Influence of the COMT Genotype on Working Memory and Brain Activity Changes During Development

Abstract

Background: The Valine158Methionine (Val158Met) polymorphism of the COMT gene leads to lower enzymatic activity and higher dopamine availability in Met carriers. The Met allele is associated with better performance and reduced prefrontal cortex activation during working memory (WM) tasks in adults. Dopaminergic system changes during adolescence may lead to a reduction of basal dopamine levels, potentially affecting Met allele benefits during development. Methods: We investigated the association of COMT genotype with behavioral (n 322) and magnetic resonance imaging data (n 81– 84) collected during performance of a visuospatial WM task and potential changes in these effects during development (reflected in age genotype interactions). Data were collected from a cross-sectional and longitudinal typically developing sample of 6-to 20-year-olds. Results: Visuospatial WM capacity exhibited an age genotype interaction, with a benefit of the Met allele emerging after 10 years of age. There was a parallel age genotype interaction on WM-related activation in the right inferior frontal gyrus and intraparietal sulcus (IPS), with increases in activation with age in the Val/Val group only. Main effects of COMT genotype were also observed in the IPS, with greater gray matter volumes bilaterally and greater right IPS activation in the Val/Val group compared with the Met carriers.

Full text

Page 1

ARCHIVAL REPORT
Influence of the COMT Genotype on Working Memory
and Brain Activity Changes During Development
Iroise Dumontheil, Chantal Roggeman, Tim Ziermans, Myriam Peyrard-Janvid, Hans Matsson,
Juha Kere, and Torkel Klingberg
Background: The Valine158Methionine (Val158Met) polymorphism of the COMT gene leads to lower enzymatic activity and higher
dopamine availability in Met carriers. The Met allele is associated with better performance and reduced prefrontal cortex activation during
workingmemory(WM)tasksinadults.Dopaminergicsystemchangesduringadolescencemayleadtoareductionofbasaldopaminelevels,
potentially affecting Met allele benefits during development.
Methods: WeinvestigatedtheassociationofCOMTgenotypewithbehavioral(n?322)andmagneticresonanceimagingdata(n?81–84)
collected during performance of a visuospatial WM task and potential changes in these effects during development (reflected in age ?
genotype interactions). Data were collected from a cross-sectional and longitudinal typically developing sample of 6- to 20-year-olds.
Results: VisuospatialWMcapacityexhibitedanage?genotypeinteraction,withabenefitoftheMetalleleemergingafter10yearsofage.
Therewasaparallelage?genotypeinteractiononWM-relatedactivationintherightinferiorfrontalgyrusandintraparietalsulcus(IPS),with
increases in activation with age in the Val/Val group only. Main effects of COMT genotype were also observed in the IPS, with greater gray
matter volumes bilaterally and greater right IPS activation in the Val/Val group compared with the Met carriers.
Conclusions: TheseresultssuggestthatCOMTgenotypeeffectsonWMbrainactivityandbehaviorarenotstaticduringdevelopment.The
fulldevelopmentalpictureshouldbeconsideredwhentryingtounderstandtheimpactofgeneticpolymorphismsonthematurecognition
of healthy adult or psychiatric populations.
Key Words: Child development, functional magnetic resonance
imaging (fMRI), genetics, gray matter, voxel-based morphometry
(VBM), working memory
T
as a promising candidate gene for studying variance in cognitive
function and mental illness (1,2). Developmental studies of this
genecouldinformusonthetypicaldevelopmentofthedopamine
system and provide a reference for the study of developmental
disorders that have been linked to dopamine system dysfunction,
suchasschizophrenia(3),forwhichCOMTfunctionmayberelevant
(2,4,5), although the evidence is inconsistent (e.g., [1]).
The common rs4680 Valine158Methionine (Val158Met) single
nucleotide polymorphism (SNP) leads to a reduction of COMT en-
zymaticactivityinMetcarriers(6–8).Theeffectofrs4680oncogni-
tion has been extensively studied in adults, both in typical and
neuropsychiatric populations. Better working memory (WM) per-
formance (e.g., in N-back tasks), fewer perseverative errors, and
higher IQ tend to be associated with the Met allele in healthy
individuals (9–11), although some studies have failed to replicate
theseeffects(10).Infunctionalmagneticresonanceimaging(fMRI)
studies, Val/Val adults tend to exhibit greater prefrontal cortex
hegenecodingforthecatechol-O-methyltransferase(COMT)
enzyme,whichmediatesthedegradationofcatecholamines,
in particular dopamine, has attracted considerable attention
(PFC) activation during WM tasks than Met/Met individuals (12,13).
Structural imaging studies report less consistent results (14–16).
In children and adolescents, effects of COMT genotype in spe-
cific age groups have been reported on brain structure (17–19),
resting brain perfusion (20), and brain activation in response to
emotional stimuli (17). Behavioral findings are mixed (21–23). This
lack of consistency may arise from age differences in the effect of
COMT genotype on cognition.
Dopamine effects on behavior follow an inverted U-shaped
dose-responsecurve,withbothdeficientandexcessiveamountsof
dopamine activity predicting poor cognitive task performance
(24,25). The Met/Met adults are thought to be near the apex of this
curve, while Val carriers lay toward the lower end because of the
increaseddopaminemetabolismrateassociatedwiththeValallele
(2,26,27).
There is some evidence that the dopamine system undergoes
changesduringchildhoodandadolescence.Intherhesusmonkey,
dopamineconcentrationinthePFCanddopaminesynthesisinthe
PFC and parietal cortex peak around puberty (2–3 years), while
dopamineconcentrationintheparietalcortexremainsstablefrom
5 months of age (28,29). Dopaminergic input in the PFC also peaks
aroundpuberty,bothintermsoflengthoftheaxonsanddensityof
varicosities (30–32). In humans, postmortem studies of the PFC
have shown that dopamine concentration is highest during early
postnatal development (33), while D1 receptor density peaks in
adolescents (age 14–18) and young adults compared with neo-
nates, infants, adults, and aged adults (33). In the living brain, a
decrease in D1 receptor binding was observed in the PFC and
parietalcortexoveranagespanof10to30years(34).Theseresults
overall suggest a decrease in dopamine levels from puberty to
adulthood (23). Little information is available regarding potential
sex differences.
A decrease in basal dopamine level between childhood and
adulthoodcouldaffectthepositionoftheVal158Metgenotypeson
the inverted U-shaped curve and lead to a differential effect of
rs4680duringdevelopment.Wetestedthehypothesisofachange
From the Neuroscience Department (ID, CR, TZ, TK), Karolinska Institutet,
Stockholm; Department of Biosciences and Nutrition (MP-J, HM, JK),
Karolinska Institutet, Huddinge; and Science for Life Laboratory (JK),
Department of Biosciences and Nutrition at Novum, Karolinska Institu-
tet, Stockholm, Sweden; and Department of Medical Genetics (JK),
Haartman Institute, University of Helsinki, and Folkhälsan Institute of
Genetics, Helsinki, Finland.
Address correspondence to Iroise Dumontheil, Ph.D., Karolinksa Institutet,
NeuroscienceDepartment,Retziusväg8,SE-17177Stockholm,Sweden;
E-mail: i.dumontheil@gmail.com.
Received Nov 19, 2010; revised Feb 24, 2011; accepted Feb 24, 2011.
BIOL PSYCHIATRY 2011;xx:xxx
© 2011 Society of Biological Psychiatry
0006-3223/$36.00
doi:10.1016/j.biopsych.2011.02.027

