Serotonin 2A Receptors, Citalopram and Tryptophan-Depletion: a Multimodal Imaging Study of their Interactions During Response Inhibition.

0.56 MB·English

by  Macoveanu, Julian

Neuropsychopharmacology(2013)38,996–1005 OPEN
&2013AmericanCollegeofNeuropsychopharmacology. Allrightsreserved0893-133X/13
www.neuropsychopharmacology.org
Serotonin 2A Receptors, Citalopram and Tryptophan-
Depletion: a Multimodal Imaging Study of their Interactions
During Response Inhibition
(cid:2)
(cid:2)
(cid:2)
(cid:2) Julian Macoveanu1,2, Bettina Hornboll1,2, Rebecca Elliott3, David Erritzoe2,4, Olaf B Paulson1,2,4,
(cid:2)
(cid:2) Hartwig Siebner1,2, Gitte M Knudsen2,4 and James B Rowe*,2,5
(cid:2)
(cid:2)
(cid:2) 1DanishResearch Centrefor MR,Copenhagen University Hospital, Hvidovre, Denmark; 2Center for Integrated Molecular Brain Imaging,
(cid:2)
(cid:2) Copenhagen,Denmark; 3Neuroscience andPsychiatry Unit,University ofManchester, Manchester,UK;4Neurobiology Research Unit,
(cid:2)
(cid:2) RigshospitaletandUniversity ofCopenhagen, Copenhagen,Denmark; 5Cambridge UniversityDepartment ofClinicalNeurosciences,
(cid:2)
(cid:2) Cambridge,UK
(cid:2)
(cid:2)
(cid:2)
(cid:2)
(cid:2)
(cid:2)
(cid:2)
(cid:2) Poorbehavioralinhibitionisacommonfeatureofneurologicalandpsychiatricdisorders.Successfulinhibitionofaprepotentresponsein
(cid:2)
(cid:2) ‘NoGo’paradigmsrequirestheintegrityofboththeinferiorfrontalgyrus(IFG)andtheserotonergicsystem.Weinvestigatedindividual
(cid:2)
(cid:2) differencesinserotonergicregulationofresponseinhibition.In24healthyadults,weused18F-altanserinpositronemissiontomographyto
(cid:2)
(cid:2) assesscerebral5-HT receptors,whichhavebeenrelatedtoimpulsivity.Wetheninvestigatedtheimpactoftwoacutemanipulationsof
(cid:2) 2A
(cid:2) brain serotonin levels on behavioral and neural correlates of inhibition using intravenous citalopram and acute tryptophan depletion
(cid:2)
(cid:2) during functional magnetic resonance imaging. We adapted the NoGo paradigm to isolate effects on inhibition per se as opposed to
(cid:2)
(cid:2)(cid:2) otheraspectsoftheNoGoparadigm.SuccessfulNoGoinhibitionwasassociatedwithgreateractivationoftherightIFGcomparedto
(cid:2)
(cid:2) control trials with alternative responses, indicating that the IFG is activated with inhibition in NoGo trials rather than other aspects of
(cid:2)
(cid:2) invoked cognitive control. Activation of the left IFG during NoGo trials was greater with citalopram than acute tryptophan depletion.
(cid:2)
(cid:2) Moreover,withtheNoGo-typeofresponseinhibition,therightIFGdisplayedaninteractionbetweenthetypeofserotonergicchallenge
(cid:2)
(cid:2) andneocortical5-HT receptorbinding.Specifically,acutetryptophandepletion(ATD)producedarelativelylargerNoGoresponsein
(cid:2) 2A
(cid:2) the right IFGin subjects with low5-HT BP but reduced theNoGo response in those withhigh 5-HT BP . Theselinks between
(cid:2) 2A P 2A P
(cid:2)(cid:2) serotonergicfunctionandresponseinhibitioninhealthysubjectsmayhelptointerpretserotonergicabnormalitiesunderlyingimpulsivity
(cid:2)
(cid:2) in neuropsychiatricdisorders.
(cid:2)
(cid:2) Neuropsychopharmacology (2013)38, 996–1005;doi:10.1038/npp.2012.264;publishedonline 30January2013
(cid:2)
(cid:2)
(cid:2) Keywords: inhibition; serotonin; citalopram;5-HT receptor;altanserin; fMRI
2A
INTRODUCTION Chikazoe et al, 2007; Langenecker et al, 2007; Simmonds
et al, 2008; Zheng et al, 2008; Chikazoe et al, 2009), and
Impaired response inhibition is common in neuropsychia-
electrophysiology (Swann et al, 2009). Many studies
tric and neurodegenerative disorders. One of the most
emphasize right IFG, but left IFG is also involved (Rubia
studied forms of inhibition is restraint of a prepotent
et al, 2001; Swick et al, 2008).
responseexemplifiedby‘NoGo’tasks(IversenandMishkin,
Responseinhibitionalsoshowsneurochemicalspecificity.
1970;Konishietal,1999).Specificanatomicalstructuresare
NoGo inhibition is strongly associated with integrity of
implicated in response inhibition including inferior frontal
serotonergic system in humans and animals (Eagle et al,
gyrus (IFG) (Wager et al, 2005; Levy and Wagner, 2011),
2008). This contrasts with noradrenergic modulation of the
evidenced by IFG lesions (Aron et al, 2003; Swick et al,
stop response required by stop-signal tasks (Chamberlain
2008), neuroimaging (Konishi et al, 1998; Konishi et al,
et al, 2009). Global depletion of serotonin (5-HT) leads to
1999;Asahietal,2004;Del-Benetal,2005;Rubiaetal,2005;
impulsivityinrats(Harrisonetal,1999;Masakietal,2006)
andhumans(Walderhaugetal,2002).Althoughbehavioral
*Correspondence:DrJBRowe,CambridgeUniversityDepartmentof
NoGo effects of serotonergic interventions are often mild
Clinical Neurosciences, Herchel-Smith Building for Brain and Mind
or absent in humans, neuroimaging has revealed altered
Sciences, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK. Tel:
activity of underlying fronto-striatal circuits (Rubia et al,
[phone],Fax:[phone],
E-mail: [email protected] 2005; Evers et al, 2006). For example, acute tryptophan
Received 27 March 2012; revised 1 November 2012; accepted 5 depletion (ATD) reduces frontal cortical activations (Rubia
November2012;acceptedarticlepreviewonline19December2012 et al, 2005; Lamar et al, 2009), whereas the selective
Serotoninmodulationandinhibition
JMacoveanuetal
997
serotonin uptake inhibitor (SSRI) citalopram enhances MATERIALS AND METHODS
them (Del-Ben et al, 2005).
