjunction of activation maps for the contrasts episodic liegt true and belief

liegt true questions) are rendered on standard brain horizontal sections (middle) MNI coordinates are presented at the bottom of each section Group mean param-

eter estimates for lie and true activation in episodic and belief questions (arbitrary units) (A) Left parietal cortex (B) right parietal cortex (C) superior frontal gyrus

Bars represent standard error

120 | Social Cognitive and Affective Neuroscience 2017 Vol 12 No 1

Cunningham et al 2004) Thus we tested whether the right tem-poral pole activations seen in the present study were higheroverall for belief compared with episodic questions Activationsin this region were higher for belief compared to episodic ques-tions [as evidenced by a main effect for question typeF(116)frac14 76 Pfrac14 001] Furthermore in line with a role for this re-gion in the processing of belief-laden material right temporalpole activation was greater overall for beliefs judged as lsquostrongrsquocompared with those judged as lsquoweakrsquo [F(114)frac14 57 Pfrac14 003] Wealso found a trend towards an interaction between deceptionand belief strength in this region [F(114)frac14 36 Pfrac14 008] with alarger effect of deception for weak beliefs

In contrast in the precuneus there was higher liegt truth activa-tion when lying for personal beliefs This was evident by a maineffect of deception for belief questions [F(114)frac14 122 Pfrac14 0004] thatdid not differ by belief strength [F(114)frac14 01 Pfrac14 091] In contrastactivation in this region showed no effect of deception for episodicquestions [F(116)frac14 01 Pfrac14 076] and no interaction of deception forepisodic questions by memory [F(116) 01 Pfrac14 072]

An anterior medial temporal lobe region that included the rightamygdala showed reduced activation for all experimental con-ditions compared with baseline [trsquos(16)gt 23 Prsquoslt 004] howeverthe pattern of activation across these conditions revealed a de-ception effect for episodic questions Specifically we identifieda main effect of deception for episodic questions [F(116)frac14 80Pfrac14 001] that did not differ by memory type [F(116)frac14 20Pfrac14 017] There was no main effect for deception for belief ques-tions [F(114)frac14 14 Pfrac14 025] or modulation of deception beliefstrength [F(114)frac14 19 Pfrac14 019] however activation in this regionwas marginally higher (closer to baseline) for lsquostrongrsquo comparedwith lsquoweakrsquo beliefs [F(114)frac14 45 Pfrac14 005]

Deception activations in a priori defined ROIs We tested deceptioneffects in a set of 11 literature-based ROIs identified in a meta-analysis of deception studies (Christ et al 2009) In the majorityof these regions we identified a main effect of deception[Frsquos(116)gt 57 Plt 003] with no interaction of deception by ques-tion type [Frsquos(116)lt 27 Prsquosgt 02] These included bilateral lateralparietal cortex bilateral inferior frontal gyrus and bilateral an-terior prefrontal cortex Moreover there was no interaction ofdeception and either episodic memory type [Frsquos(116)lt 38

Prsquosgt 007] or belief strength [F(114)rsquoslt 13 Prsquosgt 027] in these re-gions further supporting the general role of these regions indeception

Activation in the left and right insula ROIs also identifiedmain effects of deception [left F(116)frac14 50 Pfrac14 004 rightF(116)frac14 65 Pfrac14 002] however these effects differed by questiontype [left F(116)frac14 57 Pfrac14 003 trend in the right F(116)frac14 39Pfrac14 006] Follow-up examinations showed deception effectswhen subjects responded to belief questions [t(16)gt 29 Plt 001]but not to episodic questions [jtj(16)lt 16 Pgt 014] Deception ef-fects in the insula for belief questions did not differ by beliefstrength [Frsquos(114)lt 025 Prsquosgt 063]

In three of the ROIs identified in the meta-analysis we wereunable to identify a deception effect or an interaction of decep-tion by question type These include the anterior cingulate cortexright intraparietal sulcus and an anterior right insula region [de-ception Frsquos(116)lt 21 Prsquosgt 01 interaction Frsquos(116)lt 01 Prsquosgt 07]