Page 2

in the effect of COMT genotype with age on behavioral and neuro-
imaging data associated with visuospatial WM, a cognitive ability
that develops during childhood and adolescence (35).
Methods and Materials
Participants and Genetic Data
Participants in nine different age groups (6, 8, 10, 12, 14, 16, 18,
20, and 25 years) were recruited using random sampling from the
population registry in Nynäshamn in Sweden (Brainchild study
[36]). Informed consent was obtained from the participants and
from the parents of children under 18. The study was approved by
the local ethics committee of the Karolinska University Hospital,
Stockholm. See Figure 1 and Table 1 for a description of included
participants. Deoxyribonucleic acid was extracted either from
blood or saliva. The COMT genotype is located on chromosome
22q11.2andSNPrs4680isinexon4ofthegene(seeSupplement1
for details of the genetic analyses and excluded participants).
Assessment of Working Memory
Participants completed a large neuropsychological battery ad-
ministered individually and in a quiet room. Visuospatial WM was
assessed using a grid task (Dot Matrix) from the Automated Work-
ing Memory Assessment battery (37). This computerized task in-
volves remembering the location and order of dots displayed se-
quentially in a four-by-four grid, for 1000 milliseconds each, with a
500-millisecondintervalbetweendots.Aftertrainingwithone,two,
andthreedots,theteststartedwithonedot.Eachlevelconsistedof
sixtrials.Fourcorrectanswerswererequiredformovingtothenext
level, where one more dot needed to be remembered. The test
terminated when three errors were committed on one level. The
score used was the total number of correct trials.
Dot Matrix Statistical Analyses
Linear mixed model analyses, which allow the inclusion of data
from participants who have not attended all testing waves and
adjust for intercorrelation between testing waves (38-40), were
performed using the PASW 18.0 statistical package (41). A com-
pound symmetry covariance structure was used and Dot Matrix
score was treated as a repeated measure (rounds 1 and 2). We first
identified how to best model changes in WM capacity with age,
includingthelevelingofperformanceduringmid-adolescence.We
comparedenteringtheagevariableasagewithouttransformation
(lineareffect),theinverseofage(age?1),andthenaturallogarithm
of age (ln[age]) (42,43). In all cases, age was transformed into Z-
scorestoevaluateeffectsofinterestatthemeanageofoursample
andreducecollinearitybetweenmaineffectsandinteractions(44).
Sex and sex ? age were entered as covariates. The best model was
identifiedusingAkaike’sinformationcriterionandSchwartz’sinfor-
mation criterion (45,46).
We then tested whether including COMT genotype and age ?
COMT genotype as additional fixed effects improved the model.
Genotype ? sex and genotype ? age ? sex interactions were also
included to test for potential sex differences in the effect of COMT
genotype (47). Although the effect of rs4680 on dopamine degra-
dation is additive, the effect on cognition is suggested to follow a
nonlinear, inverted-U curve. To allow the detection of such effects,
weevaluatedbothadditive(0?Met/Met,1?Val/Met,2?Val/Val)
and dominance models (Val dominance: 0 ? Met/Met, 1 ? Val
carriers; Met dominance: 0 ? Met carriers, 1 ? Val/Val) (similar to
[9,48]).
Brain Imaging: fMRI
Magnetic resonance imaging data were collected on a 1.5 T
Siemens scanner (Siemens, Erlangen, Germany) (Supplement 1).
Participants performed two 5-minute sessions, each including 16
WM and 16 control trials in a pseudorandomized order. Dots were
presented sequentially in a four-by-four grid. To reduce potential
age or genotype differences in behavior, the task included load 2
(twodots)and4(fourdots)trialsonly.Afterthelastdotwasshown,
a number was presented in the grid. Participants were asked to
indicate whether the number and its position in the grid matched.
For example, “2?” would prompt the participant to indicate
Figure 1. Distribution of the data points across round 1
and round 2 according to age and COMT genotype in
eachofthethreeanalyses.Foreachagegroup,thenum-
berofparticipantsfromround1andround2isindicated
above each bar (round 1 n / round 2 n). fMRI, functional
magnetic resonance imaging; Met, Methionine; Val, Va-
line; VBM, voxel-based morphometry.
Table 1. Distribution of the Participants and Longitudinal Measures According to COMT Genotype in Each of the Three Analyses
Behavioral AnalysisfMRI AnalysisVBM Analysis
Participants Total n M/FParticipantsTotal nM/F ParticipantsTotal nM/F
Met/Met
Met/Val
Val/Val
Total
98 (79)
147 (115)
77 (66)
322 (260)
177
262
143
582
87/90
136/126
70/73
293/289
22 (9)
38 (19)
21 (16)
81 (44)
31
57
37
125
16/15
32/25
16/21
64/61
22 (12)
39 (21)
23 (14)
84 (47)
34
60
37
131
18/16
33/27
17/20
68/63
The numbers in brackets indicate the number of subjects with repeated measures.
F, female; fMRI, functional magnetic resonance imaging; M, male; Met, Methionine; Val, Valine; VBM, voxel-based morphometry.
2 BIOL PSYCHIATRY 2011;xx:xxx
I. Dumontheil et al.
www.sobp.org/journal