Participants, Task, and Pharmacological Challenges
There are marked individual differences in both im-
pulsivityandserotonin.Impulsiveclinicalpopulationswith Twenty-four right-handed adults (age 20–40 years) were
serotonergic deficits include ADHD (Zepf et al, 2008), recruited from a cohort of volunteers with 18F-altanserin
borderline personality disorder (Leyton et al, 2001), and PET brain imaging. Subjects were part of the Center for
frontotemporal dementia (Huey et al, 2006), whereas 5- Integrated Molecular Brain Imaging (CIMBI) database that
HT2AreceptorabnormalitieshavebeenlinkedtoTourette’s includes extensive enquiries into the participants’ past and
syndrome (Haugbol et al, 2007) and obsessive–compulsive present diagnosis or treatment of psychiatric illness. Based
disorder (Adams et al, 2005). Within the healthy popula-
on these entry records, which were reconfirmed at study
tion,variationincerebral5-HT receptorbindingispartly
2A recruitment, we excluded subjects with a history of
genetically determined (Pinborg et al, 2008) with an
stimulant abuse, neurological disorder, or a psychiatric
influence on behavioral impulsivity (Nomura and
disorder requiring specialist referral or treatment. Written
Nomura, 2006) if not self-report impulsivity (Frokjaer
informed consent was obtained and the study approved by
et al, 2008). Individual differences may not be marked
the Copenhagen Ethics Committee (KF01-2006–20). One
under normal testing conditions, but chronic serotonergic
subject was excluded due to an outlier 5-HT BPp value
status (whether a genetically determined trait or a result of 2A
(42.5SD)identifiedonre-estimationofPETdatafollowing
chronic environmental influences) may influence the
recruitment. This subject also had outlying error rates on
change of impulsivity in response to acute challenges, such
the behavioral task (2.5–4SD from the group mean). A
as stress, depression, or medication. This requires analysis
secondsubject withdrew fromthestudybeforecompletion.
of interactions between acute and chronic serotonergic
Thus complete data sets from 22 participants (eight female
states (including state–trait interactions) that we investi-
participants, mean age±SD of 31.5±6.2) were included in
gated in this study.
further analyses.
This study addressed two issues. The first aim was to
Participants performed a modified NoGo task with three
better understand the functional, anatomical, and pharma-
trial types: (a) ‘Go’ trials, pressing a button to a visual cue
cological basis of response inhibition. Our principal
(yellow square); (b) AltGo, pressing a different button to a
hypothesis was that chronic 5-HT receptor availability,
2A
inferred from 18F-altanserin steady-state binding measure- different visualcue(yellow circle); (c)‘NoGo’, requiring no
ments (BP ), influences the effect of acute state manipula- response (yellow triangle). Trial frequencies were 70, 15,
P
tions on response inhibition. Such an interaction would and 15%, respectively. Stimuli were presented for 1000ms
explain some of the behavioral and imaging differences with 500ms intervals in a pseudorandom order during two
between healthy individuals (Del-Ben et al, 2005; Rubia blocks of 5min.
etal,2005;Eversetal,2006;Lamaretal,2009)andpatients Errorratesandreactiontimeswereenteredintorepeated
(LeMarquand et al, 1998; Zepf et al, 2008). We studied measures analyses of variance (PASW-SPSS17 software,
healthy subjects with 18F-altanserin positron emission Chicago). Drug session and trial-type were within-subjects
tomography (PET) and functional MRI sessions that factors with three and two levels, respectively. 5-HT BPp
2A
differed only in 5-HT levels by (a) increased 5-HT was a between-subjects covariate. Huynh–Feldt correction
neurotransmission by intravenous administration of the fornon-sphericity wasused whereappropriate andpo0.05
SSRIcitalopram;(b)reducedbrain5-HTsynthesisviaacute considered significant.
dietary depletion of the 5-HT precursor tryptophan (ATD); Subjects underwent three sessions of pharmacological
(c) a control state without drug intervention. We were fMRI, at least 1 week apart, and a fully counterbalanced
primarily interested in interactions between acute changes orderbypseudorandompermutation.Thepharmacological
in serotonergic transmission, 5-HT BPp, and functional conditions were: (a) Control session, without intervention;
2A
magnetic resonance imaging (fMRI) correlates of response (b) SSRI session; and (c) ATD session. For SSRI sessions,
inhibition. citalopram was administered intravenously at 20mg/h
The second issue was to dissect cognitive components of starting 2h before scanning, with maintenance dose during
the NoGo task and identify whether 5-HT specifically the fMRI session at 8mg/h (B50mg in total). For the ATD
modulates the response inhibition component separable session,participantswereaskedtofollowalow-proteindiet
from other aspects of the task. For example, NoGo trials on the day before scanning. On the test day, they ingested
includelow frequency events thattriggerreflexivereorient- 75g tryptophan-free powdered mixture of amino acids
ing to task relevant stimuli, require greater cognitive dissolved in water (XLYS, TRY Glutaridon, SHS Interna-
control, and lead to response adjustment (Ridderinkhof tional) and performed the fMRI session 5h later. Blood
etal,2004;Kenneretal,2010;LevyandWagner,2011).The samples were taken upon arrival and immediately before
response adjustment in classical NoGo trials is to withhold the fMRI to determine plasma amino acids and prolactin.
action,butthesubjectsmightalternativelybeaskedtoshift The blood sample acted both as a control on our
to a different response (Mostofsky and Simmonds, 2008). pharmacological treatments, and as a biochemical index
We therefore included a low frequency of trials in which oftheserotonergicbasisofimagingeffectsintheabsenceof
subjects update to a different motor response (‘alternative marked behavioral change. Prolactin levels may be in-
go’, AltGo). We predicted that serotonergic manipulations creased by acute tryptophan depletion in susceptible
would mainly influence the inhibitory component of the individuals(Wingroveetal,1999)astheycanbycitalopram
NoGo paradigm, revealed by differential activations (Attenburrow et al, 2001; Del-Ben et al, 2005; McKie et al,
between NoGo and AltGo trials. 2005). In many previous studies, the behavioral and
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
998
cognitive changes induced by ATD have been attributed to were realigned to the first volume and co-registered to the
specific effects on the serotonergic system (Ardis et al, structural MPRAGE brain scan. The MPRAGE scan was
2009). This view has been challenged recently (van normalized to a T1 template in standard space (MNI
Donkelaar et al, 2011), but we correlated the changes in template) using linear and non-linear transformations.
tryptophanandprolactinwith5-HT BPpinsupportofthe Normalization parameters were applied to functional
2A
hypothesis that the ATD effect is at least partly due to volumes before smoothing with an isometric Gaussian
effects on the serotonergic systems. On each of the fMRI kernel with full width half maximum 8mm.
sessions, participantscompleted amodifiedDanish version Subject-specific first-level general linear models of fMRI
of the Profile of Mood States (POMS) questionnaire included four separate regressors for ‘Go’, ‘AltGo’, ‘NoGo’,
(McNair et al, 1971) thrice to assess current mood and ‘NoGo commission errors’, and 40 nuisance regressors
uponarrivalimmediatelybeforefMRIandaftercompletion to correct for physiological noise related to pulse ((cid:2)10),
of the tests. respiration((cid:2)6),andmovement((cid:2)24)(Lundetal,2005).
Contrasts of interest were entered into group level random
18F-altanserin PET effects analyses using repeated measures ANOVA.