Preparing to lie Brain activation in bilateral parietal and occipi-tal regions was greater when participants prepared to lie com-pared with when they prepared to tell the truth (contrastpreparation-liegtpreparation-true) (Figure 4A and Table 2A)None of these regions showed deception effects when partici-pants responded to the questions [Frsquos(116)lt 19 Prsquosgt 02] Thesefindings suggest that the brain regions recruited when a personis preparing to lie are distinct from those engaged when the per-son generates and delivers the specific lie

Lie preparation effects in a priori defined ROIs Of all the ROIsdefined based on the deception-related meta-analysis (Christet al 2009) only the anterior right insula showed differential ac-tivation during the period of lie preparation [t(16)frac14 23 Pfrac14 003other ROIs jtjrsquos(16)lt 11 Prsquosgt 03] This ROI was one of the threethat did not show a deception effect during the response periodsuggesting that this regionrsquos involvement in lie preparation con-tributed to its identification in the meta-analysis

Activations during lie preparation linked to deception cost score Weinvestigated the behavioral correlates of regional brain activa-tions during the preparation to lie (preparation-liegtprepar-ation-true) by assessing the relation of activations during the

Table 1 Common (A) and distinct (B) activations for lying (lie gt true) about episodic or belief questions

BA x y z No voxels

(A) Deception effects common to episodic and belief questionsndashconjunction analysisL Supramarginal gyrus 39 62 52 28 1030

Inferior parietal lobule 40 54 54 30Superior temporal gyrus 22 58 62 26

R Supramarginal gyrus 39 56 48 30 543Inferior parietal lobule 40 56 42 44

R Superior frontal gyrus 6832 14 12 58 405L Superior frontal gyrus 6 10 14 58

(B) Interaction of deception effects by question typeT

R Precentral gyrus 64 46 10 42 524 221R Middlesuperior temporal gyrus 21 54 4 40 440 168

38 46 12 30 420R Globus pallidum na 26 18 2 397 234

Amygdala na 30 2 14 368RL Precuneuscuneus 7 6 78 40 425 235

Notes Peak coordinates (x y z) are based on MNI brain BA Brodmannrsquos area L left R right

N Ofen et al | 121

preparation to lie to deception cost scores [calculated as 100 3

(mean reaction time to lie ndash mean reaction time to respondtruthfully)mean reaction time to respond truthfully] (Figure 4Band Table 2B) Across participants the magnitudes of activa-tions during the preparation to lie in the posterior cingulate andin the left temporal and frontal cortices were correlated withlower deception cost scores (Figure 4C see additional informa-tion provided in Supplementary data Figure S2) This meansthat participants who recruited these brain areas more whenpreparing to lie than when preparing to tell the truth alsotended to show less deception cost suggesting that recruitmentof these regions during lie preparation was linked to lying moreefficiently Across all participants activations in these regionswere not different in preparation for lie compared with trueconditions [jtjrsquos(16)lt 20 Prsquosgt 007] Moreover across all

participants in only one of these regions the left dorsolateralprefrontal cortex (BA 46) there was a significant deception ef-fect [increased activation for lie than truth during the responseF(116)frac14 62 Pfrac14 003] The deception effect in the left dorsolateralprefrontal cortex did not differ by question type [F(116)frac14 16Pfrac14 022]

Distinct regions for responding truthfully to episodic or beliefquestions Although not the focus of this report the present ex-periment allowed us to compare brain activations associatedwith two distinct types of personal knowledge We found strik-ing dissociations in the brain regions involved in respondingtruthfully to questions about personal episodic knowledge com-pared to personal beliefs and opinions (described in detail inSupplementary materials)