Page 3

whether the second circle had appeared in the grid position filled
bythenumber.Inthecontrolcondition,thedotswerepresentedin
the corners of the grid and the cue (“8?”) always required a “no”
response.
Preprocessing and statistical analysis (36) were carried out with
SPM5 (http://www.fil.ion.ucl.ac.uk/spm/software/spm5). Contrast
imagescomparingWMandcontrolconditions,irrespectiveofload,
for each participant were used in flexible factorial design second-
level analyses that modeled whether the contrasts were from the
same or different participants by including subject and testing
round as factors. This permitted the inclusion in a single analysis of
participantswithandwithoutlongitudinaldata.Ageinmonthswas
transformed into Z-score (ageZ).
First, ageZand sex were entered as covariates and the main
WM-control contrast was performed, correcting for multiple com-
parisonacrossallvoxelswithinthebrain,withafalsediscoveryrate
(FDR) threshold of p ? .05, to identify regions recruited during the
task. Then, genotype and genotype ? ageZwere entered as addi-
tionalcovariates.Genotypewasenteredwithanadditive,Metdom-
inant or Val dominant effect. In each model, a single F-test was
performed to test whether any main effect of genotype or age ?
genotype interaction could be observed (FDR, p ? .05). Contrasts
were inclusively masked by the WM-control main effect previously
defined.Resultswereplottedonasurface-basedhumanatlas(Pop-
ulation-Average, Landmark- and Surface-Based atlas) (49,50) using
the Caret software (http://www.nitrc.org/projects/caret/) (51).
Mean parameter estimates from the first level contrasts were ob-
tained in the significant clusters using MarsBar (52) and further
analyzed using linear mixed models with PASW 18.0.
Brain Imaging: Voxel-Based Morphometry
Structural T1-weighted spin echo images were acquired with a
three-dimensionalmagnetizationpreparedrapidgradientechose-
quence(fieldofview?256?256mm,256?256grid,1mm3voxel
size).Voxel-basedmorphometry(VBM)wasperformedusingSPM5.
Following segmentation of the T1-weighted images, high-dimen-
sional normalization was performed using the Diffeomorphic Ana-
tomicalRegistrationThroughExponentiatedLiealgebratoolboxin
statistical parametric mapping (53). The modulated warped gray
matter (GM) images were then smoothed with an 8 mm Gaussian
kernel.Thesameflexiblefactorialanalysesandstatisticalthresholds
asthoseusedforthefMRIdatawereusedfortheVBMGMdata,with
the addition of total GM volume as a covariate.
Results
Analyses combined data from all participants and both testing
rounds using linear mixed models. The complicated nonindepen-
dent design did not allow the calculation of standardized effect
statistics (54). Effect sizes from analyses performed on the round 1
data only are reported in Supplement 1.
Behavioral Analysis
WM Capacity Development. Behavioral data were available
for 322 participants, including 260 with longitudinal data. We first
compared the information criteria of three different models (with
the age variable as ageZ, age?1
model of WM capacity development. Linear mixed model analyses
included Dot Matrix score as the dependent variable, and sex, age,
and their interaction were entered as fixed effects. The best model
was where WM capacity was explained by age?1
changeinWMcapacityoveryoungeragesandthenflatteningtoan
asymptote.
Z, or ln[age]Z) to identify the best
Z, with a steeper
Influence of COMT Genotype on WM Capacity Devel-
opment. We then tested whether including the main effect of
COMTgenotypeandthegenotype?ageinteractionimprovedthis
model. Genotype ? sex and genotype ? age ? sex interactions
werealsoincluded.Genotypewasmodeledwithanadditiveeffect,
Val dominance, or Met dominance.
OnlytheValdominancemodelledtoasignificantlyimprovedfit
of the data compared with the age and sex only model (likelihood-
ratio test, D ? 9.77, 4 degrees of freedom, p ? .045). The only
significant effects observed in this full model were a main effect of
age?1
Z[F(1,502.7) ? 171.01, p ? .001] and an interaction between
genotype (Met/Met vs. Val carriers) and age?1
p ? .036]. The Met/Met individuals switched from being poorer
performers during childhood to being the better performers from
mid-adolescence onward (Figure 2). According to the model esti-
mates,thecross-overoccurredat10.2yearsofage.Neithersexnor
interactions between sex and other variables had a significant ef-
fect on behavior.
Z[F(1,500.1) ? 4.40,
fMRI Analysis
In-ScannerPerformance. Functionalmagneticresonanceim-
aging data were available for 81 participants, including 44 with
longitudinaldata.Accuracywashighoverall.Participantswereless
accurateintheWMcondition(WM:mean90.6%?SD10.6;control
condition:98.4%?3.5)[t(124)?9.52,p?.001),aswellasslowerto
respond (correct trials only, WM: 1277 msec ? 360; control condi-
tion: 738 msec ? 215) [t(124) ? 25.11, p ? .001]. The difference
betweenWMandcontrolconditionsdecreasedwithageintermsof
both accuracy and reaction time [main effect of age?1
tively, F(1,87.8) ? 10.06, p ? .002, and F(1,97.1) ? 32.05, p ? .001];
however,therewasnomaineffectofgenotypeoragebygenotype
interaction (all p’s ? .29). Genotype thus did not affect WM versus
control condition performance. Furthermore, only correct trials
were used in the brain activity analyses.
WM-Control Main Effect. Functional magnetic resonance im-
aging data were analyzed using a flexible factorial second-level
Z, respec-
Figure2.Workingmemorycapacityasafunctionofeachageyeargroupat
testing and COMT Valine158Methionine polymorphism. The bars are in-
cluded for illustrative purposes only and represent the mean and standard
error of working memory capacity, as measured by the Dot Matrix score, in
each age group, collapsing across longitudinal and cross-sectional data.
Participantstestedatbothroundsareincludedastwodatapoints.Analyses
were performed with age as a covariate rather than considering the age
groupsseparately.ThelinesrepresentthemeanpredictedDotMatrixscore
foreachgenotypegroup,asobtainedfromthemixedmodelanalysisusing
the inverse of age as a factor and genotype effects with Valine dominance.
Met, Methionine; SE, standard error; Val, Valine.
I. Dumontheil et al.
BIOL PSYCHIATRY 2011;xx:xxx 3
www.sobp.org/journal