Two second-level flexible factorial models were used to
18F-altanserinPETwasusedtoestimatethesubject-specific examine main effects of trial-type and drug as well as
neocortical 5-HT receptor binding relative to plasma interactions between individual differences and experimen-
2A
(BP ) using standardized published protocols (Pinborg tal conditions. The first model included regressors expres-
p
et al, 2003; Adams et al, 2004; Svarer et al, 2005) and a sing ‘NoGo vs Go’ and ‘AltGo vs Go’ contrasts in each of
maximum dose of 3.7MBq/kg bodyweight. Reconstruction, three drug session (six total). This model also included
attenuation, and scatter correction procedures were con- mean corrected linear and quadratic functions of the
ducted according to this published protocol using cerebel- average neocortical 5-HT BP (12 total). These functions
2A P
lum as a reference region (Pinborg et al, 2003). The were separate for each drug condition (enabling the
outcome parameter for regional 5-HT receptor binding exploration of interactions between acute and chronic
2A
was the binding potential relative to plasma (BP ). To serotonergic functions). This model explains individual
p
estimate regional BP , PET images and structural T1- differences in terms of prior 5-HT BPp.
p 2A
weightedMRimageswereco-registered(Adamsetal,2004) Asecondmodel included ‘subject’ asanadditionalfactor
and normalized. We applied automatic parcelation of PET with 22 subject-specifying regressors. This adjusts for
images using standardized volumes of interest delineated individual subject differences that may or may not be
on transaxial MRI slices (Svarer et al, 2005) and derived related to trait serotonergic function, but prevents inter-
neocortical estimatesusing the volume-weightedaverageof pretation of the main effects of BP (a between subjects
P
eight regions (orbitofrontal, medial inferior frontal, super- factor). Both of the second-level ANOVAs were corrected
ior frontal, superior temporal, medial inferior temporal, for non-sphericity using pooled estimates of non-indepen-
sensorymotor,parietal,andoccipitalcortex).Asinalarger dent unequal variance over suprathreshold voxels.
sample(Erritzoeetal,2010),therewerehighcorrelationsof For whole-brain comparisons, we applied family-wise
BP among cortical regions of interest (ROIs) (all r40.8). errorcorrectionpo0.05.ForhypothesesregardingtheIFG,
P
We therefore used an average neocortical gray matter BP wedefinedabilateralROIthatincludedtheoperculumand
P
for subsequent correlations with the functional MRI data pars triangularis of the IFG (Automatic Anatomical Label-
(this correlated with BP in the IFG, r40.95). ing and WFU-Pickatlas (Lancaster et al, 2000), and used
P
‘small volume correction’ family-wise error po0.05. Owing
tolikelyfunctionalandanatomicalheterogeneitywithinthis
MRI Acquisition and Analysis
IFGregion,wereportvoxel-wiseresultsanddonotaverage
WeusedaSiemens3TTrioscannerwitheight-channelhead responses across this region.
array coil for functional MRI with whole-brain coverage,
high-resolution structural MRI, and arterial spin labeling
(ASL) perfusion-weighted images. fMRI used BOLD-sensi-
tiveT2*-weightedecho-planarimages(repetitiontime2.5s, RESULTS
echotime 26ms, flipangle 901) with41(cid:2)3mm slices (25%
Behavior, Biochemistry, and Perfusion
gaps), 192(cid:2)192mm field-of-view. High-resolution 3D
structural T1-weighted spin-echo images used an Magneti- CommissionerrorsdifferedbetweenNoGoandAltGotrials
zation-Prepared Rapid Acquisition Gradient Echo (Figure 1a; effect of trial-type F1,20¼5.0, po0.05), but
(MPRAGE) sequence (TI/TE/TR¼800/3.93/1540ms, flip errors were similar across all three acute drug states (effect
angle 91, 256(cid:2)256(cid:2)192 isotropic voxels). ASL perfusion- of drug F2,40¼2.5, ns). There was neither an interaction
weighted images (TR¼3.4s, TE¼19.3ms, TI¼200–3000, between trial-type and drug effects on error rates
200ms intervals, 26 slices, voxel size¼5.0(cid:2)5.0(cid:2)4.0mm, (F2,40¼2.3, ns) nor a higher order interaction between
320(cid:2)160(cid:2)104mm field-of-view) used vessel suppression drug,trial-type,and5-HT BP (F2,40¼2.0,ns).Although
2A P
with bipolar gradients (b¼6s/mm2). ASL images were therewasnomaineffectof5-HT BP onerrorrate(Fo1),
2A P
calculated using FABBER with spatial priors (www.fmrib. therewasaninteractionbetweentrial-typeand5-HT BP
2A P
ox.ac.uk/fsl/fabber)andpermutationstestingfordifferences (F1,20¼5.2, po0.05) in their effect on error rates: 5-HT
2A
between drug conditions. BP correlatedweaklypositivewitherrorratesintheNoGo
P
Preprocessing and statistical analysis used SPM5 (www. trials (r(22)¼0.26, ns) without a significant correlation on
fil.ion.ucl.ac.uk/spm/software/spm5). Functional images the AltGo trials (r(22)¼ (cid:3)0.1, ns).
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
999
Figure 1 (a) Commission error rates (group mean and SD) on the Figure 2 (a) The difference in prolactin levels between baseline (T0)
NoGo and AltGo trials, under control condition (no drug) and after beforeadministrationofcitalopram(Cit)orATDandimmediatelypriorto
citalopram or ATD. (b) Reaction times (group mean and SD) for AltGo MRIscanning(T1).GroupmeanvalueswithSEM.(b)ATDreducedserum
andGotrials,inthethreedrugsessions(Statisticalanalysisinresultssection: tryptophan by 75%, comparable with previous studies of this method.
Behavior,biochemistry,andperfusion). Group mean values with SEM (Statistical analysis in results section:
Behavior,biochemistry,andperfusion).
The reaction time was longer for AltGo than Go trials
(F1,20¼6.7, po0.05, Figure 1b). There was no main effect AnrmANOVAofthePOMSquestionnairewasconducted
of drug (Fo1, Figure 1b) or 5-HT BP (Fo1). There with two factors: 5-HT intervention (ATD, citalopram, and
2A P
were no interactions between 5-HT BP and trial-type or control) and time (at session baseline and after fMRI
2A P
drug,andnohighorderinteractionbetweenallthreefactors session).Therewasaneffectoftimeforanger/hostilitywith
(all Fo1). lower scores after scanning compared to pre-scanning
Baseline prolactin levels correlated highly between baseline (F(12)¼6.98, p¼0.02). There was no significant
sessions (r¼0.80, n¼19, po0.001). ANOVA of prolactin interaction between intervention and time in any of the
levelsrevealednomaineffectofdrug(Fo1)ortimewithin reported mood states.
session from baseline to scanning (F1,17¼2.86, ns). There
was however a trend of interaction between drug session
and the change in prolactin from session baseline to onset
fMRI Results
of scanning (F1,17¼3.0, po0.1, Figure 2a), such that
prolactin only increased following citalopram. The 5-HT Inhibiting vs updating a motor response. We first
2A
BP did not correlate with prolactin levels (Fo1) but did examined the low frequency events in which visual cues
P
tend to influence the effect of drug on the change in trigger reorientation to task relevant stimuli, invoke
prolactin between baseline and pre-scanning (F1,17¼3.3, cognitive control, and a response adjustment, including
po0.1). Post hoc tests confirmed that higher 5-HT BP (on NoGo trials) the inhibition of the prepotent response.