554 0

02

04

06

08

1

12

14

episodic belief

-16 -14 -12

-1 -08 -06 -04 -02

0

episodic belief

-15

-1

-05

0

05

1

15

episodic belief 40

4

A

A

C

B

Fig 3 Distinct deception effects for episodic or belief questions Activation maps (paired t-test of activation maps for contrasts episodic liegt true from and belief liegt -

true questions) are rendered on standard brain sagittal and horizontal sections (left) MNI coordinates are presented at the bottom of each section Group mean param-

eter estimates for lie and true activation in episodic and belief questions (arbitrary units) (A) right temporal pole (B) precuneus (C) right amygdala Bars represent

standard error

122 | Social Cognitive and Affective Neuroscience 2017 Vol 12 No 1

Discussion

Using a novel paradigm to assess the neural correlates of decep-tion we identified three major findings First consistent withprevious reports (Christ et al 2009 Lisofsky et al 2014) wefound that frontal and parietal regions are engaged during theexecution of a deceptive response Frontal and parietal brain re-gions support processes related to cognitive control and execu-tive functioning Thus our findings are consistent with theinvolvement of cognitive control and executive functioning indeception Second we identified regions that showed adifferential deception contrast (LiegtTruth) when lying aboutepisodic knowledge or personal beliefs suggesting thatdeception-related neural correlates are sensitive to the contentone is lying about (Greene and Paxton 2009) Finally we identi-fied regions that may be involved in lsquopreparatoryrsquo processes andwhose engagement accounts for some individual differences in

behavioral measures of the lie response (Ito et al 2012) Takentogether these findings further elucidate the neural correlatesof several important aspects of deception and highlight the im-portance of considering the content one is lying about as well asthe processes involved in preparing to lie as those aspects arelinked to unique neural correlates

In an act of deception one allegedly first determines thetruth and then engages cognitive effort in suppressing the urgeto make a truthful response (Spence 2004) This process iswidely referred to as response inhibition and it occupies a cen-tral role in the mental toolkit known as cognitive control Weobserved a deception cost-effect in participant response timessuch that lie responses were overall slower compared with truthresponses This deception cost-effect suggests that participantsin this experiment either employed more cognitive control orperformed additional computations (eg response reversal)when lying compared with telling the truth

Although there was an overall deception effect in responsetimes interestingly this effect was almost entirely driven by theslower deceptive responses when asked about personal beliefswhereas participants responded equally quickly when lying ortelling the truth about past personal experiences Interpretingthis differential behavioral effect in the context of the broaderdeception literature is difficult Although most studies of decep-tion find that lying is slower than truth telling this is not thecase for all types of lies Faster responses for lies are sometimesreported (Langleben et al 2005) and may reflect a more auto-matic repeated and rehearsed lie response In this study it ispossible that the participants were less invested in lies aboutepisodic memory or that these questions were simply less cog-nitively complex compared with belief questions and thereforealtering the response was not accompanied by a noticeable in-crease in reaction time In the current design each questionwas presented only once eliminating the possibility of an auto-matic lie response Under these circumstances truth should bethe more automatic and thus faster response when consider-ing either personal experiences or beliefs The differential be-havioral effect may reflect a differential ability to monitorcognitive effort when lying about personal experiences vs be-liefs The participants in our study were asked to try to lie insuch a way that the experimenter would not be able to tellwhether they were lying or telling the truth In informal inter-views conducted after the test period a few of the participantsindicated that they had been deliberately trying to equate thereaction times between the lie and true conditions as a strategyto make it harder for the experimenter to detect deceptive trialsThis deliberate attempt to equate reaction times may have beenmore successful when responding to episodic questions It istherefore possible that participants were overall better able tomonitor their own responses during episodic but not during be-lief trials Alternatively it is possible that differences in the na-ture of evaluating episodic knowledge vs personal beliefsaccounted for the differences in deception reaction time By thisaccount the greater deception cost in response time for beliefquestions is due to difficulty espousing a moral opinion that iscontradictory to onersquos own Overall the finding of a deceptioncost effect in participantsrsquo reaction times suggests that the ma-nipulation we used in the current design effectively requiredparticipants to engage additional cognitive resources whenlying compared to telling the truth at least in the case of beliefstatements