Page 4

analysis of the WM-control contrast, which coded subject and test-
ing round as factors and ageZand sex as covariates. Because of the
smallersamplesizeofthefMRIdata,onlythelineareffectofagewas
tested in statistical parametric mapping analyses. The resulting
WM-control contrast image (FDR, p ? .05) was saved and used as a
maskforthesubsequentanalyses.Alargefrontoparietalnetworkof
regions was more active in the WM than the control condition
(Figure 3A).
EffectofCOMTGenotypeontheWM-ControlContrast. The
COMT genotype and genotype ? ageZwere added as additional
covariatesintheflexiblefactorialanalysis.Genotypewasenteredas
an additive effect, or with Val or Met dominance. Interactions be-
tweengenotypeandsexandgenotype,sex,andagewereincluded
in the follow-up analyses only. In each analysis, a single F-test as-
sessed whether genotype or genotype ? ageZhad significant ef-
fects (FDR, p ? .05, within WM-control mask). Only the Met domi-
nance model resulted in significant clusters located in the right
inferior frontal gyrus (IFG) and in the posterior section of the right
intraparietal sulcus (IPS)/angular gyrus (Figure 3B and Table 2; Fig-
ure S1 in Supplement 1).
Mean parameter estimates from the individual first-level WM-
control contrasts were obtained for these two clusters. Linear
mixedlevelmodelanalyses(seeTable2forstatistics)demonstrated
that the parietal cluster exhibited both a main effect of genotype,
withgreateractivationinVal/ValparticipantsthanMetcarriers,and
an age by genotype interaction, with an increase in activation with
age in the Val/Val participants only [Val/Val, F(1,22.5) ? 21.5, p ?
.001; Met carriers, p ? .2]. The frontal cluster exhibited an age by
genotype interaction with a similar pattern [Val/Val, F(1,20.3) ?
25.9, p ? .001; Met carriers, p ? .4] (Figure 3B). Note that for both
Figure 3. Functional magnetic resonance imaging (fMRI) and voxel-based morphometry results. (A) Render of the fMRI main effect working memory
(WM)-control (false discovery rate [FDR], p ? .05) on a surface-based human atlas (see Methods and Materials). From left to right: lateral view of the left and
right hemispheres, dorsal view of the left and right hemispheres. This contrast was used as an inclusive mask for the tests of the effect of COMT genotype
presented in (B) and (C). (B) F-test of COMT genotype or genotype ? age effects on the fMRI WM-control contrast (FDR, p ? .05). Mean parameter estimates
oftheWM-controlcontrastatthefirstlevelwerecalculatedandplottedfortheparietalandfrontalclustersobservedintherighthemisphere.Fulllinesconnect
thoseparticipantswhoparticipatedinbothtestingrounds.Dashedlinesarethepredictedfitfromlinearmixedmodelsincludingsex,ageZ,COMTgenotype
(withMethioninedominance),andgenotype?ageZinteractionascovariates.Theagebygenotypeinteractionwassignificantinbothclusters.Therewasan
additional main effect of genotype in the parietal cluster. (C) F-test of COMT genotype or genotype ? age effects on the voxel-based morphometry gray
matter data (FDR, p ? .05). Mean gray matter volume was calculated and plotted for the two largest parietal clusters (Table 2). Full lines connect those
participantswhoparticipatedinbothtestingrounds.Dashedlinesarethepredictedfitfromlinearmixedmodelsincludingsex,ageZ,totalgraymattervolume
(Z),COMTgenotype(withMethioninedominance),andgenotype?ageZinteractionascovariates.Themaineffectofgenotypebutnottheage?genotype
interaction was significant in both clusters (Figure S1 in Supplement 1). COMT, catechol-O-methyltransferase; GM, gray matter; Met, Methionine; Val, Valine.
4 BIOL PSYCHIATRY 2011;xx:xxx
I. Dumontheil et al.
www.sobp.org/journal

Page 5

clusters the genotype effects were significant when using age?1
but the information criteria indicated that the models including
ageZwereabetterfitofthedata.Figure4relatesourfindingstothe
suggestion that PFC functioning and WM performance follow an
inverted-U curve function of dopamine levels.
The main effect of genotype and age ? genotype interaction
remained significant (p’s ? .001) when including genotype ? sex,
age?sex,andgenotype?age?sexinteractionsinthemodel.In
the IPS, the main effect of genotype tended to be greater in male
subjects, while in the IFG the interaction between age and geno-
type tended to be greater in male subjects (Table 2).
Further analyses (Figures S2 and S3 in Supplement 1) indicated
that activations in the IFG and IPS clusters were more strongly
correlated in the Val/Val group and that these participants also
showed a positive correlation between IFG activation and Dot Ma-
trix performance.
Z,
Voxel-Based Morphometry
Structural magnetic resonance imaging data were available for
84 participants, including 47 with longitudinal data. The effects of
