An fMRI investigation of memory encoding in PTSD: Influence of symptom severity

Page 1

A

emtfpmai©

K

1

TmPtfius&she

LT

0d

Neuropsychologia 46 (2008) 1522–1531

An fMRI investigation of memory encoding in PTSD:Influence of symptom severity

Erin W. Dickie a,∗, Alain Brunet a,b, Vivian Akerib a, Jorge L. Armony a,b

a Douglas Mental Health University Institute, Canadab Department of Psychiatry, McGill University, Montreal, Quebec, Canada

Received 13 November 2007; received in revised form 18 December 2007; accepted 6 January 2008Available online 19 January 2008

bstract

Previous studies have shown memory deficits in Post-Traumatic Stress Disorder (PTSD) patients, as well as abnormal patterns of brain activity,specially when retrieving trauma-related information. This study extended previous findings by investigating the neural correlates of successfulemory encoding of trauma-unrelated stimuli and their relationship with PTSD symptom severity. We used the subsequent memory paradigm, in

he context of event-related functional magnetic resonance imaging, in 27 PTSD patients to identify the brain regions involved in the encoding ofearful and neutral faces. Symptom severity was assessed by the Clinically Administered PTSD Scale (CAPS) scores. It was found that memoryerformance was negatively correlated with CAPS scores. Furthermore, a negative correlation was observed between CAPS scores and ventral

edial prefrontal cortex (vmPFC) activity elicited by the subsequently forgotten faces. Finally, symptom severity predicted the contribution of the

mygdala to the successful encoding of fearful faces. These results confirm the roles of the vmPFC and the amygdala in PTSD and highlight themportance of taking into account individual differences when assessing the behavioural and neural correlates of the disorder.

2008 Elsevier Ltd. All rights reserved.

ory;

fc12leL

otiB

eywords: Post-Traumatic Stress Disorder; Magnetic resonance imaging; Mem

. Introduction

Several of the clinical symptoms associated with Post-raumatic Stress Disorder (PTSD) – including intrusiveemories, flashbacks and psychogenic amnesia (Americansychiatric Association, 2000) – suggest the existence of dis-

urbances in memory function in this disorder. Numerousunctional neuroimaging studies have examined brain activ-ty associated with the recall of the traumatic event, mostlysing symptom-provocation paradigms through personalizedcripts, images or sounds (for reviews see Francati, Vermetten,

Bremner, 2007; Rauch, Shin, & Phelps, 2006). Most of these

tudies have shown that PTSD patients, when compared toealthy controls or trauma-exposed individuals without PTSD,xhibit decreases in activity within regions of the medial pre-

∗ Corresponding author at: Douglas Mental Health University Institute, 6875aSalle Boulevard, F.B.C. Pavillon, Verdun, Quebec H4H 1R3, Canada.el.: +1 514 761 6131x3394; fax: +1 514 888 4099.

E-mail address: [email protected] (E.W. Dickie).

Jtsa(alng

028-3932/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.oi:10.1016/j.neuropsychologia.2008.01.007

Encoding; Medial frontal cortex; Amygdala

rontal cortex, including the anterior cingulate and orbital frontalortices (Bremner, Narayan, et al., 1999; Bremner, Staib, et al.,999; Britton, Phan, Taylor, Fig, & Liberzon, 2005; Lanius et al.,001; Rauch et al., 1996; Shin, Orr, et al., 2004), and, althoughess consistently, increased amygdala activation (e.g., Liberzont al., 1999; Shin, Orr, et al., 2004; but see Britton et al., 2005;anius et al., 2001).

Importantly, a growing body of literature suggests that mem-ry dysfunction in PTSD may not be limited to material relatedo the traumatic event, but it may extend to trauma-unrelated,ncluding emotionally neutral, information (for reviews seerewin, Kleiner, Vasterling, & Field, 2007; Isaac, Cushway, &

ones, 2006). However, unlike the case of traumatic memories,he neural correlates of memory deficits for trauma-unrelatedtimuli in PTSD have not been greatly explored. To date, onlyhandful of studies have examined this issue. Bremner et al.

2003) scanned survivors of early sexual abuse with PTSD

nd healthy controls using PET while they recalled previouslyearned emotionally charged and neutral word pairs. Althougho differences in memory performance were observed betweenroups, PTSD participants showed decreased activity in the

Page 2

ychol

me(wtfHeiiWtamgordh

dPmabtutm

tco(ibottsseKChmWrBr

biiaPw

btSiPdeHom(adfsidrTotose2imwr

2

2

rLttPPaCdrd(cyT(ntstm

E.W. Dickie et al. / Neurops

edial prefrontal and orbital frontal cortices while recalling themotional word pairs. In another PET study of memory retrievalShin, Shin, et al., 2004), firefighters with and without PTSDere scanned while they were cued to retrieve neutral words

hat had been previously encoded. Again, no behavioural dif-erences were found between the PTSD patients and controls.owever, the PTSD group showed decreased hippocampal and

nhanced prefrontal activity, relative to controls, when compar-ng the recall of deeply vs. shallowly encoded words. Finally,n a recent fMRI study, Geuze, Vermetten, Ruf, de Kloet, and

estenberg (2007) scanned PTSD patients and matched con-rols during both encoding and retrieval of neutral verbal pairedssociates. Behaviourally, a trend towards a deficit in perfor-ance was observed for PTSD patients. In addition, the PTSD

roup exhibited reduced activity, relative to controls, in regionsf the frontal lobe together with larger activation in the tempo-al lobe during the encoding phase. At retrieval, they showedecreased activity in areas of the frontal lobe and the posteriorippocampus.

Taken together, these findings provide further support for aysfunction of the fronto-temporal circuit in individuals withTSD. In addition, they suggest that a common neural substrateay underlie the abnormal memory patterns for both traumatic

nd neutral information observed in these individuals. However,ecause all of these studies employed a block design, in whichhe activity associated with both remembered and forgotten stim-li was grouped together in the analysis, the relation betweenhe observed differences in brain activation and the behavioural

emory deficits present in PTSD patients remains unclear.Furthermore, PTSD is a complex disorder, characterized by

he simultaneous presence of a number of symptoms. The spe-ific combination of symptoms, as well as the intensity of eachf them can vary among patients, spanning a considerable rangeWeathers, Keane, & Davidson, 2001). In addition to the obviousnfluence on an individual’s quality of life and emotional well-eing (Johansen, Wahl, Eilertsen, Weisaeth, & Hanestad, 2007),verall PTSD symptom severity has been shown to be an impor-ant predictor of the magnitude of the cognitive impairments andheir neural correlates, associated with the disorder. For instance,everal PTSD studies have reported that increased symptomeverity is associated with larger memory deficits (Bremnert al., 1993; Gilbertson, Gurvits, Lasko, Orr, & Pitman, 2001;ivling-Boden & Sundbom, 2003; Lindauer, Olff, van Meijel,arlier, & Gersons, 2006). In addition, PTSD symptom severityas been shown to be negatively correlated with activity in thePFC (Bryant et al., 2007; Kim et al., 2007; Shin et al., 2005;illiams et al., 2006) and positively correlated with amygdala

esponses to emotional stimuli (Armony, Corbo, Clement, &runet, 2005; Rauch et al., 2000) or during traumatic memory

etrieval (Shin, Orr, et al., 2004).Thus, in addition to studying the differences in brain activity

etween PTSD patients and control groups in episodic memory,t seems important to investigate the influence of symptom sever-

ty in this process. Indeed, studying individual differences withinPTSD group, by taking into account not only the presence ofTSD symptoms but also their intensity, may reveal effects thatould be “washed-out” in a between-groups comparison.