In line with a general role for cognitive control in deceptionwe found robust deception effects in frontal and parietal regionsthat did not differ by the type of knowledge (episodic or

A

B

C

Fig 4 Neural correlates of lie preparation (A) Activations during lie preparation

(contrast preparation-liegtpreparation-true) (B) Activations during lie prepar-

ation that were correlated with deception cost score across participants (see text

for details) MNI coordinates are presented at the bottom of each section

(C) Scatterplot of the correlation depicted in B from one of the clusters located in

the posterior cingulate cluster The data in the scatterplot depicts the correl-

ation between activation during the preparation to lie and the behavioral decep-

tion cost score calculated as ratio between reaction time when responding with

truth compared to lie Increased activation in this region was related to smaller

deception cost score

N Ofen et al | 123

personal belief) about which participants lied Regions in thefrontal and lateral parietal cortex are typically recruited duringtasks that require participants to exhibit high levels of cognitivecontrol (Bunge 2004 Ridderinkhof et al 2004 Aron andPoldrack 2006 Dosenbach et al 2006) Prior neuroimaging stud-ies of deception have implicated the involvement of brain re-gions linked to cognitive control (Abe et al 2006 Christ et al2009) Indeed we identified deception effects within the major-ity of regions that were previously identified in a meta-analysis

Different brain regions responded specifically for deceptionabout episodic vs belief knowledge The right temporal pole wasactivated only for deception about episodic personal experienceThis may reflect the recruitment of this region in retrieval ofcomplex detailed memories Consistent with evidence about thisregion being involved in reasoning about beliefs (Goel and Dolan2003 Cunningham et al 2004) we found that it was active dur-ing both lie and truth conditions when participants evaluatedtheir personal beliefs In contrast the precuneus was active onlyfor deception about personal beliefs Another fMRI study alsofound activation of the precuneus when participants lied aboutpersonal information (Marchewka et al 2012) The amygdala aregion linked to fear anxiety and emotion (LeDoux 2007) dis-played a deception effect in our study However the amygdalawas less active in all four conditions than at baseline suggestingthat this region may have been inhibited throughout the taskFinally we also identified a question type difference in the leftand right insula when using ROIs from a prior meta-analysis(Christ et al 2009) The insula has been implicated in certainemotional states (Klucken et al 2012 Sip et al 2012) and it ispossible that these activations indicate an increased sense of

disgust with onesrsquo lie about a personal belief that is not as evi-dent when one lies about past events

The study also included a period of preparation on each trialduring which participants prepared to either tell the truth or tolie before they knew the material they would be asked to lieabout The inclusion of a preparatory period in the experimentaldesign allowed us to test whether the processes involved in de-ception could be dissociated from the processes involved indetermining a specific truth and selecting a particular lie re-sponse We were able to identify regions in the superior parietallobule that were more active during preparation of a lie com-pared with preparation of a truth response These regions weredistinct from the parietal and frontal regions that showedincreased activation during the execution of the lie response Toour knowledge only one published report has examined theneural correlates of preparing to lie (Ito et al 2012) That studydid not show significant differences in brain activation betweenpreparing to tell the truth and preparing to deceive Howeverseveral differences in the design make a direct comparison diffi-cult Ito and colleagues (Ito et al 2012) included a condition inwhich no preparation is possible and the main findings as foractivation specific for preparation were observed when compar-ing certain to uncertain cue In our design participants weregiven a specific cue Second the instruction to either deceive ornot was in only limited context of the participantrsquos memory forpreviously studies pictures of objects whereas in the currentstudy we included two different types of questions in both theparticipant was to reflect on their past experiences or beliefs be-yond the context of specific experimental induced mnemoniccontent Finally the cue used by Ito et al (2012) was a colored