rs4680 on GM volumes were analyzed in the same way as for the
fMRI data. Total GM volumeZwas included as an additional covari-
ate to permit the study of regional rather than global effects. How-
ever, similar results were obtained when total GM volume was not
included.
Again, only the Met dominance model resulted in significant
clusters, located in the lateral bank of the IPS bilaterally (Figure 3C
and Table 2; Figure S1 in Supplement 1). Individual mean GM vol-
umes were calculated for each cluster. Linear mixed level model
analyses indicated that all three clusters showed a main effect of
genotype, with greater GM volumes in Val/Val participants than
Met carriers. There was a weak trend in the largest cluster for a
genotype by age interaction, with a steeper decrease in GM with
age in the Val/Val individuals than Met carriers. The main effects of
genotypeinthethreeclustersremainedsignificant(allp’s?.0001)
whenincludinggenotype?sex,age?sex,andgenotype?age?
sex interactions in the model. Significant interactions between sex
andgenotypewereobservedbilaterally(Table2).Inbothcases,the
effect of genotype was greater in males than females.
Discussion
This longitudinal and cross-sectional study of 6- to 20-year-olds
combined a large sample and full age range of participants to
investigate developmental changes in the effect of a functional
polymorphismoftheCOMTgeneoncognition,brainstructure,and
brain function. Previous studies of children and adolescents ob-
served inconsistent COMT genotype effects and were restricted by
either a small age range (21) or a small sample size (22,23). Our
results show that the adult pattern of the effects of the COMT
Val158Metpolymorphismonbrainandbehaviorduringtheperfor-
manceofavisuospatialWMtaskemergesduringadolescence.The
Met/Met individuals showed a steeper increase in performance
with age than Val carriers, while Val/Val individuals exhibited an
increase in activation in the right IFG and right IPS/angular gyrus
withage.Additionalmaineffectsofgenotypewereobservedinthe
parietal cortex for both fMRI and gray matter data, with larger
activationandGMvolumesinVal/ValparticipantsthanMetcarriers.
Differencesingenotypeeffectsonthebehavioralandneuroim-
agingdata,suchastheValversusMetdominanceeffects,mayhave
beendrivenbyarangeoffactors.Forexample,thefMRIvisuospatial
WM task included only relatively low WM loads and a control task,
while the behavioral score reflected the performance achieved up
to the participants’ maximum WM load level. In addition, behavior
may reflect more complex and global dopamine effects on brain
activity and structure in a wide network of brain regions.
BehavioralresultsoftheDotMatrixtaskindicatedthatMet/Met
homozygotes switched from being the underperformers to follow
theadultpatternofbetterWMperformanceataroundtheageof10
to 12 years old, typically considered as the start of puberty. Barnett
etal.(21)reportedagreatereffectofCOMTgenotypeonverbalIQin
8-year-oldboyswhowerepubertalatage9comparedwithprepu-
bertal boys or girls. This pattern was not observed on other mea-
sures,includingacountingspanmeasureofWM.Thepresentstudy
Table 2. COMT Genetic Effects on the fMRI and VBM Data
SPM
Linear Mixed Model Analyses
Labelx, y, z (MNI)nZ
F-test
pFDR
rs4680a
rs4680 ? ageZ
a
rs4680 ? sexb
rs4680 ? ageZ? sexb
fMRI
Posterior Section
IPS/Angular Gyrus
Inferior Frontal Gyrus
VBM: Gray Matter Volumes
Lateral bank IPS
Lateral bank IPS
Lateral bank IPS
40 ?74 3677 5.05 .007F(1,69.2) ? 18.15c
F(1,89.8) ? 11.96d
F(1,67.3) ? 3.55e
ns
54 8 24 174.19 .023 nsF(1,83.0) ? 22.84f
nsF(1,79.3) ? 3.10e
?42 ?36 38
47 ?33 45
42 ?43 35
276
64
4.67
4.30
3.93
.023
.023
.028
F(1,82.2) ? 18.94c
F(1,81.2) ? 18.64c
F(1,82.6) ? 15.66d
F(1,64.9) ? 2.77e
ns
ns
F(1,80.4) ? 4.39h
ns
F(1,80.6) ? 7.44g
ns
ns
ns6
ClusterswereidentifiedfromF-testsofeffectsofCOMTgenotypeorageZ?COMTgenotype.ContrastswereperformedatanFDR-correctedthresholdof
p ? .05 and inclusively masked by the fMRI main effect of WM-control (FDR, p ? .05). The right section of the table reports the statistical results obtained in
follow-up linear mixed model analyses.
FDR,falsediscoveryrate;fMRI,functionalmagneticresonanceimaging;GM,graymatter;IPS,intraparietalsulcus;MNI,MontrealNeurologicalInstitute;n,cluster
size;ns,nonsignificant;pFDR,FDRcorrectedpvalue;SPM,statisticalparametricmapping;VBM,voxel-basedmorphometry;WM,workingmemory;Z,Z-score.
aFixed effects comprised rs4680, ageZ, sex, and rs4680 ? ageZ(and total GM volumeZfor VBM analyses).
bFixed effects comprised, in addition toa, rs4680 ? sex, ageZ? sex, and rs4680 ? ageZ? sex.
cp ? .0001.
dp ? .001.
ep ? .10.
fp ? .00001.
gp ? .01.
hp ? .05.
I. Dumontheil et al.
BIOL PSYCHIATRY 2011;xx:xxx 5
www.sobp.org/journal