d

fMn

ogia 46 (2008) 1522–1531 1523

We therefore conducted a study to investigate the relationetween memory for faces (fearful and neutral) and PTSD symp-om severity, as measured by the Clinically Administered PTSDcale (CAPS) score. We specifically focused on memory encod-

ng, as it has been suggested that the memory deficits shown inTSD may in part be due to attention and working memoryifficulties at the time of encoding (Brandes et al., 2002; Isaact al., 2006; Jenkins, Langlais, Delis, & Cohen, 2000; Koso &ansen, 2006; Vasterling, Brailey, Constans, & Sutker, 1998). Inrder to directly isolate the neural regions involved in successfulemory formation, we used the subsequent memory paradigm

Brewer, Zhao, Desmond, Glover, & Gabrieli, 1998; Wagner etl., 1998) in the context of an event-related analysis. This proce-ure allowed us to categorize each stimulus as remembered ororgotten based on each subjects’ behavioural response in a sub-equent memory test administered outside the scanner. Recentnvestigations from our laboratory have shown that the amyg-ala is critically involved in the successful encoding of fearful,elative to neutral faces (Sergerie, Lepage, & Armony, 2006).herefore, fearful and neutral faces were chosen as stimuli inrder to explore the potential interactions between PTSD symp-oms and this emotion-specific, amygdala-mediated modulationf memory encoding. We hypothesized that PTSD symptomeverity should predict overall memory performance (Bremnert al., 1993; Gilbertson et al., 2001; Kivling-Boden & Sundbom,003; Lindauer et al., 2006) with concomitant changes in activ-ty in medial prefrontal and temporal regions as a function of

emory success. Furthermore, we predicted that CAPS scoresould correlate with the amygdala responses to subsequently

emembered fearful faces.

. Methods

.1. Participants

Thirty-two individuals (range: 20–60 years) suffering from PTSD wereecruited from two Montreal clinics (the Traumatys Clinic and the CharleseMoyne Hospital). All recruitment and testing procedures were approved by

he ethical review boards of the Douglas Mental Health University Institute,he Montreal Neurological Institute and the McGill University Health Centre.TSD diagnosis was verified in all participants using the Clinically AdministeredTSD Scale (CAPS; Blake et al., 1995), which was administered and scored byclinical psychologist (V.A.). Inclusion criteria required participants to have aAPS score greater than 45 and no history of neurological, learning or psychoticisorders. Five subjects were excluded from the final analysis for the followingeasons: excessive movement during the scan (n = 2), below-chance performanceuring the memory test (n = 1) and no responses for one of the stimulus categoriesn = 2; see fMRI data acquisition and analysis below). Thus, the final sampleonsisted of 27 individuals (19 female, 8 male; age M = 36 years, S.D. = 11ears), all with CAPS scores greater than 50 (M = 83, S.D. = 16, median: 87).he nature of the trauma for these participants included motor vehicle accidents

n = 13), physical assault and/or death threats (n = 7), sexual assault (n = 1), wit-essing a violent physical assault or death (n = 5), and combat exposure duringhe Rwanda Genocide (n = 1). The time elapsed between their trauma and MRIcanning ranged from 5 to 62 weeks (M = 16, S.D. = 12), with the exception ofhe latter participant, who was exposed to trauma 11 years prior to the experi-

ent. Post hoc analyses showed that removal of this subject from the analysis

id not significantly change the pattern of results reported here (data not shown).

The severity of participants’ subjective traumatic experiences was assessedurther using the Peritraumatic Distress Inventory (Brunet et al., 2001;

= 31.6, S.D. = 11.7) and the Peritraumatic Dissociative Experience Question-aire (Birmes et al., 2005; M = 29.7, S.D. = 10.0). Psychiatric comorbidity was

Page 3

1 ychol

aawspsmathmm

WlCBdsoa

2

vF(l

s(stlost

csh

fwleas(twtfa

ohPoatn

2

1ui(t

524 E.W. Dickie et al. / Neurops

ssessed with the MINI International Neuropsychiatric Interview (Sheehan etl., 1998): 10 subjects were free of co-morbid diagnoses, while 17 presentedith one or more co-morbid mood or anxiety disorders, including major depres-

ion (n = 12), another anxiety disorder (n = 12), and eating disorder (n = 1). Noarticipant had suffered from alcohol or drug abuse for 6 months prior to thecan. In terms of medication, fourteen participants were not taking psychoactiveedications at the time of testing, while the other 13 were taking prescribed

ntidepressants (n = 12) and/or anxiolytics (n = 6). An independent samples t-est showed a non-significant trend such that those patients taking medicationad higher average CAPS scores (M = 88.4, S.D. = 13.8) than those not takingedication (M = 77.8, S.D. = 16.6; t(25) = 1.78, p = 0.09), Therefore, the use ofedication was also included in post hoc analyses as a covariate.

Depressive symptoms, assessed using the Beck Depression Inventory (Beck,ard, Mendelson, Mock, & Erbaugh, 1961), ranged from healthy to clinical

evels (M = 10.6, S.D. = 6.1, range 0–26) and were significantly correlated withAPS scores (r = 0.57, p < 0.005). Therefore, all analyses were repeated usingDI scores as a covariate in order to statistically control for the effects ofepressive co-morbidity. In addition, post hoc analyses were also conducted totatistically control for the effects of time-since-trauma (after exclusion of oneutlier on this measure, see above), presence of a co-morbid anxiety disorder,nd use of psychoactive medications.