Table 2 Activations related to lie preparation (prepare-lie gt prepare-truth) across participants (A) were related to behavioral index of deceptioncost (B) (see text for details)

BA x y z T value No voxels

(A) Preparation-lie gt preparation-trueL Superiorinferior parietal lobule 7 26 56 36 706 1188

Precuneus 39 30 70 50 452R Middleinferior occipital lobe 18 36 84 4 547 314L Middleinferior occipital lobe 18 44 92 8 514 322

46 82 2 507R Superior parietal lobule 7 34 56 50 486 918

Precuneus 7 26 70 32 415Inferior parietal lobule 40 28 64 38 388

L Caudate na 22 24 24 498 446Middle frontalprecentral gyrus 6 32 4 50 406

R Middle frontalprecentral gyrus 6 34 4 44 431 461Caudate na 22 10 26 397

L Superiormedial frontal gyrus 10 18 60 8 428 139

(B) Correlation with deception cost indexR Posterior cingulate cortex 302923 4 50 14 531 1015L Superior temporal gyrus 22 56 44 8 511 505

Middle temporal gyrusfusiform 37 52 56 2 464L Middleinferior frontal gyrus 469 50 30 26 485 519R Middlesuperior temporal gyrus 22 70 38 4 480 163L Thalamusmidbrain na 18 22 4 477 314L Cuneusmiddle occipital gyrus 1817 16 94 8 448 392R Inferiormiddle frontal gyrus 464510 58 32 8 450 234L Lingual gyrusinferior occipital gyrus 1718 16 96 14 445 117L Inferior frontal gyrus 4746 44 32 0 440 297L Medialsuperior frontal gyrus 10 8 62 16 388 142

Notes Peak coordinates (x y z) are based on MNI brain BA Brodmannrsquos area L left R right

124 | Social Cognitive and Affective Neuroscience 2017 Vol 12 No 1

circle and therefore a possible difference between the studies isin the more direct instruction and possible associations that goalong with processing the cue word lsquoLiersquo or lsquoTurersquo Indeed inthis study we identified regions in the occipital lobe that weremore active during preparation of a lie compared with prepar-ation of a truth response One possible explanation is that thisoccipital activation is reflecting the different visual properties ofthe cue words Nonetheless our results are consistent with theaforementioned paper in that we have identified preparatoryactivation in regions in the prefrontal cortex however unlikethese prior findings we have identified differential preparatoryresponses for lying compared with truth-telling in several re-gions including the superior frontal gyrus

Separating the assessment of neural correlates that supportlie preparation from the assessment of those supporting lie re-sponse allowed us to make an additional unique contribution tothe literature by identifying a specific role for a portion of theright anterior insula in lie preparation The right anterior insulais one of the regions that was identified in meta-analysis (Christet al 2009) as involved in deception but in this study activationin this specific region only differed between the lie and truthconditions during the preparation period and not during the re-sponse Thus the design may have succeeded in dissociatingsome of the processes involved in normal deception

Moreover we found that activation during lie preparation cor-related across participants with individualsrsquo ability to lie effi-ciently We approximated efficient lying with reduced behavioraldeception cost scores Thus we found that individual behavioraldeception costs (measure of reaction time increase during lie vstrue conditions) correlate inversely with individual neural activa-tion during lie preparation Specifically greater preparatory activa-tion in left dorsolateral cortex a region previously linked to bothworking memory (Curtis and DrsquoEsposito 2003 Funahashi 2006)and deception (Nunez et al 2005 Priori et al 2008 Mameli et al2010) was associated with more efficient deception One limitationof our design pertains to the ecological validity of the preparationperiod Specifically deception in the real world does not typicallyinvolve preparing to lie before one has determined what he or shewill be lying about Nonetheless our findings of regions in whichthe level of lie preparation activation correlated with a behavioralindex of lying suggests that the processes carried out during thisperiod may indeed contribute to the act of lying