2A P
wasassociatedwithasmallerchangeinprolactinafterATD Averaging ‘NoGo and AltGo vs Go’ trials across the three
(r¼ (cid:3)0.39, n¼21, po0.05) but not after citalopram pharmacological sessions (Figure 3a) identified transient
(r¼0.0, ns). The ATD protocol reduced serum tryptophan activations of prefrontal, premotor, parietal, and inferior
levels by 75% (paired t-test, t¼11.2, df¼21, po0.001, occipitotemporalcortex,andextensiveactivationintheIFG
Figure 2b), suggesting reductions in central tryptophan bilaterally (Table 1, Figure 3a).
bioavailability (Williams et al, 1999; Blokland et al, 2002). Next, to identify activations related to inhibition, we
ASL showed no significant differences in cerebral compared regional activity between the NoGo and AltGo
perfusion during drug vs no-drug sessions, either with trials (‘NoGo vs AltGo’, averaging across all drug sessions,
whole-brainanalysis(FWEpo0.05corrected)orwithinthe ROI analysis, Table 2, Figure 3b). There was differential
IFG-ROI (FWE po0.05 corrected or uncorrected threshold activation of the right IFG (activations for each drug
po0.001). condition are shown in Figure 3c and for each trial-type in
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1000
Figure3 (a)SPM(t)mapthresholdedatFWEpo0.05forthecontrastof(NoGoandAltGo)trialsvsGotrials.FurtheractivationsinIFGwereseen
within the predefined ROI (see text). Activations are rendered onto a representative brain in MNI space, in lateral views, and confirm the widespread
differentialactivationfollowinglowfrequencystimuliandresultingcognitiveprocessesofcontrol,updating,orinhibitingaresponse.(b)ActivationsoftheIFG
relatedtoresponseinhibition,especiallyontheright,arerevealedbythecontrastofNoGovsAltGotrials,hereillustratedatpo0.001(peaksarepo0.05
FWEcorrectedwithin theIFGregion ofinterest).TheinsetsliceshowsrightIFGaty¼26.(c)Parameterestimates forvoxel40,26,4intherightIFG
(contrastofNoGovsAltGobydrugsession)undercontrol(nodrug),citalopram(cit)andATD.(d)Parameterestimatesforvoxel40,26,4inrightIFGfor
NoGoandAltGotrials(averagingoverdrugsession).(e)Parameterestimatesforvoxel40,26,4inrightIFGshownseparatelyforeachcombinationofdrug
and trial-type (NG¼NoGo, Alt¼AltGo, N¼no drug,C¼citalopram, A¼ATD). Relativeactivation wasgreater forNoGo trials vsAltGo trialsunder
control and citalopram sessions. Parameter estimates shown in C–E sum to zero across the group, from the flexible factorial design, and are scaled in
arbitraryunits.Pinkerrorbarsindicate90%confidenceintervalsofthemeans.
Figure 3d). In the right IFG, ATD abolished the differences Serotonergic challenges. Based on previous studies
between NoGo and AltGo trials (see below and Figure 3e). (Del-Ben et al, 2005; Rubia et al, 2005), we predicted that
The reverse contrast between AltGo and NoGo trials citalopram and ATD would have differential effects on
revealedactivationsassociatedwithupdatingorswitchinga activations associated with response inhibition. For the
response in a broad ‘motor network’ of left motor and NoGo (vs Go) trials, there was greater activation in left IFG
premotor cortex, right cerebellum, putamen, thalamus, and withcitalopramthanATD((cid:3)48,10,12,t¼3.97,33voxels,
smaller clusters of activation in the left cingulate cortex, FWE p¼0.05). For AltGo trials (vs Go), there were no
bilateral insula, and left cerebellum. Bilateral activations significant interactions between response updating and
were also seen posterolaterally in the IFG bilaterally at the citalopram vs ATD. However, the difference between NoGo
junctions between areas 6 and 44 (Table 3). andAltGotrialsdidnotitselfdifferbetweencitalopramand
We looked for regions in which the effects of response ATD sessions. In other words, for the inhibition of a
inhibition and response updating were modulated by response, but not the updating of a response, there were
5-HT BP .ForNoGo(vsGo)andforAltGo(vsGo)trials; voxels in the left IFG that were differentially sensitive to
2A P
no significant interactions were found. When examined citalopram and ATD. This was corroborated by an
separatelyforNoGoandAltGotrials(vsGo);therewereno F-contrast testing the effect ofeither serotonergic challenge
linear or quadratic effects of 5-HT BP . The difference (vs no-drug or the other challenge) within either AltGo–Go
2A P
between NoGo and AltGo trials did not depend on 5-HT or NoGo–Go contrast: identifying the left IFG at—48, 8, 12
2A
BP when averaged across drug sessions. (trend po0.1 corrected, F4,114¼6.8).
P
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1001
Table1 SignificantClusterPeaksfromtheContrastof‘NoGoand Table 3 Significant Cluster Peaks from the Contrast of ‘AltGo vs
AltGo vs Go Trials’ (See Also Figure 3), Thresholded at po0.05 NoGo’, Thresholded at po0.05 (Corrected for Whole-Brain
(t45.0, Corrected for Whole-Brain Comparisons, Cluster Multiple Comparisons or Where Marked by **, Corrected for
Minimum 10 Voxels) Multiple Comparisons Within the IFG ROI)
Region x y z t Inferiorfrontalgyrus (cid:3)58 8 20 7.94
60 10 20 5.42**
Inferiorfrontalgyrus 42 6 31 9.51
Precentralgyrus/IFG (cid:3)58 2 29 9.61
54 6 35 9.47
Sensorimotorcortex (cid:3)38 (cid:3)28 59 16.5
38 24 29 6.9
Supramarginalgyrus (cid:3)58 (cid:3)26 47 13.4
36 20 13 6.63
Rolandicoperculum (cid:3)42 (cid:3)8 21 13.2
(cid:3)46 0 29 10.7
60 20 21 8.86
(cid:3)32 24 3 8.08
SMA (cid:3)2 (cid:3)14 57 7.38
Frontaloperculum 32 22 0 11.41
Putamen (cid:3)28 (cid:3)17 3 7.10
(cid:3)30 22 4 11.03
Cerebellum 20 (cid:3)54 (cid:3)13 14.8
Middlefrontalgyrus 42 30 22 7.42
(cid:3)26 (cid:3)60 (cid:3)15 6.74
Precentralgyrus (cid:3)28 (cid:3)16 63 23.4
(cid:3)52 (cid:3)4 41 13.3
30 (cid:3)1 59 15.6 significant voxels were identified by these contrasts using
either whole-brain correction or just within the IFG ROI.
SMA (cid:3)6 4 51 16.2
However, a higher order interaction was observed: the
4 (cid:3)2 55 14.1
effect of ATD (ATD vs no-drug or ATD vs citalopram) on
Sensorycortex (cid:3)46 (cid:3)38 51 22.0
activations related to specific response inhibition (NoGo vs
Superiorparietallobule (cid:3)26 (cid:3)62 57 21.2 AltGo)dependedonthe5-HT receptorBP (Figure4a)in
2A P
38 (cid:3)52 63 15.6 the right IFG at 38, 14, 28 (F2,114¼13.18, po0.05 FWE
Inferiorparietallobule (cid:3)34 (cid:3)44 47 20.8 corrected within the ROI). Figure 4b shows parameter
34 (cid:3)48 51 15.0 estimates for the relationship between 5-HT2A BPP and IFG
activation as function of drug and trial-type. This means
Middleoccipitalgyrus (cid:3)42 (cid:3)82 3 16.4
that the BP level determined the effect of ATD on
44 (cid:3)80 1 18.5 P
activations associated with inhibition in the NoGo trials
Superiortemporalgyrus 64 (cid:3)44 25 11.4
(Figure 4c).