Page 6

did not evaluate pubertal state. However, WM capacity showed no
significanteffectofsexorinteractionbetweensexandageorgene.
WethereforeinterprettheobservedinteractionofCOMTgenotype
by age as reflecting a changing dopamine system during develop-
ment, independent of possible additional interactions with hor-
monaleffects.Note,however,thatsubstanceusewasnotrecorded
inthisstudyandisapotentialconfoundoftheageeffectobserved
here(55).InSweden,dailyoralmostdailysmokingincreasesduring
adolescencefrom?1%atage11to12yearsto12%atage18to20
years(56,57),andtheCOMTValallelehasbeenshowntobeasmall
risk factor (odds ratio 1.12) for smoking (58). Smoking is associated
with lower WM in adults, while the evidence is mixed in younger
samples(59).TheincreasinglypoorerWMperformanceinValcarri-
ers compared with Met/Met individuals may thus partly reflect in-
creased smoking during adolescence, particularly in Val carriers.
However,smokingrateislowandtheassociationsbetweentheVal
allele and smoking and between smoking and WM are weak and
therefore unlikely to explain the entire effect observed here.
The effects of genotype and genotype ? age interactions also
remained significant in the neuroimaging data when interactions
withsexwereincludedinthemodels.IntheIPS,bothbrainactivity
(trend in the right IPS) and GM volumes (bilateral IPS) exhibited an
interaction between sex and genotype, with greater genotype ef-
fects in males than females. In addition, a greater interaction effect
on IFG activation tended to be observed in males. There is wide-
spreadbutweakevidenceofaninteractionbetweensexandrs4680
onbehavior,brainstructure,andtheincidenceofpsychiatricdisor-
ders (47). Estrogens have been implicated to explain these sex
differences, as they are thought to downregulate COMT activity
(47). The current study included only small age groups when split
for sex. This limitation and the lack of pubertal stage information
prevented us from investigating sex differences during develop-
ment in more detail.
Our findings of rs4680 effects on GM volumes are consistent
with previous evidence for an effect of this polymorphism on both
adolescent and adult GM volumes (14,16) and cortical thickness
(15,18). However, the locations of the observed effects differ be-
tween studies, which may be due to differences in methodology
and age of the participants. The SNP rs4680 is a functional variant
and could be thought to affect brain function but not brain struc-
ture.However,thepruningofsynapsesandaxonsduringdevelop-
ment is affected by synaptic functioning (60,61). Thus, during mat-
uration,genotypeeffectsonbrainfunctioncouldgraduallyimpact
onbrainstructureandGMvolumesinparticular.Theweakergeno-
type by age interaction observed in the VBM compared with the
fMRI and behavioral data may be because these processes occur
over a protracted period and are not as flexible and rapid as func-
tional changes in brain activity.
It is unclear yet what the greater parietal GM volumes observed
in Val/Val individuals across ages precisely reflect and why the Val
carriersarethosewithgreatervolumes.Furtherinvestigationusing
a combination of cortical thickness, cortical surface area, and vol-
ume data may provide more insight on the underlying develop-
mental changes in structure (62). A possible interpretation of the
results of this study relates them to the altered trajectories of brain
structure maturation reported in developmental psychiatric disor-
ders (63). For example, delayed peaks in cortical thickness in atten-
tion-deficit/hyperactivity disorder (64), a disorder associated with
visuospatial WM impairments (65), lead to increased cortical thick-
ness in attention-deficit/hyperactivity disorder during late child-
hood and early adolescence. In the current study, Val/Val partici-
pants had greater gray matter volumes in the parietal cortex and
showed poorer WM capacity from mid-adolescence onward.
The effects of COMT genotype we observed were not limited to
thePFC.Inlinewiththeseresults,dopamineD1receptoravailability
(34) and correlation between D1 receptor availability and COMT
genotype(66)aresimilarthroughoutthecortex,anddopamineD1
receptor density and change in WM capacity after training are cor-
related in both prefrontal and parietal cortex (67). Recent multi-
modal studies combining positron emission tomography and fMRI
datatostudythedopaminesysteminadultshaverepeatedlyiden-
tified parietal and lateral PFC regions similar to the clusters ob-
servedinthepresentstudyandprovideindependentevidencethat
Figure4.COMTgenotype,workingmemory(WM)capac-
ity, and WM activity and their relation to a hypothetical
basal dopamine concentration. Brain activity is from the
right lateral prefrontal cortex (PFC). Eight- and 18-year-
old age groups were chosen to illustrate the changes in
the influence of COMT genotype on WM function during
development.Onthetoprow,WMcapacityZ-scoreswere
calculated using the Dot Matrix scores of the 8- and 18-
year-oldsseparatelyandareplottedintheinsetbarcharts
for each COMT genotype. The three COMT genotypes
were accordingly located on the hypothetical inverted-U
curveofprefrontalfunctioningasafunctionofdopamine
concentrationproposedfromadultandanimaldata(e.g.,
[25,27]). Here, an arbitrary Gaussian curve was used. On
the bottom row, similar steps were applied to the WM-
control activation in the right lateral PFC. Note that
greater brain activation is associated with worse perfor-
mance (27), and thus the y axis is inverted to highlight the
similarities between the behavioral and imaging results.
Eight-year-olds showed a detrimental effect of the Methio-
nine(Met)allele,while18-year-oldsshowedthetypicaladult
pattern of a beneficial effect of the Met allele, with higher
WM capacity and lower brain activity in the frontal cortex.
This pattern of change in genotype effect with age can be
regardedas a shift during development of the position of
the COMT Valine158Met genotypes on the inverted-U
curve of PFC function relative to dopamine levels in the
direction of lower basal dopamine levels. Met, Methio-
nine; PFC, prefrontal cortex; Val, Valine.
6 BIOL PSYCHIATRY 2011;xx:xxx
I. Dumontheil et al.
www.sobp.org/journal