In this study we used a novel paradigm to assess the neuralcorrelates of deception The most obvious limitation of this andother studies with this goal is the limited ability of experimentalmanipulations to truly mimic real-life deception Participantswere explicitly asked to lie (or tell the truth) and thus even whenthey lied they were in fact complying with the experimenterrsquosinstruction In this regard we greatly minimized the emotionalcomponent in deception In fact the lie detection techniqueswidely in use today rely on physiological measures (Lykken1959) that reflect sympathetic arousal such as skin conductanceresponse These measures are believed to reflect the emotionalcomponent of lying We took measures against this limitation byinstructing the participants to try to respond in such a way thatwe would not be able to tell whether they were lying or tellingthe truth (Uncapher et al 2015) Moreover we found deceptioneffects in the amygdala and anterior insula regions typicallyassociated with emotional processing (LeDoux 2007 WagnerNrsquoDiaye et al 2011 Hamann 2012 Klucken et al 2012)

Another limitation is that we ultimately chose a small set ofquestions (total of 120) to elicit memories of specific past experi-ences or judgments about specific personal beliefs To a large ex-tent the findings of this study are limited by our choice of

specific example questions The comparison between episodicand belief questions for example is limited by minimal controlof the strength or emotional valence that was evoked in the se-lected sets of questions Relatedly the findings regarding decep-tion about past experiences may have been specifically limitedby weak memories of some of those past events If a participantfor example had only a weak memory when evaluating a ques-tion (eg lsquoHave you received a parking ticket in the last monthrsquo)it may be difficult for him or her to engage in intentional decep-tion about this event Although a possible limitation we arguethat the intention to deceive would remain a critical driver ofwhat we observed even in cases when the actual memory isweak Support for this notion comes from a study demonstratingthat the active aspect of deception is critical for engagingfrontal-parietal regions when participants were asked to feignmemory impairment only intentional faked responses and noterrors committed unintentionally were associated with activa-tion in prefrontal and parietal regions (Lee et al 2009)

Conclusions

The findings described here suggest that the brain signature of alie is influenced by the type of knowledge one is lying about Theopen question for those interested in developing techniques for liedetection is whether lying evokes a reliable neural signature thatcannot be manipulated by the deceiver and yet can be reliably de-tected with neuroimaging The cumulative evidence suggests thata few brain regions are consistently more active when a person islying However these regions also appear to be engaged by othercognitive tasks making their use in lie detection algorithms sus-ceptible to false alarms We believe that the importance of decep-tion studies is in identifying the underlying cognitive processes asthey may be linked to specific brain networks that play a role indeception Here we identified differential contributions of brain re-gions to lying about personal past experiences and personal be-liefs Moreover our findings suggest that preparatory processesare directly related to behavioral responses and are supported byregions that are distinct from those that support the actual act oflying These findings add to a growing body of knowledge aboutthe underpinnings of deception in the human brain

Acknowledgments

We wish to thank Max E Gray and J Ben Hutchinson for helpwith data collection and analyses Paul Mazaika Heesoo Kimand Boris Volfson for help with data analysis and AsafBachrach and Itamar Kahn for helpful discussions

Supplementary data

Supplementary data are available at SCAN online

Funding

This work was supported by the Defense AdvancedResearch Projects Agency to JDEG

Conflict of interest None declared

ReferencesAbe N (2009) The neurobiology of deception evidence from

neuroimaging and loss-of-function studies Current Opinion inNeurology 22(6) 594ndash600

N Ofen et al | 125

Abe N Okuda J Suzuki M et al (2008) Neural correlates oftrue memory false memory and deception Cerebral Cortex18(12) 281ndash9

Abe N Suzuki M Mori E Itoh M Fujii T (2007) Deceivingothers distinct neural responses of the prefrontal cortex andamygdala in simple fabrication and deception with socialinteractions Journal of Cognitive Neuroscience 19(2) 287ndash95

Abe N Suzuki M Tsukiura T et al (2006) Dissociable roles ofprefrontal and anterior cingulate cortices in deceptionCerebral Cortex 16(2) 192ndash9