.2. Experimental design

Face stimuli (112 fearful, 112 neutral) were selected from severalalidated databases: the Karolinska Directed Emotional Faces (Lundqvist,lykt, & Ohman, 1998), the Psychological Image Collection at Stirlinghttp://pics.psych.stir.ac.uk), the CVL Face Database (http://lrv.fri.uni-j.si/facedb.html), the AR Face Database (http://rvl1.ecn.purdue.edu/aleix/aleixface DB.html), the Ekman set (Ekman & Friesen, 1976), the PennCNPet (http://www.med.upenn.edu/bbl/), and the NimStim Face Stimulus Sethttp://www.macbrain.org). Of these 224 faces, 112 were pseudo-randomlyelected for each subject and divided into two sets of 28 fearful and 28 neu-

ral faces each (half male, half female) to act either as encoding stimuli or asures during the recognition test (i.e., old and new stimuli). Fifty-six indoor andutdoor scenes were also presented, intermixed with the faces during the fMRIcanning (data not reported here). All images were modified using Adobe Pho-oshop 7.0 (Adobe Systems, San Jose, CA) to achieve consistency in size and

ccTca

Fig. 1. Schematic of the experimental paradigm

ogia 46 (2008) 1522–1531

ontrast, cut into a 158-by-225-pixel oval and converted to grayscale. This ovalhape occluded the presentation of gender typical features in the faces, such asair and face shape (see Fig. 1).

A schematic of the experimental procedure is shown in Fig. 1. During theMRI scan (encoding session), 56 faces were presented twice and 56 scenesere presented once for 1500 ms each, preceded by a 1000 ms cue. The stimu-

us onset asynchrony was 2000 ms on average; longer intervals (so-called nullvents) were also included in order to obtain an adequate estimate of baselinectivity. Stimulus onset was desynchronized with respect to the scan acqui-ition time to ensure an equal sampling of the entire hemodynamic responseJosephs, Turner, & Friston, 1997). During the scan, participants were instructedo indicate, pressing a button, whether the face presented was male or female, orhether the scene was of an interior or an exterior. In addition, subjects were told

o memorize all images in preparation for a memory test. This task was scannedollowing the acquisition of another functional scan (results not presented here)nd a structural MRI image.

After the scanning session, participants completed the face recognition mem-ry test. One hundred and twelve faces were presented on a PC laptop screen,alf of which had been seen in the scanner, the other half never seen before.articipants were instructed to indicate, by pressing a key, whether the face wasld (previously presented) or new. Faces were presented for 1500 ms each with2000 ms inter-stimulus interval. Following the memory test, participants rated

he faces that they had seen for emotional valence on a visual analog scale (“veryegative/tres negative” to “very positive/tres positive”).

.3. fMRI acquisition and analysis

Scanning was conducted at the Montreal Neurological Institute (MNI) with a.5 Tesla Siemens Sonata whole-body system equipped with a standard head coil,sing gradient EPI sequences. A vacuum cushion was used to stabilize the partic-pants’ head. Stimuli were generated by a PC laptop computer running E-PRIMEPsychology Software Tools, Pittsburg, PA) and displayed using an LCD projec-or, a screen and a mirror system. An optical mouse connected to the computer

ollected the participants’ responses. Functional T2*-weighted images wereollected with blood oxygen level dependent (BOLD) contrast (360 volumes,R = 2450 ms, TE = 50 ms, Flip angle 90◦, FOV = 256 mm, Matrix = 64 × 64),overing the entire brain (30 interleaved slices, 4 mm thickness, parallel to thenterior-posterior commissural plane; voxel size 4 mm × 4 mm × 4 mm). Prior

. Only the encoding session was scanned.

Page 4

ychologia 46 (2008) 1522–1531 1525

tgs

ClfitpialeditmwioimmupaaL(tasttaa

3

3

pApeet

rAfa2ms(2tcrst

Fm

sisfC

3

rrapaatasee2

3

sanstss

E.W. Dickie et al. / Neurops

o the functional scan a T1-weighted anatomical volume was collected using aradient echo pulse sequence (TR = 22 ms, TE = 9.2 ms, Flip angle = 30◦, voxelize 1 mm × 1 mm × 1 mm).

Functional data was preprocessed using SPM2 (Wellcome Department ofognitive Neurology, London, UK; see http://www.fil.ion.ucl.ac.uk/spm), fol-

owing standard procedures. Briefly, images were time corrected, realigned to therst volume and spatially normalized (final voxel size 2 mm × 2 mm × 2 mm)

o the stereotaxic space of Talairach and Tournoux (1988) using the MNI tem-late (Evans, Kamber, Collins, & MacDonald, 1994) and smoothed using ansotropic 8 mm FWHM Gaussian kernel. A high-pass filter (cut of 128 s) waspplied to remove low frequency temporal drifts in fMRI signal. Data was ana-yzed with SPM2 using a general linear model with correlated errors, in whichvents were modeled by a synthetic hemodynamic response and its temporalerivative. Five conditions were defined, based on each participant’s responsesn the memory recognition test: subsequently remembered fearful and neu-ral faces, subsequently forgotten fearful and neutral faces, and scenes. The

ovement parameters obtained during preprocessing (realignment procedure)ere also included in the model. For each participant, three linear contrasts of

nterest were calculated, corresponding to the main effects of subsequent mem-ry (Remembered–Forgotten), emotional expression (Fearful–Neutral) and theirnteraction. These linear contrasts were taken to second level random-effects

odels using whole-brain analysis at a threshold of p < 0.001, uncorrected forultiple comparisons, together with a cluster threshold of p < 0.05. Small vol-

me correction was applied to the amygdala and hippocampus because of our ariori hypothesis regarding amygdala and hippocampus involvement in PTSDnd memory, and the small volume of these regions. Masks for the left and rightmygdala and hippocampus were created using the WFU pickatlas (Maldjian,aurienti, Kraft, & Burdette, 2003) in accordance with Tzourio-Mazoyer et al.

2002). Two types of random effects analyses were conducted, one sample t-testso determine those areas activated or deactivated during a particular contrast byll patients, and linear regressions to determine those areas in which the BOLDignal, for a particular contrast, was significantly correlated with PTSD symp-om severity (CAPS). Once activations of interest were identified according tohe significance thresholds listed above, data from the associated cluster (usingthreshold of p < 0.005) was extracted and averaged for subsequent post hoc

nalyses and plotting.

. Results

.1. Memory performance

Accuracy (percent correct) for the gender decision task com-leted during encoding was very high (M ± S.D. = 89 ± 4%).nalysis of accuracy and reaction time data from the encodinghase as a function of later memory performance, emotionalxpression and CAPS scores did not reveal any significant mainffects or interactions (for accuracy all p > 0.22, for reactionimes all p > 0.39).