Page 7

theseregionsaresensitivetobasaldopaminelevelsanddopamine
levels in interconnected subcortical regions (68–70).
InboththebehavioralandfMRIdatasets,weobservedagradual
emergence from age 12 of the adult pattern of COMT genotype
effects. Prefrontal cortex functioning and WM performance have
been suggested to follow an inverted-U curve function of dopa-
mine levels (24), with adult rs4680 Met/Met individuals near the
apex of this curve (2). The pattern observed in children (Figure 4)
resemblesthereductionoflateralPFCactivationandimprovement
in WM performance observed in Val/Val adults administered am-
phetamine, which increases basal dopamine levels (27). Thus, our
findings are compatible with a decrease in basal dopamine levels
duringadolescence(23),providingindirectevidenceforthedevel-
opment of the dopamine system in humans.
Increasedprefrontalactivitycouldreflecttop-downboostingof
WM capacity in the parietal cortex (71). The Val/Val individuals
showed an increase in prefrontal and parietal activation during
adolescence, a greater correlation between the activation in these
two clusters than Met carriers, and a correlation between greater
WM activity in the right IFG and better WM capacity outside the
scanner.Together,theseresultssuggestthatatop-downexcitation
from frontal to parietal cortex may be gradually implemented dur-
ing adolescence in Val/Val individuals to compensate for deficient
amounts of dopamine levels and parietal functioning.
Giventhattheonsetofmanypsychiatricconditionsoccursdur-
ingadolescence(72)andthatgenotypevariationshavebeenasso-
ciated with the incidence of developmental psychiatric disorders,
elucidating the role of genetics in determining brain function dur-
ing childhood and adolescence is critical to our understanding of
the development of these disorders. The findings presented here
show that the full developmental picture should be considered
when trying to understand the impact of genetic polymorphisms
onthematurecognitionofhealthyadultorpsychiatricpopulations.
This work was supported by Knut and Alice Wallenberg Founda-
tion,TheSwedishResearchCouncil,andaSwedishRoyalBankTercen-
tennial Foundation Grant in the program “Learning and Memory in
ChildrenandYoungAdults”toTKandJK.Datafromround1oftesting
waspreviouslypublishedinSöderqvistetal.(36).
WewouldliketothankJensGisselgård,YlvaSamuelsson,Douglas
Sjöwall, and Sissela Bergman Nutley for help with study administra-
tion; Keith Humphreys for statistics help; Kerstin Eriksson and Tomas
Jonsson for scanning; and the Mutation Analysis Core Facility at the
KarolinskaInstitutet.
The authors report no biomedical financial interests or potential
conflictsofinterest.
Supplementarymaterialcitedinthisarticleisavailableonline.
1. Craddock N, Owen MJ, O’Donovan MC (2006): The catechol-O-methyl
transferase (COMT) gene as a candidate for psychiatric phenotypes:
Evidence and lessons. Mol Psychiatry 11:446–458.
2. Meyer-Lindenberg A, Weinberger DR (2006): Intermediate phenotypes
and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci
7:818–827.
3. HowesOD,KapurS(2009):Thedopaminehypothesisofschizophrenia:
Version III--the final common pathway. Schizophr Bull 35:549–562.
4. Gothelf D, Hoeft F, Ueno T, Sugiura L, Lee AD, Thompson P, Reiss AL
(2011): Developmental changes in multivariate neuroanatomical pat-
ternsthatpredictriskforpsychosisin22q11.2deletionsyndrome.JPsy-
chiatrRes45:322–331.
5. VorstmanJA,MorcusME,DuijffSN,KlaassenPW,Heineman-deBoerJA,
Beemer FA, et al. (2006): The 22q11.2 deletion in children: High rate of
autistic disorders and early onset of psychotic symptoms. J Am Acad
ChildAdolescPsychiatry45:1104–1113.
6. Chen J, Lipska BK, Halim N, Ma QD, Matsumoto M, Melhem S, et al.
(2004): Functional analysis of genetic variation in catechol-O-methyl-
transferase (COMT): Effects on mRNA, protein, and enzyme activity in
postmortemhumanbrain.AmJHumGenet75:807–821.
7. Männistö PT, Kaakkola S (1999): Catechol-O-methyltransferase (COMT):
Biochemistry,molecularbiology,pharmacology,andclinicalefficacyof
the new selective COMT inhibitors. Pharmacol Rev 51:593–628.
8. Weinshilboum RM (2006): Pharmacogenomics: Catechol-O-methyl-
transferase to thiopurine S-methyltransferase. Cell Mol Neurobiol 26:
539–561.
9. Barnett JH, Jones PB, Robbins TW, Müller U (2007): Effects of the cate-
chol-O-methyltransferaseVal158Metpolymorphismonexecutivefunc-
tion: A meta-analysis of the Wisconsin Card Sort Test in schizophrenia
and healthy controls. Mol Psychiatry 12:502–509.
10. BarnettJH,ScorielsL,MunafòMR(2008):Meta-analysisofthecognitive
effects of the catechol-O-methyltransferase gene Val158/108Met poly-
morphism. Biol Psychiatry 64:137–144.
11. Flint J, Munafò MR (2007): The endophenotype concept in psychiatric
genetics. Psychol Med 37:163–180.
12. Dickinson D, Elvevåg B (2009): Genes, cognition and brain through a
COMT lens. Neuroscience 164:72–87.
13. MierD,KirschP,Meyer-LindenbergA(2010):Neuralsubstratesofpleio-
tropic action of genetic variation in COMT: A meta-analysis. Mol Psychi-
atry 15:918–927.
14. Cerasa A, Gioia MC, Labate A, Liguori M, Lanza P, Quattrone A, et al.
(2008):Impactofcatechol-O-methyltransferaseVal(108/158)Metgeno-
type on hippocampal and prefrontal gray matter volume. Neuroreport
19:405–408.
15. Cerasa A, Cherubini A, Quattrone A, Gioia MC, Tarantino P, Annesi G, et
al. (2010): Met158 variant of the catechol-O-methyltransferase geno-
type is associated with thicker cortex in adult brain. Neuroscience 167:
809–814.
16. Honea R, Verchinski BA, Pezawas L, Kolachana BS, Callicott JH, Mattay
VS, et al. (2009): Impact of interacting functional variants in COMT on
regional gray matter volume in human brain. Neuroimage 45:44–51.
17. Mechelli A, Tognin S, McGuire PK, Prata D, Sartori G, Fusar-Poli P, et al.
(2009):Geneticvulnerabilitytoaffectivepsychopathologyinchildhood:
A combined voxel-based morphometry and functional magnetic reso-
nance imaging study. Biol Psychiatry 66:231–237.
18. Shaw P, Wallace GL, Addington A, Evans A, Rapoport J, Giedd JN, et al.
(2009): Effects of the Val158Met catechol-O-methyltransferase poly-
morphismoncorticalstructureinchildrenandadolescents.MolPsychi-
atry 14:348–349.
19. Thomason ME, Dougherty RF, Colich NL, Perry LM, Rykhlevskaia EI,
LouroHM,etal.(2010):COMTgenotypeaffectsprefrontalwhitematter
pathways in children and adolescents. Neuroimage 53:926–934.
20. ThomasonME,WaughCE,GloverGH,GotlibIH(2009):COMTgenotype
and resting brain perfusion in children. Neuroimage 48:217–222.
21. Barnett JH, Heron J, Ring SM, Golding J, Goldman D, Xu K, et al. (2007):
Gender-specific effects of the catechol-O-methyltransferase Val108/
158Met polymorphism on cognitive function in children. Am J Psychia-
try 164:142–149.
22. Diamond A, Briand L, Fossella J, Gehlbach L (2004): Genetic and neuro-
chemicalmodulationofprefrontalcognitivefunctionsinchildren.AmJ
Psychiatry 161:125–132.
23. Wahlstrom D, White T, Hooper CJ, Vrshek-Schallhorn S, Oetting WS,
Brott MJ, et al. (2007): Variations in the catechol O-methyltransferase
polymorphism and prefrontally guided behaviors in adolescents. Biol
Psychiatry 61:626–632.
24. VijayraghavanS,WangM,BirnbaumSG,WilliamsGV,ArnstenAF(2007):
Inverted-U dopamine D1 receptor actions on prefrontal neurons en-
gaged in working memory. Nat Neurosci 10:376–384.
25. Arnsten AF (1997): Catecholamine regulation of the prefrontal cortex.
JPsychopharmacol(Oxford)11:151–162.
26. Giakoumaki SG, Roussos P, Bitsios P (2008): Improvement of prepulse
inhibition and executive function by the COMT inhibitor tolcapone
depends on COMT Val158Met polymorphism. Neuropsychopharmacol-
ogy 33:3058–3068.
27. Mattay VS, Goldberg TE, Fera F, Hariri AR, Tessitore A, Egan MF, et al.
(2003): Catechol O-methyltransferase val158–met genotype and indi-
vidual variation in the brain response to amphetamine. Proc Natl Acad
SciUSA100:6186–6191.
I. Dumontheil et al.
BIOL PSYCHIATRY 2011;xx:xxx 7
www.sobp.org/journal