Aron AR Poldrack RA (2006) Cortical and subcortical contri-butions to stop signal response inhibition role of the subthala-mic nucleus Journal of Neuroscience 26(9) 2424ndash33

Bunge SA (2004) How we use rules to select actions a review ofevidence from cognitive neuroscience Cognitive Affective ampBehavioral Neuroscience 4(4) 564ndash79

Christ SE Van Essen DC Watson JM Brubaker LEMcDermott KB (2009) The contributions of prefrontal cortexand executive control to deception evidence from activationlikelihood estimate meta-analyses Cerebral Cortex 19(7)1557ndash66

Cunningham WA Raye CL Johnson MK (2004) Implicit andexplicit evaluation FMRI correlates of valence emotional in-tensity and control in the processing of attitudes Journal ofCognitive Neuroscience 16(10) 1717ndash29

Curtis CE DrsquoEsposito M (2003) Persistent activity in the pre-frontal cortex during working memory Trends in CognitiveSciences 7(9) 415ndash23

Davatzikos C Ruparel K Fan Y et al (2005) Classifying spa-tial patterns of brain activity with machine learning methodsapplication to lie detection Neuroimage 28(3) 663ndash8

Dosenbach NU Visscher KM Palmer ED et al (2006) A coresystem for the implementation of task sets Neuron 50(5)799ndash812

Falkiewicz M Sarzynska J Babula J Szatkowska IGrabowska A Necka E (2015) Explicit instructions increasecognitive costs of deception in predictable social contextFrontiers in Psychology 6 1863

Funahashi S (2006) Prefrontal cortex and working memoryprocesses Neuroscience 139(1) 251ndash61

Ganis G (2015) Deception detection using neuroimaging InGranhag PA Vrij A Verschuere B editors DetectingDeception Current Challenges and Cognitive Approaches (pp105ndash21) Chicago IL John Wiley

Ganis G Kosslyn SM Stose S Thompson WL Yurgelun-Todd DA (2003) Neural correlates of different types of decep-tion an fMRI investigation Cerebral Cortex 13(8) 830ndash6

Ganis G Morris RR Kosslyn SM (2009) Neural processesunderlying self- and other-related lies an individual differ-ence approach using fMRI Society for Neuroscience 4(6) 539ndash53

Glover GH Law CS (2001) Spiral-inout BOLD fMRI forincreased SNR and reduced susceptibility artifacts MagneticResonance in Medicine 46(3) 515ndash22

Goel V Dolan RJ (2003) Explaining modulation of reasoningby belief Cognition 87(1) B11ndash22

Greene JD Paxton JM (2009) Patterns of neural activity asso-ciated with honest and dishonest moral decisions Proceedingsof the National Academy of Sciences of United States America106(30) 12506ndash11

Hamann S (2012) Mapping discrete and dimensional emotionsonto the brain controversies and consensus Trends inCognitive Sciences 16(9) 458ndash66

Ito A Abe N Fujii T et al (2012) The contribution of thedorsolateral prefrontal cortex to the preparation for deceptionand truth-telling Brain Research 1464 43ndash52

Johnson KA Kozel FA Laken SJ George MS (2007) Theneuroscience of functional magnetic resonance imaging fMRIfor deception detection American Journal of Bioethics 7(9) 58ndash60

Klucken T Schweckendiek J Koppe G et al (2012) Neuralcorrelates of disgust- and fear-conditioned responsesNeuroscience 201 209ndash18

Kozel FA Johnson KA Mu Q Grenesko EL Laken SJGeorge MS (2005) Detecting deception using functional mag-netic resonance imaging Biological Psychiatry 58(8) 605ndash13

Langleben DD Loughead JW Bilker WB et al (2005) Tellingtruth from lie in individual subjects with fast event-relatedfMRI Human Brain Mapping 26(4) 262ndash72