Memory performance, indexed by hits minus false alarmates, was significantly above chance (t(26) = 8.63, p < 0.001).

repeated-measures ANOVA on memory performance, withacial expression as the within-subjects factor and CAPS scoress a covariate, showed a significant main effect of CAPS (F(1,5) = 4.48, p < 0.05), reflecting a negative correlation betweenemory performance and CAPS scores (r = −0.40, p < 0.05;

ee Fig. 2). There was no main effect of stimulus emotionF(1, 25) = 0.12) or stimulus emotion by CAPS interaction (F(1,5) = 2.01). Including BDI scores as a covariate in the analysiso control for the effects of comorbid depression did not signifi-

antly change the results (CAPS: partial r = −0.36, BDI: partial= 0.04). Similarly, possible mediating effects of time elapsedince trauma, peritraumatic distress, peritraumatic dissociation,he presence of a co-morbid anxiety disorder and the use of pre-

swa(

ig. 2. Scatterplot depicting the negative correlation between memory perfor-ance (hit minus false alarm rates) and PTSD symptom severity (CAPS).

cribed anxiolytic or anti-depressant medication, were tested byncluding these variables as covariates during in a linear regres-ion equation, either individually or all together. No variable wasound to affect the relation between memory performance andAPS scores (all p > 0.65).

.2. Emotional valence ratings

Participants’ mean valence ratings were entered into aepeated measures ANOVA with facial expression as theepeated measure and CAPS scores as a covariate. As expected,

main effect of emotion was observed (F(1, 25) = 275.3,< 0.001), as fearful faces were consistently rated as more neg-tive than neutral ones. In addition, there was a trend towardsmain effect of CAPS (F(1, 25) = 3.45, p = 0.08), suggesting

hat subjects with higher CAPS scores were more likely to ratell the faces more negatively. This trend persisted after BDIcores were included in the above analyses as a covariate (mainffect of emotional expression: F(1, 24) = 264.7, p < 0.001; mainffect of CAPS F(1, 24) = 3.56, p = 0.07; main effect of BDI F(1,4) = 0.43, ns).

.3. fMRI data

Results for the main effects of memory and emotional expres-ion, as well as their interaction, obtained in the whole-brainnalysis are shown in Table 1. Brain regions showing a sig-ificant correlation between the contrast of interest and CAPScores are reported in Table 2. Interestingly, the activity withinhe ventral medial prefrontal cortex (vmPFC) associated withubsequent memory success significantly correlated with CAPScores (xyz = [0 42 −8], Z = 3.86, p(cluster) = 0.04, r = 0.65), as

hown in Fig. 3a, b. Post hoc analysis revealed that this effectas driven by a significant negative correlation between vmPFC

ctivity for subsequently forgotten items and CAPS scoresr = −0.51, p < 0.01; Fig. 3d). In contrast, no association was

Page 5

1526 E.W. Dickie et al. / Neuropsychologia 46 (2008) 1522–1531

Table 1Clusters showing significant activations (main effects)

Region x y z Z score (peak voxel) Cluster size p-Value (cluster)

Remembered–forgotten contrastR amygdala* 18 −6 −20 3.04L hippocampus* −26 −14 −22 2.94R inferior temporal gyrus/fusiform gyrus 42 −64 −4 3.93 40 0.08

Fearful–neutral contrastL fusiform gyrus −38 −60 −18 4.04 80 0.02L inferior temporal gyrus −48 −68 −4 3.64 56 0.05

Neutral (remembered–forgotten)–fearful (remembered–forgotten)R inferior parietal lobule 50 −54 56 4.36 229 <0.001

−12

L

aiiieattalFiPiCcdtafp

tiC

(pw

bwZapp(tfAtmp

ioip

TC

R

R

F

F

B

L orbital frontal cortex −28 40

: left and R: right.* p < 0.05 small-volume correction.

pparent between neural activity for subsequently rememberedtems and CAPS scores (r = −0.08, p = 0.70; Fig. 3c). Thats, high CAPS scores were associated with reduced activityn vmPFC for faces that would be later forgotten. To furtherxplore this intriguing effect, subjects were divided into high-nd low-CAPS groups according to a median split. Those inhe high-CAPS group showed a significant negative correla-ion between memory performance and the memory-relatedctivity in the vmPFC (r = −0.67, p < 0.01; Fig. 3f) while theow-CAPS group showed no such relationship (r = 0.04, ns;ig. 3e). This group difference was confirmed by a three-way

nteraction between vmPFC activity, memory performance andTSD subgroup (t(24) = 2.25, p < 0.05). In order to assess the

mpact of depressive co-morbidity on the relationship betweenAPS scores and fMRI signal, multiple regression analyses wereonducted on memory-based activity from this cluster as theependent variable, using both CAPS and BDI scores as predic-ors. CAPS scores remained a significant predictor of vmPFCctivity (t(24) = 3.4, p = 0.002, partial r = 0.57) while BDI scoresailed to add significantly to the equation (t(24) = 0.12, p = 0.9,artial r = −0.03).

Region-of-interest analyses were conducted to investigate

he contributions of amygdala and hippocampal activ-ty to memory encoding, as well as interactions withAPS scores. As expected, activity in the right amygdala

pai

able 2lusters showing a significant positive correlation with CAPS scores

egion x y z

emembered–forgotten contrastB ventral medial prefrontal cortex 0 42 −8

earful–neutral contrastNo significant clusters

earful (remembered–forgotten)–neutral (remembered–forgotten) contrastL amygdala* −18 −6 −14L superior temporal sulcus −46 −42 8

−64 −50 12L superior temporal sulcus/insula −46 −2 −14

: bilateral; L: left; R: right.* p < 0.05 small-volume correction.

4.02 65 0.03

xyz = [18 −6 −20], Z = 3.04, p = 0.03, corrected) and left hip-ocampus (xyz = [−26 −14 −22], Z = 2.94, p = 0.04, corrected)as greater for remembered items than for forgotten ones.In terms of emotional memory, a three-way interaction

etween subsequent memory, emotional expression and CAPSas observed in the left amygdala (xyz = [−18 −6 −14],= 2.94, p = 0.03, corrected), with a trend for similar interaction

lso present in the right amygdala (xyz = [22 −8 −12], Z = 2.60,= 0.07), as shown in Fig. 4a. This interaction with CAPS scoresersisted after BDI scores were incorporated into the analysist(24) = 2.51, p = 0.02, partial r = 0.46). In order to understandhis effect, the cluster’s BOLD signal relating to memory successor fearful and neutral items was extracted and further analyzed.s shown in Fig. 4b, there was a significant positive correla-

ion between left amygdala activity associated with successfulemory for fearful faces and PTSD symptom severity (r = 0.46,= 0.02).