Inferiortemporalgyrus 44 (cid:3)60 (cid:3)5 14.0
Thalamus (cid:3)12 (cid:3)18 23 7.16
10 (cid:3)18 21 6.94
DISCUSSION
Clustersubpeaksareseparatedby420mm,exceptforpeakswithintheROI
We have shown that individual differences in neocortical
definedbytheinferiorfrontalgyriandfrontalopercula,separation48mm.
5-HT receptorbindingarerelatedtothesubsequenteffect
Coordinatex,y,zvaluesinstandardspaceusingtheMNItemplate,with 2A
accompanyingt-statistics. ofacute5-HTmanipulationsontheneuralcorrelatesofthe
successfulresponseinhibitioninhumanIFG.Weconfirmed
Table 2 SignificantCluster Peaksfromthe Contrast of‘NoGovs the association of NoGo inhibitionwiththe IFG, and found
AltGo’ (See Also Figure 3b), Thresholded at po0.05 (Corrected that citalopram and ATD differentially modulated IFG
for Muliple Comparisons within an ROI Defined by the Inferior activation for NoGo inhibition. In relation to our principal
Frontal Gyri and Frontal Opercula, Peaks 48mm Apart) hypothesis,thenovelfindingwasthattheNoGoresponsein
IFG under the different pharmacological challenges was
Region x y z t differentially related to the neocortical 5-HT receptor
2A
binding.Critically,theinteractionbetween5-HT receptor
2A
Inferiorfrontalgyrus 40 26 4 4.81 binding and drug effects was confined to the difference in
52 20 34 3.70 regional activity between the NoGo and AltGo trials,
(cid:3)46 25 26 3.67 providing evidence for serotonergic modulation of the
Inferiorfrontalsulcus 42 18 30 4.28 inhibitory component within the NoGo trials.
Response inhibition and switching to an alternative
32 34 28 4.36
response did not elicit activations that correlated with
(cid:3)40 16 28 3.20
5-HT receptor binding. However, when contrasting these
2A
two trial types, the difference in error rates and the drug
interaction with NoGo-specific activations both correlated
We tested for interactions between the effects of with 5-HT receptor binding. Why were the effects of
2A
serotonergic challenges and the 5-HT BP , on activations 5-HT BP not more prominent, given the previous
2A P 2A P
associated with response inhibition and updating. SPM(t) evidence for serotonergic modulation of inhibition? We
contrasts were used to identify voxels in which BP suggest that long-term autoregulation of post-synaptic
P
modulated the differential effect of citalopram or ATD (vs efficacy may reduce the baseline effects of 5-HT BPp on
2A
no-drug), for NoGo (vs Go), and for AltGo (vs Go). No inhibitory control, at least within the normal range, and in
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1002
Figure 4 The inhibition of responses (NoGo vs AltGo) was associated with a variable degree of activation in the right IFG (peak voxel 38, 14, 28)
accordingtotheserotonergicchallengeandthe5-HT BP .Althoughcitalopramdidnotdifferfromcontrolsessions,theeffectofATDwasreducedwith
2A P
increasing 5-HT2A BPP. (a) Surface rendered SPM(t) map (thresholded po0.05 FWE within IFG) indicating the significant cluster in right IFG. (b) The
parameterestimatesforthetwotrialtypesatthepeakrightIFGvoxel:NG¼NoGoandAlt¼AltGo,foreachofthethreedrugsessions,N¼Nodrug,
C¼citalopram, A¼ATD. (c) The difference in bold response at the peak IFG voxel according to 5-HT BP (normalized volume weighted mean
2A P
neocortical5-HT BP fromaltanserinPETdata)foreachdrugsession.Linearregressionlinesareshownseparatelyforeachdrugsession,withposthoc
2A P
correlationsshownseparatelyforeachpharmacologicalchallengeasR2.ATDdifferedsignificantlyinslopefromcitalopram(po0.05,correctedstatistical
inferencefromtheSPMcontrast).
contrast to the neuropsychiatric populations (Leyton et al, effects, we identified the neural processes that are
2001; Huey et al, 2006; Zepf et al, 2008). specifically related to withholding a prepotent response by
Although ATD did not result in a change in response comparing NoGo and AltGo trials. The restraint from a
times or accuracy, ATD produced a relatively larger prepotent response (NoGo) was associated with activation
response in the right IFG of subjects with low 5-HT of the right IFG, even after controlling for other factors
2A
BP . Conversely, ATD reduced activation in right IFG in including the presentation of low-frequency stimuli, pro-
P
those with high 5-HT BP . This differential response to cessing of these task relevant stimulus attributes, engage-
2A P
ATDsuggestsanon-linearrelationshipbetweenneocortical mentofadditionalcognitivecontrol,andachange(update)
5-HT andtheefficiencyofrightIFG inresponseinhibition, of response.
with opposite effects of ATD in subjects with low vs high The right IFG has been previously associated with
levels of 5-HT BP One implication is that drug effects switching to alternative responses and withholding re-
2A P.
maybeexaggeratedinsomepatients.ForATD,thisappears sponses (Kenner et al, 2010; Dodds et al, 2011). One
to be the case for disorders of impulsivity (Leyton et al, interpretation is that updating to an alternate response
2001; Huey et al, 2006; Zepf et al, 2008) and mood (Ruhe utilizes the same neural systems as inhibition, but the
et al, 2007). A corollary is that for subjects with very high critical process may instead be detection of behaviorally
5-HT BP , ATD might actually enhance the inhibition relevantcues(Hampshireetal,2010)orincreasedresponse
2A P
efficiency (LeMarquand et al, 1998). control demands (Dodds et al, 2011; Levy and Wagner,
Howmightsuchanon-linearresponsearise?Inprinciple, 2011). However, the IFG areas which were reported in
it requires opposing downstream mechanisms driven by previous studies include sites in the lateral prefrontal
differential rates of 5-HT neurotransmission, for example, convexity and frontal operculum extending onto anterior
from two receptor subtypes with different affinity for an insula.Inthisextendedregionweobserved anatomicaland
inhibitory autoreceptor. If the 5-HT BP reflects the pharmacological differences between AltGo and NoGo,
2A P
average serotonin level as determined by the raphe nucleus although both trial-types require detection of salient cues.
output (Erritzoe et al, 2010) then other receptor subtypes, Interestingly, there were no main effects of drug on
most notably the inhibitory 5-HT receptor, may con- performance, but 5-HT levels affected error rates
1A 2A
tribute to the interaction between ATD, IFG, and 5-HT differentially between NoGo and AltGo trials. Although
2A
BP (Hannon and Hoyer, 2008). there was no effect of 5-HT on the AltGo condition,
P 2A
Theinteractionsbetweentheacuteserotonergicstatus(as individuals with higher 5-HT levels showed increased
2A
induced by the drug interventions) and chronic serotoner- error rates during NoGo trials.