Langleben DD Schroeder L Maldjian JA et al (2002)Brain activity during simulated deception an event-relatedfunctional magnetic resonance study Neuroimage 15(3)727ndash32

LeDoux J (2007) The amygdala Current Biology 17(20) R868ndash74Lee TM Au RK Liu HL Ting KH Huang CM Chan CC

(2009) Are errors differentiable from deceptive responseswhen feigning memory impairment An fMRI study Brain andCognition 69(2) 406ndash12

Lieberman MD Cunningham WA (2009) Type I and Type IIerror concerns in fMRI research re-balancing the scale SocialCognitive and Affective Neuroscience 4(4) 423ndash8

Lilienfeld SO Andrews BP (1996) Development and prelimin-ary validation of a self-report measure of psychopathic per-sonality traits in noncriminal populations Journal of PersonalityAssessment 66(3) 488ndash524

Lisofsky N Kazzer P Heekeren HR Prehn K (2014)Investigating socio-cognitive processes in deception a quanti-tative meta-analysis of neuroimaging studiesNeuropsychologia 61 113ndash22

Lykken DT (1959) The GSR in the detection of guilt Journal ofApplied Psychology 43(6) 385ndash8

Mameli F Mrakic-Sposta S Vergari M et al (2010)Dorsolateral prefrontal cortex specifically processes general -but not personal - knowledge deception multiple brainnetworks for lying Behavioural Brain Research 211(2) 164ndash8

Marchewka A Jednorog K Falkiewicz M Szeszkowski WGrabowska A Szatkowska I (2012) Sex lies and fMRIndashgender differences in neural basis of deception PLoS One 7(8)e43076

Nunez JM Casey BJ Egner T Hare T Hirsch J (2005)Intentional false responding shares neural substrates with re-sponse conflict and cognitive control Neuroimage 25(1)267ndash77

Pascual L Rodrigues P Gallardo-Pujol D (2013) How does mor-ality work in the brain A functional and structural perspectiveof moral behavior Frontiers in Integrative Neuroscience 7 65

Priori A Mameli F Cogiamanian F et al (2008) Lie-specific in-volvement of dorsolateral prefrontal cortex in deceptionCerebral Cortex 18(2) 451ndash5

Ridderinkhof KR Ullsperger M Crone EA Nieuwenhuis S(2004) The role of the medial frontal cortex in cognitive con-trol Science 306(5695) 443ndash7

Rissman J Greely HT Wagner AD (2010) Detecting individ-ual memories through the neural decoding of memory statesand past experience Proceedings of the National Academy ofSciences of United States America 107(21) 9849ndash54

126 | Social Cognitive and Affective Neuroscience 2017 Vol 12 No 1

Sip KE Roepstorff A McGregor W Frith CD (2008)Detecting deception the scope and limits Trends in CognitiveSciences 12(2) 48ndash53

Sip KE Skewes JC Marchant JL McGregor WB RoepstorffA Frith CD (2012) What if I Get Busted Deception Choiceand Decision-Making in Social Interaction Frontiers inNeuroscience 6 58

Spence SA (2004) The deceptive brain Journal of the RoyalSociety of Medicine 97(1) 6ndash9

Uncapher MR Boyd-Meredith JT Chow TE Rissman JWagner AD (2015) Goal-directed modulation of neural

memory patterns implications for fmri-based memory detec-tion Journal of Neuroscience 35(22) 8531ndash45

Volz KG Vogeley K Tittgemeyer M von Cramon DYSutter M (2015) The neural basis of deception in strategicinteractions Frontiers in Behavioral Neuroscience 9 27

Vrij A Fisher R Mann S Leal S (2006) Detecting deceptionby manipulating cognitive load Trends in Cognitive Sciences10(4) 141ndash2

Wagner U NrsquoDiaye K Ethofer T Vuilleumier P (2011) Guilt-specific processing in the prefrontal cortex Cerebral Cortex21(11) 2461ndash70

N Ofen et al | 127