In addition to comorbid depression, we included covariatesnto linear regression models to control for the possible effectsn the fMRI results of (1) time elapsed since trauma, (2) per-traumatic distress and (3) peritraumatic dissociation (4) theresence of a co-morbid anxiety disorder and (5) the use of

rescribed medication. All the effects reported for the vmPFCnd amygdala remained significant after entering these variablesnto the regression analyses either separately or simultaneously

Z score (peak voxel) Cluster size p-Value (cluster)

3.86 57 0.04

2.943.58 55 0.043.79 42 0.073.64 37 0.08

Page 6

E.W. Dickie et al. / Neuropsychologia 46 (2008) 1522–1531 1527

Fig. 3. (a) Statistical parametric map depicting the cluster within the vmPFC (peak voxel [xyz = 0 42 −8]) exhibiting a significant correlation between PTSD symptomseverity (CAPS) and the contrast representing remembered minus forgotten faces. Parameter estimates from this cluster were extracted and plotted against CAPS fort e, andr t mina

(r

4

iiecm

iioGLa

he (b) remembered minus forgotten contrast, (c) remembered faces vs. baselinemembered minus forgotten contrast as a function of memory performance (himedian split).

for vmPFC all partial r > 0.57, p < 0.01; for amygdala all partial> 0.50, p < 0.05).

. Discussion

In this study, we investigated the influence of symptom sever-ty on the neural correlates of memory encoding for faces

n PTSD. By using the subsequent memory paradigm in anvent-related fMRI design, we were able to isolate the specificontributions of brain regions to successful and unsuccessfulemory formation.

scis

(d) forgotten faces vs. baseline. Scatterplots of the vmPFC activation for theus false alarm rates) for the (e) low- and (f) high-CAPS groups (determined by

We observed that increasing PTSD symptom severity, asndexed by CAPS scores, predicted a reduced performancen face recognition memory, consistent with previous studiesf memory function in PTSD patients (Bremner et al., 1993;ilbertson et al., 2001; Kivling-Boden & Sundbom, 2003;indauer et al., 2006). Furthermore, activity in the vmPFC,s a function of memory success, was correlated with PTSD

ymptom severity. This effect was primarily due to a negativeorrelation between CAPS and vmPFC activity for forgottentems. Therefore, it appears that in patients with more severeymptomatology, a reduction in stimulus-elicited activity of the

Page 7

1528 E.W. Dickie et al. / Neuropsychologia 46 (2008) 1522–1531

Fig. 4. (a) Statistical parametric map depicting the activation within the left (xyz = [−18 −6 −14]) and right (xyz = [−18 −6 −14]) amygdala (threshold at p = 0.05 fore tom sa metec

vbtttwpa1sfmItm

4

ictt

ebe2sfrmvoas

scsfmw

asy visualization), representing the significant interaction between PTSD sympnd stimulus emotional expression (fearful vs. neutral). Scatterplots of the paraontrast against CAPS scores for (b) fearful and (c) neutral faces.

mPFC is associated with a greater likelihood to forget what iseing presented. While the vmPFC activity during our memoryask varied according to PTSD symptom severity, we consis-ently saw greater activity for remembered items, in comparisono forgotten ones, in several brain areas previously associatedith memory encoding, including the amygdala, the hippocam-us, and regions of the inferior temporal cortex (Brewer etl., 1998; Wagner, Koutstaal, & Schacter, 1999; Wagner et al.,998). Interestingly, although the relationship between PTSDymptoms and memory performance did not differ for fear-ul and neutral faces, symptom severity was associated withemory-related activity in the amygdala for fearful faces only.

mportantly, all the results reported above persisted after sta-istically controlling for symptoms of depression, as well as

edication and time since trauma.

.1. Prefrontal cortex and successful memory encoding

Previous studies have often reported decreased activation

n regions of the PFC (including the rostral anterior cingulateortex and the orbital frontal cortex) in PTSD in response torauma-related or unrelated emotional material, compared torauma-exposed or healthy controls (Francati et al., 2007; Rauch

K1oh

everity (CAPS), subsequent memory performance (remembered vs. forgotten)r estimates, for the left amygdala cluster, for the remembered minus forgotten

t al., 2006). Furthermore, the magnitude of this reduction haseen shown to correlate with PTSD symptom severity (Bryantt al., 2007; Kim et al., 2007; Shin et al., 2005; Williams et al.,006). In our study, symptom severity, as measured by CAPScores, predicted vmPFC activity associated with subsequentlyorgotten items. In addition, memory-related activity in thisegion was negatively correlated with overall memory perfor-ance, particularly in the high-CAPS group. Therefore, if the

mPFC is, directly or indirectly, necessary for successful mem-ry formation, it is conceivable that a reduced activity of thisrea in highly symptomatic PTSD patients may lead to a poortimulus encoding and thus to reduced memory performance.

The mechanisms linking vmPFC activity, PTSD symptomeverity and memory function remain to be determined. Wean, however, speculate on a number of possible ways in whichymptom severity may affect memory through impaired vmPFCunctioning. It has been proposed that memory deficits in PTSDay be partly explained by related dysfunctions in attention andorking memory (Brandes et al., 2002; Jenkins et al., 2000;

oenen et al., 2001; Koso & Hansen, 2006; Vasterling et al.,998, 2002), which may be in turn related to frontal lobe pathol-gy (Koso & Hansen, 2006; Vasterling et al., 1998). The vmPFCas been identified as part of the brain’s “default network”, a

Page 8

ychol

gbradKfg2sardimmptp2Pa

PieeP(2ovaf(tcsfehaaMettft

4

ttBD

1ahg(TdCbPbHfms

idti2Dorpobiow(bibrcsrhiEmbe

iibhIdl

E.W. Dickie et al. / Neurops

roup of brain regions which deactivate in comparison to restingaseline, reflecting the redirection of attentional and cognitiveesources from the monitoring of internal states and thoughts ton external cognitive task (Gusnard & Raichle, 2001). Theseeactivations are modulated by task difficulty (McKiernan,aufman, Kucera-Thompson, & Binder, 2003) and have been

ound to be greater for subsequently remembered items than for-otten ones in healthy individuals (Daselaar, Prince, & Cabeza,004; Turk-Browne, Yi, & Chun, 2006). It may therefore be pos-ible that the relationship we observed between vmPFC activitynd memory performance in highly symptomatic patients rep-esented an alteration in the allocation of attention resourcesuring the memory task that interfered with successful encod-ng of the stimuli. Alternatively, abnormal activity of the vmPFC

ay disrupt memory performance through its connectivity withedial temporal lobe structures, such as the amygdala and hip-

ocampus. Indeed, abnormal functional connectivity betweenhe mPFC and amygdala has been previously reported in PTSDatients (Gilboa et al., 2004; Shin, Orr, et al., 2004; Shin et al.,005). Further neuroimaging studies of memory encoding inTSD, specifically designed to conduct connectivity analyses,re necessary to directly test this hypothesis.