gic status (as indexed by neocortical 5-HT BP ) were AltGo cues prompted a switch of behavioral response. In
2A P
localizedtotheIFG.Tofacilitatetheinterpretationofthese prefrontal cortex, 5-HT is central to switching responses
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1003
withfeedback(Clarkeetal,2004)andresponsesetshifting. interventions. We also focus our interpretation on the
However, an important distinction is that the NoGo/AltGo contrast between the two interventions. The behavioral
paradigm did not require a shift between attentional set or differences were found to be minimal and there were no
task set between trials. Both AltGo and NoGo trials were differences in side effects (eg, nausea), and no differences
associated with activations in parietal, occipitotemporal, between drug conditions in terms of self-rated mood. It is
andsupplementarymotorarea(SMA).AltGowasassociated unlikely, therefore, that placebo effects can account for the
with greater activation in the motor network: motor, functionally specific and anatomically specific effects
premotor, parietal cortex, striatum, and cerebellum. This observed. The correlations with serotonergic markers also
is not simply due to the execution of movement in AltGo confirmtheserotonergicroleinNoGo-typeinhibition,over
trials, as this network was more active for AltGo than Go and above potential placebo effects. These factors together
trials. We suggest instead that late updating of a motor suggestthatobservedeffectsofATDandcitalopramareless
response is associated with either increased attention to likely to be due to differences in anxiety, discomfort, or
action (Rowe et al, 2002) or co-activation of both standard nausea, even where these might theoretically be 5-HT-
andalternativemotorprograms.Thislatterinterpretationis mediated. Although 5-HT receptor binding estimated
2A
compatible with the hypothesis that AltGo trials induced a from18F-altanserinPETisstableover2years(Marneretal,
‘race’ in the motor system between alternative responses 2009), the interval between PET and MRI in our subjects
(Logan et al, 1984). introduces additional variance. Moreover, despite clear
TheeffectsofcitalopramonIFGactivationwereminimal heritability of 5-HT receptor binding (Pinborg et al,
2A
in contrast with two previous studies (Del-Ben et al, 2005; 2008), environmental factors may contribute to adult
Langenecker et al, 2007) and the effects of the 5-HT variance, and we cannot infer that the interactions between
2A/2C
receptor agonist, mCPP (Anderson et al, 2002). The acuteandchronicserotonergicfactorsarenecessarilystate–
difference may be a type II error, despite a powerful trait interactions. In addition, it may also be that
repeated measures design, or differences in study popula- citalopram’s behavioral effects are not mediated by 5-
tions. However, differences in drug regime must also be HT receptors,andthereforedonotcorrelatewith5-HT
2A 2A
considered. For example, Del-Ben et al (2005) used a receptor binding. As such, citalopram and ATD cannot be
smallertotaldose(7.5mgduring7.5min)thanourinfusion considered as simple opposite interventions. Finally, one
(B50mg during the entire session B3h). The effects of must consider the potential confounds in pharmacological
high and low dose citalopram are not necessarily compar- fMRI studies. By seeking ‘trial-type by drug’ interactions
able.Forexample,inhumans,lowdosecitalopram5–10mg and regionally specific effects together with quantitative
increases prolactin levels from baseline and compared to arterial spin labeling perfusion studies to exclude non-
placebo(Attenburrowetal,2001;Del-Benetal,2005;McKie specific perfusion effects, we aimed to minimize such
et al, 2005). However, the effects of 20mg citalopram vary confounds.
across studies (Henning and Netter, 2002; Pinborg et al,
2004) perhaps due to high variability of absolute increases
andtimetopeak(Pinborgetal,2004).Prolactinlevelswere Conclusions
only measured twice in our study, preventing us from
The IFG was associated with response inhibition, and this
capturing smaller or transient increases. The complex
effect was not fully attributable to the engagement of
pharmacology of 5-HT and acute SSRIs may also explain
cognitive control and updating of a motor response
the null result for citalopram: although acute citalopram
following a low frequency stimulus. The activity of IFG
increases cortical extracellular 5-HT at high doses (Moret
that was associated with NoGo inhibition was itself
and Briley, 1996), it may not translate into enhanced
modulated by the interaction between acute and chronic
cortical post-synaptic stimulation with intermediate or low
serotonergic factors. Specifically, the neural response to
doses.Thisuncertaintyispartlyduetoacutestimulationof
ATDdependedonindividualdifferencesin5-HT receptor
somatodendritic 5-HT inhibitory autoreceptors in the 2A
1A binding, with implications for studies of clinical popula-
raphe nucleus.
tions with abnormal 5-HT levels.
Despite the lack of effect of citalopram (vs no-drug),
citalopram and ATD exerted significantly different effects
ontheinhibitionsignalexpressedinleftIFG,mirroringthe
ACKNOWLEDGEMENTS
IFGresponsereportedpreviously(Del-Benetal,2005).This
apparent laterality difference in thresholded activations The Lundbeck Foundation is gratefully acknowledged for
across studies should not be interpreted as evidence of their financial support of the study. Dr Fin S Larsen,
significant laterality effects in cognitive or neuronal Department of Hepatology, Rigshospitalet, is thanked for
functions (Henson, 2006). Indeed, lesion studies indicate providing measurements of serum tryptophan. SHS Inter-
bilateral IFG contributions to inhibition (Aron et al, 2003; national Ltd, Liverpool, UK is acknowledged for providing
Swick et al, 2008). the tryptophan-free amino acid mixture. JBR is supported
There are some limitations to our study. We did not use by the Wellcome Trust (088324).Financial support was
full placebo control of intravenous and oral challenges. provided primarily from the Lundbeck Foundation to the
Placebo effects, differential expectations, and non-specific CenterforIntegratedMolecularBrainImaging(Cimbi).The
drug induced changes in mood, nausea, or anxiety might John and Birthe Meyer Foundation donated funding for
contributetotheeffectsseen.Inordertominimizethenon- PET-scanner and cyclotron. The Copenhagen University
specific drug effects, we did not give subjects prior Hospitals Rigshospitalet and Hvidovre also supported the
information about differences in side effects of the study.TheSpiesfoundationdonatedfundingtothe3TTrio
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1004
MRI scanner. HRS was supported by a grant of excellence network using functional magnetic resonance imaging. Cereb
by the Lundbeck Foundation on the Control of Action Cortex 21: 1155–1165.
(ContAct, Grant no. R59 A5399). Eagle DM,Bari A, Robbins TW(2008). Theneuropsychopharma-
cology of action inhibition: cross-species translation of the
stop-signalandgo/no-go tasks.Psychopharmacology (Berl) 199:
439–456.
DISCLOSURE Erritzoe D, Holst K, Frokjaer VG, Licht CL, Kalbitzer J,
Nielsen FA et al (2010). A nonlinear relationship between
HRS has received honoraria as reviewing editor for Neuro-
cerebral serotonin transporter and 5-HT(2A) receptor binding:
image, as speaker for Biogen Idec Denmark A/S, and
an in vivo molecular imaging study in humans. J Neurosci 30:
scientificAdvisorfromLundbeckA/S,Valby,Denmark.The
3391–3397.
remaining authors declare no conflicts of interest. Evers EA, van der Veen FM, van Deursen JA, Schmitt JA, Deutz
NE, Jolles J (2006). The effect of acute tryptophan depletion on
the BOLD response during performance monitoring and
response inhibition in healthy male volunteers. Psychopharma-
REFERENCES
cology(Berl) 187: 200–208.