Page 8

28. Goldman-Rakic PS, Brown RM (1982): Postnatal development of mono-
amine content and synthesis in the cerebral cortex of rhesus monkeys.
BrainRes256:339–349.
29. Goldman-RakicPS,BrownRM(1981):Regionalchangesofmonoamines
in cerebral cortex and subcortical structures of aging rhesus monkeys.
Neuroscience 6:177–187.
30. Lambe EK, Krimer LS, Goldman-Rakic PS (2000): Differential postnatal
developmentofcatecholamineandserotonininputstoidentifiedneu-
rons in prefrontal cortex of rhesus monkey. J Neurosci 20:8780–8787.
31. RosenbergDR,LewisDA(1994):Changesinthedopaminergicinnerva-
tion of monkey prefrontal cortex during late postnatal development: A
tyrosine hydroxylase immunohistochemical study. Biol Psychiatry 36:
272–277.
32. Rosenberg DR, Lewis DA (1995): Postnatal maturation of the dopami-
nergic innervation of monkey prefrontal and motor cortices: A tyrosine
hydroxylase immunohistochemical analysis. J Comp Neurol 358:383–
400.
33. Weickert CS, Webster MJ, Gondipalli P, Rothmond D, Fatula RJ, Herman
MM, et al. (2007): Postnatal alterations in dopaminergic markers in the
human prefrontal cortex. Neuroscience 144:1109–1119.
34. JucaiteA,ForssbergH,KarlssonP,HalldinC,FardeL(2010):Age-related
reduction in dopamine D1 receptors in the human brain: From late
childhoodtoadulthood,apositronemissiontomographystudy.Neuro-
science 167:104–110.
35. LunaB,PadmanabhanA,O’HearnK(2010):WhathasfMRItoldusabout
thedevelopmentofcognitivecontrolthroughadolescence?BrainCogn
72:101–113.
36. Söderqvist S, McNab F, Peyrard-Janvid M, Matsson H, Humphreys K,
Kere J, Klingberg T, et al. (2010): The SNAP25 gene is linked to working
memorycapacityandmaturationoftheposteriorcingulatecortexdur-
ing childhood. Biol Psychiatry 68:1120–1125.
37. AllowayTP(2007):AutomatedWorkingMemoryAssessmentManual.Ox-
ford, UK: Harcourt.
38. McCulloch CE, Searle SR (2001): Generalized, Linear, and Mixed Models.
New York: Wiley-Interscience.
39. Molenberghs G, Verbeke G (2001): A review on linear mixed models for
longitudinal data, possibly subject to dropout. Stat Model 1:235–269.
40. Starr JM, Fox H, Harris SE, Deary IJ, Whalley LJ (2007): COMT genotype
and cognitive ability: A longitudinal aging study. Neurosci Lett 421:
57–61.
41. SPSS,Inc.(2005):LinearMixed-EffectsModelinginSPSS:anIntroductionto
theMIXEDProcedure.SPSSTechnicalReport.Chicago:SPSS,Inc.
42. KailRV,FerrerE(2007):Processingspeedinchildhoodandadolescence:
Longitudinal models for examining developmental change. Child Dev
78:1760–1770.
43. LunaB,GarverKE,UrbanTA,LazarNA,SweeneyJA(2004):Maturationof
cognitive processes from late childhood to adulthood. Child Dev 75:
1357–1372.
44. Aiken LS, West SG (1991): Multiple Regression: Testing and Interpreting
Interactions. Newbury Park, CA: Sage.
45. Akaike H (1974): A new look at the statistical model identification. IEEE
TransAutomatContr19:716–723.
46. Schwartz G (1978): Estimating the dimension of a model. Ann Stat
6:461–464.
47. Harrison PJ, Turnbridge EM (2008): Catechol-O-methyltransferase
(COMT):Agenecontributingtosexdifferencesinbrainfunction,andto
sexual dimorphism in the predisposition to psychiatric disorders. Neu-
ropsychopharmacology 33:3037–3045.
48. Hersrud SL, Stoltenberg SF (2009): Epistatic interaction between COMT
and DAT1 genes on eating behavior: A pilot study. Eat Behav 10:131–
133.
49. Van Essen DC (2005): A Population-Average, Landmark- and Surface-
based (PALS) atlas of human cerebral cortex. Neuroimage 28:635–662.
50. VanEssenDC,DierkerDL(2007):Surface-basedandprobabilisticatlases
of primate cerebral cortex. Neuron 56:209–225.
51. VanEssenDC,DruryHA,DicksonJ,HarwellJ,HanlonD,AndersonCH,et
al. (2001): An integrated software suite for surface-based analyses of
cerebralcortex.JAmMedInformAssoc8:443–459.
52. BrettM,ValabregueR,PolineJ-B(2002):Regionofinterestanalysisusing
an SPM toolbox [abstract]. Presented at the 8th International Confer-
ence on Functional Mapping of the Human Brain, June 2–6, Sendai,
Japan.
53. Ashburner J (2007): A fast diffeomorphic image registration algorithm.
Neuroimage 38:95–113.
54. NakagawaS,CuthillIC(2007):Effectsize,confidenceintervalandstatis-
tical significance: A practical guide for biologists. Biol Rev Camb Philos
Soc 82:591–605.
55. Saraceno L, Munafó M, Heron J, Craddock N, van den Bree MB (2009):
Genetic and non-genetic influences on the development of co-occur-
ring alcohol problem use and internalizing symptomatology in adoles-
cence: A review. Addiction 104:1100–1121.
56. HvitfeldtT,GripeI(2009):Rapport118SkoleleversDrogvanor2009.Stock-
holm: Centralförbundet för Alkohol-Och Narkotikaupplysning.
57. Galanti MR, Rosendahl I, Wickholm S (2008): The development of to-
baccouseinadolescenceamong”snusstarters”and”cigarettestarters”:
An analysis of the Swedish ”BROMS” cohort. Nicotine Tob Res 10:315–
323.
58. TammimäkiAE,MännistöPT(2010):AregeneticvariantsofCOMTasso-
ciated with addiction? Pharmacogenet Genomics 20:717–741.
59. Durazzo TC, Meyerhoff DJ, Nixon SJ (2010): Chronic cigarette smoking:
Implications for neurocognition and brain neurobiology. Int J Environ
ResPublicHealth7:3760–3791.
60. Webb SJ, Monk CS, Nelson CA (2001): Mechanisms of postnatal neuro-
biological development: Implications for human development. Dev
Neuropsychol 19:147–171.
61. Low LK, Cheng HJ (2006): Axon pruning: An essential step underlying
thedevelopmentalplasticityofneuronalconnections.PhilosTransRSoc
LondBBiolSci361:1531–1544.
62. Winkler AM, Kochunov P, Blangero J, Almasy L, Zilles K, Fox PT, et al.
(2010): Cortical thickness or grey matter volume? The importance of
selecting the phenotype for imaging genetics studies. Neuroimage 53:
1135–1146.
63. Giedd JN, Rapoport JL (2010): Structural MRI of pediatric brain develop-
ment:Whathavewelearnedandwherearewegoing?Neuron67:728–734.
64. ShawP,EckstrandK,SharpW,BlumenthalJ,LerchJP,GreensteinD,etal.
(2007): Attention-deficit/hyperactivity disorder is characterized by a
delayincorticalmaturation.ProcNatlAcadSciUSA104:19649–19654.
65. Westerberg H, Hirvikoski T, Forssberg H, Klingberg T (2004): Visuo-spa-
tial working memory span: A sensitive measure of cognitive deficits in
children with ADHD. Child Neuropsychol 10:155–161.
66. SlifsteinM,KolachanaB,SimpsonEH,TabaresP,ChengB,DuvallM,etal.
(2008):COMTgenotypepredictscortical-limbicD1receptoravailability
measured With [11C]NNC112 and PET. Mol Psychiatry 13:821–827.
67. McNab F, Varrone A, Farde L, Jucaite A, Bystritsky P, Forssberg H, Kling-
berg T (2009): Changes in cortical dopamine D1 receptor binding asso-
ciated with cognitive training. Science 323:800–802.
68. Fischer H, Nyberg L, Karlsson S, Karlsson P, Brehmer Y, Rieckmann A, et
al. (2010): Simulating neurocognitive aging: Effects of a dopaminergic
antagonist on brain activity during working memory. Biol Psychiatry
67:575–580.
69. LandauSM,LalR,O’NeilJP,BakerS,JagustWJ(2009):Striataldopamine
and working memory. Cereb Cortex 19:445–454.
70. Nyberg L, Andersson M, Forsgren L, Jakobsson-Mo S, Larsson A, Mark-
lund P, et al. (2009): Striatal dopamine D2 binding is related to frontal
BOLDresponseduringupdatingoflong-termmemoryrepresentations.
Neuroimage 46:1194–1199.
71. Edin F, Klingberg T, Johansson P, McNab F, Tegnér J, Compte A, et al.
(2009): Mechanism for top-down control of working memory capacity.
ProcNatlAcadSciUSA106:6802–6807.
72. Paus T, Keshavan M, Giedd JN (2008): Why do many psychiatric disor-
ders emerge during adolescence? Nat Rev Neurosci 9:947–957.
8 BIOL PSYCHIATRY 2011;xx:xxx
I. Dumontheil et al.
www.sobp.org/journal