Our results add to a growing body of knowledge linkingTSD to abnormalities in activity of frontal and temporal regions

n the context of a memory task (Bremner et al., 2003; Geuzet al., 2007; Shin, Shin, et al., 2004). However, within this lit-rature, the specific regions reported are not consistent; someTSD studies have reported differences in the hippocampusBremner et al., 2003; Geuze et al., 2007; Shin, Shin, et al.,004) and lateral frontal regions (Geuze et al., 2007), whilethers have observed, as we did, differences in activity in themPFC (Bremner et al., 2003) and amygdala (Shin, Shin, etl., 2004). These inconsistencies are likely to be related to dif-erences in the memory paradigms used in the different studiescued recall, paired associate learning and recognition memory),he memory stage (encoding and recognition), as well as theontrasts used in the analyses (deeply minus shallowly encodedtimuli, memory task minus baseline, and remembered minusorgotten). Our study also differs from previous ones in that wemployed faces rather than words. Although memory deficitsave been reported for both semantic and non-semantic material,recent meta-analysis suggests that behavioural impairments

re stronger for visual than verbal stimuli (Brewin et al., 2007).ore importantly, whereas these previous memory experiments

mployed a block design, we used an event-related analysis ofhe subsequent memory paradigm. This approach allowed uso specifically isolate the neural structures involved in success-ul memory formation from those associated with perceptual orask-related processes (Wagner et al., 1999).

.2. Amygdala and emotional memory

A large body of research in healthy humans has confirmed

he critical role of the amygdala in memory encoding for emo-ional stimuli (Adolphs, Cahill, Schul, & Babinsky, 1997; Cahill,abinsky, Markowitsch, & McGaugh, 1995; Cahill et al., 1996;olcos, LaBar, & Cabeza, 2004; Hamann, Ely, Grafton, & Kilts,

oasf

ogia 46 (2008) 1522–1531 1529

999; Sergerie et al., 2006), including fearful faces (Sergerie etl., 2006). In addition, neuroimaging studies in PTSD patientsave suggested that amygdala activity to fearful faces is exag-erated and positively correlated with PTSD symptom severityArmony et al., 2005; Rauch et al., 2000; Shin et al., 2005).he results of this study – namely a correlation between amyg-ala activation for successfully remembered fearful faces andAPS scores – combine and extend these two lines of evidencey showing that the relationship between amygdala activity andTSD symptom severity applied not only to emotion perceptionut also to the influence of emotion on other cognitive processes.owever, while the difference in amygdala activity between

earful remembered and fearful forgotten faces was greater inore highly symptomatic PTSD patients, we failed to find a

ignificant influence of CAPS scores on emotional memory.Although admittedly this result seems somewhat counter-

ntuitive, the presence of amygdala activation without significantifferences in memory performance between neutral and emo-ional stimuli has often been observed in neuroimaging studiesn healthy volunteers (e.g., Dolan, Lane, Chua, & Fletcher,000; Maratos, Dolan, Morris, Henson, & Rugg, 2001; Sharot,elgado, & Phelps, 2004; Taylor et al., 1998). In the contextf our study, we interpret our findings as suggesting that theelationship between PTSD symptoms and emotional memoryerformance may be more complex than could be measured byur simple memory recognition task. For example, it is possi-le that the observed amygdala activity may reflect differencesn subjective confidence, which may not have been apparent inur memory test (where participants were only asked to indicatehether the item was “old” or “new”). For example, Sharot et al.

2004) observed no difference in overall memory performanceetween emotional and neutral items in healthy individuals, butnstead found differences in the proportion of subjective remem-er and know responses (remember items being more vividlyecollected). Importantly, greater amygdala activity was asso-iated with the subjective feeling of remembering emotionaltimuli, compared to neutral ones (i.e., the interaction betweenemember vs. know and emotional vs. neutral). Interestingly, itas recently been shown that PTSD patients differ from controlsn their pattern of remember vs. know responses (Tapia, Clarys,l Hage, Belzung, & Isingrini, 2007). Future studies employingore complex memory paradigms, such as the use of remem-

er/know responses or confidence ratings should help furtherxplore this issue.

Alternatively, it is possible that the enhanced amygdala activ-ty we observed to subsequently remembered emotional facesn highly symptomatic PTSD individuals could have long-termehavioural consequences (Dolcos et al., 2004), which may notave been detectable in our immediate memory recognition test.ndeed, there is some evidence that PTSD patients and controlsiffer in their ability to consolidate emotional memories overong time intervals (Isaac et al., 2006).

In conclusion, the present study extends previous findings

n emotional perception and symptom provocation (Francati etl., 2007; Rauch et al., 2006), by showing that PTSD symptomeverity also modulates neural activity associated with success-ul memory formation. Specifically, we showed that activity in

Page 9

1 ychol

vhttis

A

eFtCsh

R

A

A

A

B

B

B

B

B

B

B

B

B

B

B

B

B

C

C

D

D

D

E

E

F

G

G

G

G

H

I

J

J

J

530 E.W. Dickie et al. / Neurops

entromedial prefrontal cortex predicts the forgetting of faces inighly symptomatic individuals, whereas amygdala responseso subsequently remembered fearful faces increase as a func-ion of symptom severity. These results further highlight themportance of taking into account individual differences whentudying neurocognitive processes in psychiatric populations.

cknowledgments

This project was funded by grants from Fonds de la recherchen sante du Quebec and the Canadian Psychiatric Researchoundation to J.L.A. and A.B. We are grateful to Dr. L. Gas-

on and the staff at Traumatys Clinic, as well as the nurses atharles LeMoyne Hospital for their invaluable help in recruiting

ubjects. We also thank the technical staff at the MNI for theirelp during scanning.

eferences

dolphs, R., Cahill, L., Schul, R., & Babinsky, R. (1997). Impaired declara-tive memory for emotional material following bilateral amygdala damage inhumans. Learning & Memory, 4(3), 291–300.

merican Psychiatric Association. (2000). Diagnostic and statistical manualof mental disorders (Text Revision) (4th ed.). Washington, DC: AmericanPsychiatric Association.

rmony, J. L., Corbo, V., Clement, M. H., & Brunet, A. (2005). Amygdalaresponse in patients with acute PTSD to masked and unmasked emotionalfacial expressions. American Journal of Psychiatry, 162(10), 1961–1963.

eck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961).An inventory for measuring depression. Archives of General Psychiatry, 4,561–571.

irmes, P., Brunet, A., Benoit, M., Defer, S., Hatton, L., Sztulman, H., et al.(2005). Validation of the Peritraumatic Dissociative Experiences Question-naire self-report version in two samples of French-speaking individualsexposed to trauma. European Psychiatry, 20(2), 145–151.

lake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Gusman, F. D.,Charney, D. S., et al. (1995). The Development of a Clinician-AdministeredPTSD Scale. Journal of Traumatic Stress, 8(1), 75–90.

randes, D., Ben-Schachar, G., Gilboa, A., Bonne, O., Freedman, S., & Shalev,A. Y. (2002). PTSD symptoms and cognitive performance in recent traumasurvivors. Psychiatry Research, 110(3), 231–238.

remner, J. D., Narayan, M., Staib, L. H., Southwick, S. M., McGlashan, T., &Charney, D. S. (1999). Neural correlates of memories of childhood sexualabuse in women with and without posttraumatic stress disorder. AmericanJournal of Psychiatry, 156(11), 1787–1795.