Adams KH, Hansen ES, Pinborg LH, Hasselbalch SG, Svarer C, Frokjaer VG, Mortensen EL, Nielsen FA, Haugbol S, Pinborg LH,
HolmSetal(2005).Patientswithobsessive-compulsivedisorder Adams KH et al (2008). Frontolimbic serotonin 2A
have increased 5-HT receptor binding in the caudate nuclei. receptor binding in healthy subjects is associated with person-
2A
Int JNeuropsychopharmacol 8: 391–401. ality risk factors for affective disorder. Biol Psychiatry 63:
Adams KH, Pinborg LH, Svarer C, Hasselbalch SG, Holm S, 569–576.
HaugbolSetal(2004).Adatabaseof[(18)F]-altanserinbinding Hampshire A, Chamberlain SR, Monti MM, Duncan J, Owen AM
to5-HT(2A)receptorsinnormalvolunteers:normativedataand (2010).Theroleoftherightinferiorfrontalgyrus:inhibitionand
relationshiptophysiologicalanddemographicvariables.Neuro- attentional control. Neuroimage 50: 1313–1319.
image21:1105–1113. Hannon J, Hoyer D (2008). Molecular biology of 5-HT receptors.
AndersonIM,ClarkL,ElliottR,KulkarniB,WilliamsSR,Deakin Behav BrainRes195:198–213.
JF (2002). 5-HT(2C) receptor activation by m-chloro- Harrison AA, Everitt BJ, Robbins TW (1999). Central serotonin
phenylpiperazine detected in humans with fMRI. Neuroreport depletion impairs both the acquisition and performance of a
13: 1547–1551. symmetrically reinforced go/no-go conditional visual discrimi-
Ardis TC, Cahir M, Elliott JJ, Bell R, Reynolds GP, Cooper SJ nation. Behav BrainRes100:99–112.
(2009). Effect of acute tryptophan depletion on noradrenaline HaugbolS,PinborgLH,RegeurL,HansenES,BolwigTG,Nielsen
anddopamine inthe rat brain. J Psychopharmacol 23: 51–55. FAetal(2007).Cerebral5-HT receptorbindingisincreasedin
2A
Aron AR, Fletcher PC, Bullmore ET, Sahakian BJ, Robbins TW patientswithTourette’ssyndrome. IntJNeuropsychopharmacol
(2003). Stop-signal inhibition disrupted by damage to right 10: 245–252.
inferiorfrontalgyrus inhumans. NatNeurosci 6:115–116. HenningJ,NetterP(2002).Oralapplicationofcitalopram(20mg)
Asahi S, Okamoto Y, Okada G, Yamawaki S, Yokota N (2004). and its usefulness for neuroendocrine challenge tests. Int J
Negative correlation between right prefrontal activity during Neuropsychopharmacol 5: 67–71.
response inhibition and impulsiveness: a fMRI study. Eur Arch Henson R (2006). Forward inference using functional neuroima-
Psychiatry Clin Neurosci 254: 245–251. ging: dissociations versus associations. Trends Cogn Sci 10:
Attenburrow MJ, Mitter PR, Whale R, Terao T, Cowen PJ (2001). 64–69.
Low-dose citalopram as a 5-HT neuroendocrine probe. Psycho- Huey ED, Putnam KT, Grafman J (2006). A systematic review of
pharmacology(Berl) 155: 323–326. neurotransmitter deficits and treatments in frontotemporal
Blokland A, Lieben C, Deutz NE (2002). Anxiogenic and dementia.Neurology 66: 17–22.
depressive-like effects, but no cognitive deficits, after repeated Iversen SD, Mishkin M (1970). Perseverative interference in
moderatetryptophandepletionintherat.JPsychopharmacol16: monkeys following selective lesions of the inferior prefrontal
39–49. convexity.Exp Brain Res11:376–386.
ChamberlainSR,HampshireA,MullerU,RubiaK,DelCampoN, Kenner NM, Mumford JA, Hommer RE, Skup M, Leibenluft E,
Craig K et al (2009). Atomoxetine modulates right inferior Poldrack RA (2010). Inhibitory motor control in response
frontal activation during inhibitory control: a pharmacological stopping andresponseswitching. J Neurosci 30: 8512–8518.
functionalmagneticresonanceimagingstudy.BiolPsychiatry65: Konishi S, Nakajima K, Uchida I, Kikyo H, Kameyama M,
550–555. Miyashita Y (1999). Common inhibitory mechanism in human
ChikazoeJ,JimuraK,AsariT,YamashitaK,MorimotoH,HiroseS inferior prefrontal cortex revealed by event-related functional
et al (2009). Functional dissociation in right inferior frontal MRI.Brain 122: 981–991.
cortex during performance of go/no-go task. Cereb Cortex 19: KonishiS,NakajimaK,UchidaI,SekiharaK,MiyashitaY(1998).
146–152. No-go dominant brain activity in human inferior prefrontal
Chikazoe J, Konishi S, Asari T, Jimura K, Miyashita Y (2007). cortex revealed by functional magnetic resonance imaging.
Activation of right inferior frontal gyrus during response EurJ Neurosci 10:1209–1213.
inhibition across response modalities. J Cogn Neurosci 19: Lamar M, Cutter WJ, Rubia K, Brammer M, Daly EM, Craig MC
69–80. et al (2009). 5-HT, prefrontal function and aging: fMRI of
ClarkeHF,DalleyJW,CroftsHS,RobbinsTW,RobertsAC(2004). inhibition and acute tryptophan depletion. Neurobiol Aging 30:
Cognitive inflexibility after prefrontal serotonin depletion. 1135–1146.
Science304:878–880. Lancaster J, Woldorff M, Parsons L, Liotti M, Freitas C, Rainey L
Del-BenCM,DeakinJF,McKieS,DelvaiNA,WilliamsSR,ElliottR et al (2000). Automated Talairach atlas labels for functional
et al (2005). The effect of citalopram pretreatment on neuronal brain mapping. HumBrain Mapp10:120–131.
responses to neuropsychological tasks innormal volunteers: an Langenecker SA, Kennedy SE, Guidotti LM, Briceno EM,
FMRI study. Neuropsychopharmacology 30:1724–1734. Own LS, Hooven T et al (2007). Frontal and limbic activation
Dodds CM, Morein-Zamir S, Robbins TW (2011). Dissociating during inhibitory control predicts treatment response in major
inhibition, attention, and response control in the frontoparietal depressive disorder. Biol Psychiatry 62: 1272–1280.
Neuropsychopharmacology
Serotoninmodulationandinhibition
JMacoveanuetal
1005
LeMarquand DG, Pihl RO, Young SN, Tremblay RE, Seguin JR, RoweJ,FristonK,FrackowiakR,PassinghamR(2002).Attention
Palmour RM et al (1998). Tryptophan depletion, executive to action: specific modulation of corticocortical interactions in
functions, and disinhibition in aggressive, adolescent males. humans.Neuroimage 17:988–998.