remner, J. D., Scott, T. M., Delaney, R. C., Southwick, S. M., Mason,J. W., Johnson, D. R., et al. (1993). Deficits in Short-Term-Memory inPosttraumatic-Stress-Disorder. American Journal of Psychiatry, 150(7),1015–1019.

remner, J. D., Staib, L. H., Kaloupek, D., Southwick, S. M., Soufer, R., & Char-ney, D. S. (1999). Neural correlates of exposure to traumatic pictures andsound in Vietnam combat veterans with and without posttraumatic stress dis-order: A positron emission tomography study. Biological Psychiatry, 45(7),806–816.

remner, J. D., Vythilingam, M., Vermetten, E., Southwick, S. M., McGlashan,T., Staib, L. H., et al. (2003). Neural correlates of declarative memoryfor emotionally valenced words in women with posttraumatic stress dis-order related to early childhood sexual abuse. Biological Psychiatry, 53(10),879–889.

rewer, J. B., Zhao, Z., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1998).Making memories: Brain activity that predicts how well visual experiencewill be remembered. Science, 281(5380), 1185–1187.

rewin, C. R., Kleiner, J. S., Vasterling, J. J., & Field, A. P. (2007). Mem-ory for emotionally neutral information in posttraumatic stress disorder: A

K

ogia 46 (2008) 1522–1531

meta-analytic investigation. Journal of Abnormal Psychology, 116(3), 448–463.

ritton, J. C., Phan, K. L., Taylor, S. F., Fig, L. M., & Liberzon, I. (2005). Cor-ticolimbic blood flow in posttraumatic stress disorder during script-drivenimagery. Biological Psychiatry, 57(8), 832–840.

runet, A., Weiss, D. S., Metzler, T. J., Best, S. R., Neylan, T. C., Rogers, C.,et al. (2001). The peritraumatic distress inventory: A proposed measure ofPTSD criterion A2. American Journal of Psychiatry, 158(9), 1480–1485.

ryant, R. A., Kemp, A. H., Felmingham, K. L., Liddell, B., Olivieri, G., Peduto,A., et al. (2007). Enhanced amygdala and medial prefrontal activation duringnonconscious processing of fear in posttraumatic stress disorder: An fMRIstudy. Human Brain Mapping,. May 24 [Epub ahead of print].

ahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). Theamygdala and emotional memory. Nature, 377(6547), 295–296.

ahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang, C., Keator, D., et al.(1996). Amygdala activity at encoding correlated with long-term, free recallof emotional information. Proceedings of the National Academy of Sciencesof the United States of America, 93(15), 8016–8021.

aselaar, S. M., Prince, S. E., & Cabeza, R. (2004). When less means more:Deactivations during encoding that predict subsequent memory. Neuroim-age, 23(3), 921–927.

olan, R. J., Lane, R., Chua, P., & Fletcher, P. (2000). Dissociable temporallobe activations during emotional episodic memory retrieval. Neuroimage,11(3), 203–209.

olcos, F., LaBar, K. S., & Cabeza, R. (2004). Interaction between the amygdalaand the medial temporal lobe memory system predicts better memory foremotional events. Neuron, 42(5), 855–863.

kman, P., & Friesen, W. (1976). Pictures of facial affect. Palo Alto, CA:Consulting Psychologists.

vans, A., Kamber, M., Collins, D., & MacDonald, D. (1994). In S. Shorvon,D. Fish, F. Andermann, & G. Bydder (Eds.), Magnetic resonance scanningand epilepsy (pp. 263–274). New York: Plenum.

rancati, V., Vermetten, E., & Bremner, J. D. (2007). Functional neuroimag-ing studies in posttraumatic stress disorder: Review of current methods andfindings. Depression and Anxiety, 24(3), 202–218.

euze, E., Vermetten, E., Ruf, M., de Kloet, C. S., & Westenberg, H. G. (2007).Neural correlates of associative learning and memory in veterans with post-traumatic stress disorder. Journal of Psychiatric Research,. Aug 13 [Epubahead of print].

ilbertson, M. W., Gurvits, T. V., Lasko, N. B., Orr, S. P., & Pitman, R. K. (2001).Multivariate assessment of explicit memory function in combat veteranswith posttraumatic stress disorder. Journal of Traumatic Stress, 14(2), 413–432.

ilboa, A., Shalev, A. Y., Laor, L., Lester, H., Louzoun, Y., Chisin, R., et al.(2004). Functional connectivity of the prefrontal cortex and the amygdala inposttraumatic stress disorder. Biological Psychiatry, 55(3), 263–272.

usnard, D. A., & Raichle, M. E. (2001). Searching for a baseline: Functionalimaging and the resting human brain. Nature Reviews Neuroscience, 2(10),685–694.

amann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D. (1999). Amygdalaactivity related to enhanced memory for pleasant and aversive stimuli. NatureNeuroscience, 2(3), 289–293.

saac, C. L., Cushway, D., & Jones, G. V. (2006). Is posttraumatic stress disorderassociated with specific deficits in episodic memory? Clinical PsychologyReview, 26(8), 939–955.

enkins, M. A., Langlais, P. J., Delis, D., & Cohen, R. A. (2000). Attentional dys-function associated with posttraumatic stress disorder among rape survivors.Clinical Neuropsychologist, 14(1), 7–12.

ohansen, V. A., Wahl, A. K., Eilertsen, D. E., Weisaeth, L., & Hanestad, B.R. (2007). The predictive value of post-traumatic stress disorder symptomsfor quality of life: A longitudinal study of physically injured victims ofnon-domestic violence. Health and Quality of Life Outcomes, 5, 26.

osephs, O., Turner, R., & Friston, K. (1997). Event-related fMRI. Human Brain

Mapping, 5(4), 243–248.

im, M. J., Chey, J., Chung, A., Bae, S., Khang, H., Ham, B., et al. (2007).Diminished rostral anterior cingulate activity in response to threat-relatedevents in posttraumatic stress disorder. Journal of Psychiatric Research,.May 30 [Epub ahead of print].

Page 10

ychol

K

K

K

L

L

L

L

M

M

M

R

R

R

S

S

S

S

S

S

T

T

T

T

T

V

V

W

W

WPTSD scale: A review of the first ten years of research. Depress Anxiety,

E.W. Dickie et al. / Neurops

ivling-Boden, G., & Sundbom, E. (2003). Cognitive abilities related topost-traumatic symptoms among refugees from the former Yugoslavia inpsychiatric treatment. Nordic Journal of Psychiatry, 57(3), 191–198.

oenen, K. C., Driver, K. L., Oscar-Berman, M., Wolfe, J., Folsom, S., Huang,M. T., et al. (2001). Measures of prefrontal system dysfunction in posttrau-matic stress disorder. Brain and Cognition, 45(1), 64–78.

oso, M., & Hansen, S. (2006). Executive function and memory in posttrau-matic stress disorder: A study of Bosnian war veterans. European Psychiatry,21(3), 167–173.

anius, R. A., Williamson, P. C., Densmore, M., Boksman, K., Gupta, M. A.,Neufeld, R. W., et al. (2001). Neural correlates of traumatic memories inposttraumatic stress disorder: A functional MRI investigation. AmericanJournal of Psychiatry, 158(11), 1920–1922.

iberzon, I., Taylor, S. F., Amdur, R., Jung, T. D., Chamberlain, K. R.,Minoshima, S., et al. (1999). Brain activation in PTSD in response to trauma-related stimuli. Biological Psychiatry, 45(7), 817–826.

indauer, R. J., Olff, M., van Meijel, E. P., Carlier, I. V., & Gersons, B. P. (2006).Cortisol, learning, memory, and attention in relation to smaller hippocam-pal volume in police officers with posttraumatic stress disorder. BiologicalPsychiatry, 59(2), 171–177.

undqvist, D., Flykt, A., & Ohman, A. (1998). The Karolinska DirectedEmotional Faces—CD ROM. Stockholm, Sweden: Department of ClinicalNeuroscience, Psychology Section, Karolinska Institute.

aldjian, J. A., Laurienti, P. J., Kraft, R. A., & Burdette, J. H. (2003). Anautomated method for neuroanatomic and cytoarchitectonic atlas-basedinterrogation of fMRI data sets. Neuroimage, 19(3), 1233–1239.

aratos, E. J., Dolan, R. J., Morris, J. S., Henson, R. N., & Rugg, M. D. (2001).Neural activity associated with episodic memory for emotional context.Neuropsychologia, 39(9), 910–920.

cKiernan, K. A., Kaufman, J. N., Kucera-Thompson, J., & Binder, J. R.(2003). A parametric manipulation of factors affecting task-induced deac-tivation in functional neuroimaging. Journal of Cognitive Neuroscience,15(3), 394–408.

auch, S. L., Shin, L. M., & Phelps, E. A. (2006). Neurocircuitry mod-els of posttraumatic stress disorder and extinction: Human neuroimagingresearch—Past, present, and future. Biological Psychiatry, 60(4), 376–382.

auch, S. L., vanderKolk, B. A., Fisler, R. E., Alpert, N. M., Orr, S. P., Savage, C.R., et al. (1996). A symptom provocation study of posttraumatic stress disor-der using positron emission tomography and script-driven imagery. Archivesof General Psychiatry, 53(5), 380–387.

auch, S. L., Whalen, P. J., Shin, L. M., McInerney, S. C., Macklin, M. L.,Lasko, N. B., et al. (2000). Exaggerated amygdala response to masked facialstimuli in posttraumatic stress disorder: A functional MRI study. BiologicalPsychiatry, 47(9), 769–776.

ergerie, K., Lepage, M., & Armony, J. L. (2006). A process-specific functionaldissociation of the amygdala in emotional memory. Journal of CognitiveNeuroscience, 18(8), 1359–1367.

harot, T., Delgado, M. R., & Phelps, E. A. (2004). How emotion enhances thefeeling of remembering. Nature Neuroscience, 7(12), 1376–1380.

heehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E.,et al. (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.):The development and validation of a structured diagnostic psychiatric inter-

W

ogia 46 (2008) 1522–1531 1531

view for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59(Suppl.20), 22–33 (quiz 34–57).

hin, L. M., Orr, S. P., Carson, M. A., Rauch, S. L., Macklin, M. L., Lasko,N. B., et al. (2004). Regional cerebral blood flow in the amygdala andmedial prefrontal cortex during traumatic imagery in male and female Viet-nam veterans with PTSD. Archives of General Psychiatry, 61(2), 168–176.

hin, L. M., Shin, P. S., Heckers, S., Krangel, T. S., Macklin, M. L., Orr, S.P., et al. (2004). Hippocampal function in posttraumatic stress disorder.Hippocampus, 14(3), 292–300.

hin, L. M., Wright, C. I., Cannistraro, P. A., Wedig, M. M., McMullin, K.,Martis, B., et al. (2005). A functional magnetic resonance imaging study ofamygdala and medial prefrontal cortex responses to overtly presented fearfulfaces in posttraumatic stress disorder. Archives of General Psychiatry, 62(3),273–281.

alairach, P., & Tournoux, J. (1988). Stereotactic coplanar atlas of the humanbrain. Stuttgart, Germany: Thieme.

apia, G., Clarys, D., El Hage, W., Belzung, C., & Isingrini, M. (2007).PTSD psychiatric patients exhibit a deficit in remembering. Memory, 15(2),145–153.

aylor, S. F., Liberzon, I., Fig, L. M., Decker, L. R., Minoshima, S., & Koeppe,R. A. (1998). The effect of emotional content on visual recognition memory:A PET activation study. Neuroimage, 8(2), 188–197.

urk-Browne, N. B., Yi, D. J., & Chun, M. M. (2006). Linking implicit andexplicit memory: Common encoding factors and shared representations.Neuron, 49(6), 917–927.

zourio-Mazoyer, N., Landeau, B., Papathanassiou, D., Crivello, F., Etard, O.,Delcroix, N., et al. (2002). Automated anatomical labeling of activations inSPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage, 15(1), 273–289.

asterling, J. J., Brailey, K., Constans, J. I., & Sutker, P. B. (1998). Attentionand memory dysfunction in posttraumatic stress disorder. Neuropsychology,12(1), 125–133.

asterling, J. J., Duke, L. M., Brailey, K., Constans, J. I., Allain, A. N., & Stuker,P. B. (2002). Attention, learning, and memory performances and intellec-tual resources in Vietnam veterans: PTSD and no disorder comparisons.Neuropsychology, 16(1), 5–14.

agner, A. D., Koutstaal, W., & Schacter, D. L. (1999). When encodingyields remembering: Insights from event-related neuroimaging. Philosoph-ical Transactions of the Royal Society of London Series B-BiologicalSciences, 354(1387), 1307–1324.

agner, A. D., Schacter, D. L., Rotte, M., Koutstaal, W., Maril, A., Dale, A.M., et al. (1998). Building memories: Remembering and forgetting of ver-bal experiences as predicted by brain activity. Science, 281(5380), 1188–1191.

eathers, F. W., Keane, T. M., & Davidson, J. R. (2001). Clinician-administered

13(3), 132–156.illiams, L. M., Kemp, A. H., Felmingham, K., Barton, M., Olivieri, G.,

Peduto, A., et al. (2006). Trauma modulates amygdala and medial prefrontalresponses to consciously attended fear. Neuroimage, 29(2), 347–357.