Neuropsychopharmacology 19: 333–341. Rubia K, Lee F, Cleare AJ, Tunstall N, Fu CH, Brammer M et al
Levy BJ, Wagner AD (2011). Cognitive control and right (2005). Tryptophan depletion reduces right inferior prefrontal
ventrolateral prefrontal cortex: reflexive reorienting, motor activationduringresponseinhibitioninfast,event-relatedfMRI.
inhibition, and action updating. Ann N Y Acad Sci 1224: Psychopharmacology (Berl) 179: 791–803.
40–62. Rubia K, Russell T, Overmeyer S, Brammer MJ, Bullmore ET,
LeytonM,OkazawaH,DiksicM,ParisJ,RosaP,MzengezaSetal Sharma T et al (2001). Mapping motor inhibition: conjunctive
(2001).Brainregionalalpha-[11C]methyl-L-tryptophantrapping brainactivationsacrossdifferentversionsofgo/no-goandstop
inimpulsivesubjectswithborderlinepersonalitydisorder.AmJ tasks. Neuroimage 13: 250–261.
Psychiatry 158:775–782. RuheHG,MasonNS,ScheneAH(2007).Moodisindirectlyrelated
LoganGD,CowanWB,DavisKA(1984). Onthe abilitytoinhibit toserotonin,norepinephrineanddopaminelevelsinhumans:a
simple and choice reaction time responses: a model and a meta-analysis of monoamine depletion studies. Mol Psychiatry
method. J ExpPsychol HumPerceptPerform 10: 276–291. 12: 331–359.
Lund TE,NorgaardMD,Rostrup E, RoweJB,Paulson OB (2005). SimmondsDJ,PekarJJ,MostofskySH(2008).Meta-analysisofGo/
Motionoractivity:theirroleinintra-andinter-subjectvariation No-gotasksdemonstratingthatfMRIactivationassociatedwith
infMRI. Neuroimage 26: 960–964. response inhibition is task-dependent. Neuropsychologia 46:
MarnerL,KnudsenGM,HaugbolS,HolmS,BaareW,Hasselbalch 224–232.
SG(2009).Longitudinalassessmentofcerebral5-HT receptors Svarer C, Madsen K, Hasselbalch SG, Pinborg LH, Haugbol S,
2A
inhealthyelderlyvolunteers:an(18F)-altanserinPETstudy.Eur Frokjaer VG et al (2005). MR-based automatic delineation of
J NuclMedMolImaging 36: 287–293. volumes of interest in human brain PET images using
Masaki D, Yokoyama C, Kinoshita S, Tsuchida H, Nakatomi Y, probabilitymaps. Neuroimage 24: 969–979.
Yoshimoto K et al (2006). Relationship between limbic and SwannN,TandonN,CanoltyR,EllmoreTM,McEvoyLK,DreyerS
cortical 5-HT neurotransmission and acquisition and reversal et al (2009). Intracranial EEG reveals a time- and frequency-
learning in a go/no-go task in rats. Psychopharmacology (Berl) specific role for the right inferior frontal gyrus and primary
189: 249–258. motor cortex in stopping initiated responses. J Neurosci 29:
McKie S, Del-Ben C, Elliott R, Williams S, del Vai N, Anderson I 12675–12685.
et al (2005). Neuronal effects of acute citalopram detected by SwickD,AshleyV,TurkenAU(2008).Leftinferiorfrontalgyrusis
pharmacoMRI. Psychopharmacology (Berl)180: 680–686. critical for responseinhibition. BMCNeurosci 9: 102.
McNairD,LorrM,DropplemanL(1971).ManualfortheProfileof vanDonkelaarEL,BloklandA,FerringtonL,KellyPA,Steinbusch
Mood States. Educational and Industrial Testing Services: San HW, Prickaerts J (2011). Mechanism of acute tryptophan
Diego, CA. depletion: is itonly serotonin?Mol Psychiatry 16: 695–713.
Moret C, Briley M (1996). Effects of acute and repeated WagerTD,SylvesterCY,LaceySC,NeeDE,FranklinM,JonidesJ
administrationofcitalopramonextracellularlevelsofserotonin (2005).Commonanduniquecomponentsofresponseinhibition
inrat brain. EurJ Pharmacol295:189–197. revealedby fMRI.Neuroimage 27: 323–340.
Mostofsky SH, Simmonds DJ (2008). Response inhibition and Walderhaug E, Lunde H, Nordvik JE, Landro NI, Refsum H,
response selection: two sides of the same coin. J Cogn Neurosci MagnussonA(2002).Loweringofserotoninbyrapidtryptophan
20: 751–761. depletion increases impulsiveness in normal individuals. Psy-
Nomura M, Nomura Y (2006). Psychological, neuroimaging, and chopharmacology (Berl)164:385–391.
biochemicalstudiesonfunctionalassociationbetweenimpulsive Williams WA, Shoaf SE, Hommer D, Rawlings R, Linnoila M
behavior and the 5-HT receptor gene polymorphism in (1999). Effects of acute tryptophan depletion on plasma and
2A
humans.AnnN YAcad Sci1086: 134–143. cerebrospinal fluid tryptophan and 5-hydroxyindoleacetic acid
Pinborg LH, Adams KH, Svarer C, Holm S, Hasselbalch SG, innormal volunteers. JNeurochem 72:1641–1647.
HaugbolSetal(2003).Quantificationof5-HT receptorsinthe WingroveJ,BondAJ,CleareAJ,SherwoodR(1999).Traithostility
2A
human brain using [18F]altanserin-PET and the bolus/infusion and prolactin response to tryptophan enhancement/depletion.
approach. J CerebBlood Flow Metab23: 985–996. Neuropsychobiology 40:202–206.
Pinborg LH, Adams KH, Yndgaard S, Hasselbalch SG, Holm S, ZepfFD,HoltmannM,StadlerC,DemischL,SchmittM,WockelL
Kristiansen H et al (2004). (18F)altanserin binding to human et al (2008). Diminished serotonergic functioning in hostile
5HT2A receptors is unaltered after citalopram and pindolol childrenwithADHD:tryptophandepletionincreasesbehaviour-
challenge. JCereb Blood FlowMetab24: 1037–1045. alinhibition. Pharmacopsychiatry 41:60–65.
PinborgLH,ArfanH,HaugbolS,KyvikKO,HjelmborgJV,Svarer ZhengD,OkaT,BokuraH,YamaguchiS(2008).Thekeylocusof
C et al (2008). The 5-HT receptor binding pattern in the commonresponseinhibitionnetworkforno-goandstopsignals.
2A
humanbrainisstronglygeneticallydetermined.Neuroimage40: J CognNeurosci 20:1434–1442.
1175–1180.
Ridderinkhof KR, van den Wildenberg WP, Segalowitz SJ, Carter
ThisworkislicensedunderaCreativeCommons
CS(2004).Neurocognitivemechanismsofcognitivecontrol:the
roleofprefrontalcortexinactionselection,responseinhibition, Attribution-NonCommercial-ShareAlike 3.0 Un-
performance monitoring, and reward-based learning. Brain ported License. To view a copy of this license, visit http://
Cogn56:129–140. creativecommons.org/licenses/by-nc-sa/3.0/
Neuropsychopharmacology