Dysbindin modulates brain function during visual processing in children

NeuroImage 49 (2010) 817–822

Contents lists available at ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r.com/ locate /yn img

Dysbindin modulates brain function during visual processing in children

A. Mechelli a,b,⁎, E. Viding c,d, A. Kumar b, W. Pettersson-Yeo b, P. Fusar-Poli b, S. Tognin b,M.C. O'Donovan e, P. McGuire b

a Department of Psychology, Institute of Psychiatry, King's College London, 103 Denmark Hill, London, SE5 8AF, UKb Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, 103 Denmark Hill, London, SE5 8AF, UKc Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King's College London, 103 Denmark Hill, London, SE5 8AF, UKd Division of Psychology and Language Sciences, University College London, Gower St., London, WC1E 6BT, UKe Division of Department of Psychological Medicine, School of Medicine, Cardiff University, Heath Park, Cardiff, CF14 4XN, UK

⁎ Corresponding author. PO Box 67, Institute of PsychiaDenmark Hill, London, SE5 8AF, UK.

E-mail address: [email protected] (A. Mechell

1053-8119/$ – see front matter © 2009 Elsevier Inc. Aldoi:10.1016/j.neuroimage.2009.07.030

a b s t r a c t

Article history:Received 18 May 2009Revised 10 July 2009Accepted 14 July 2009Available online 22 July 2009

Keywords:DysbindinVisual processingChildrenSchizophreniaFunctional Magnetic Resonance Imaging

Schizophrenia is a neurodevelopmental disorder, and risk genes are thought to act through disruption ofbrain development. Several genetic studies have identified dystrobrevin binding protein 1 (DTNBP1, alsoknown as dysbindin) as a potential susceptibility gene for schizophrenia, but its impact on brain function ispoorly understood. It has been proposed that DTNBP1 may be associated with differences in visualprocessing. To test this, we examined the impact on visual processing in 61 healthy children aged 10–12 yearsof a genetic variant in DTNBP1 (rs2619538) that was common to all schizophrenia associated haplotypes inan earlier UK-Irish study. We tested the hypothesis that carriers of the risk allele would show altered occipitalcortical function relative to noncarriers. Functional Magnetic Resonance Imaging (fMRI) was used to measurebrain responses during a visual matching task. The data were analysed using statistical parametric mappingand statistical inferences were made at pb0.05 (corrected for multiple comparisons). Relative to noncarriers,carriers of the risk allele had greater activation in the lingual, fusiform gyrus and inferior occipital gyri. Inthese regions DTNBP1 genotype accounted for 19%, 20% and 14% of the inter-individual variance, respectively.Our results suggest that that genetic variation in DTNBP1 is associated with differences in the function ofbrain areas that mediate visual processing, and that these effects are evident in young children. Thesefindings are consistent with the notion that the DTNBP1 gene influences brain development and can therebymodulate vulnerability to schizophrenia.

© 2009 Elsevier Inc. All rights reserved.

Introduction

Schizophrenia is a severe psychiatric disorder with a strong geneticcomponent (Owen et al., 2005, 2007; Craddock et al., 2005). Theidentification of susceptibility genes for schizophrenia has beenproblematic because of the complexity of the phenotype and themode of transmission compatible with a multi-locus model. However,recent association studies have provided converging evidence infavour of a number of positional gene linkage regions; amongst thestrongest candidates is the gene encoding for the dystrobrevin bindingprotein-1 (DTNBP1), otherwise known as dysbindin (Harrison andWeinberger, 2005; Owen et al., 2005).

Located at chromosomal position 6p22.3, DTNBP1 was firstimplicated as a candidate gene for schizophrenia by Straub et al.(2002) who undertook association analysis of a region of linkagefinding they had previously reported. A number of subsequent studiesin German (Schwab et al., 2003), Irish (Oord et al., 2003), Swedish

try, King's College London, 103

i).

l rights reserved.

(Van Den Bogaert et al., 2003), Japanese (Numakawa et al., 2004;Tochigi et al., 2006), Bulgarian (Kirov et al., 2004), British (Williams etal., 2004), and Chinese (Li et al., 2005; Tang et al., 2003), populationshave since added further support for association between schizo-phrenia and DTNBP1.

The nature of DTNBP1 function within the brains of healthyindividuals is poorly understood; one of the key findings to date hasbeen the implication of DTNBP1 in the alteration of glutamate/NMDAneurotransmission (Numakawa et al., 2004). Studies involving knock-down of endogenous DTNBP1 with small interfering RNA have showna reduction of glutamate levels in neurons in culture (Ross et al.,2006), suggesting a possible synaptic consequence for reductions inDTNBP1 levels (Numakawa et al., 2004; Talbot et al., 2004). The effectsof DTNBP1 on brain structure and function at system level arecurrently unknown as to date no neuroimaging studies haveinvestigated the impact of this gene on brain structure or functionin human participant.

Given the possible association between the DTNBP1 gene andschizophrenia, it is important to elucidate the biological mechanismsby which the DTNBP1 gene may affect brain function to increasebiological susceptibility to the disorder. Patients with schizophrenia

Table 1Participants' characteristics and task performance.

TT TA AA All subjects p-value

N 23 25 13 61Age in months 133.1 (8.0) 138.3 (7.9) 133.8 (6.2) 135.4 (7.9) 0.06Full scale IQ 107.0 (12.2) 106.4 (11.9) 102.3 (10.2) 105.7 (11.6) 0.49

Parent rated SDQ scoresHyperactivity 3.3 (1.8) 3.6 (2.8) 2.7 (2.2) 3.3 (2.3) 0.55Conduct problems 1.7 (2.1) 1.2 (1.3) 1.6 (1.5) 1.5 (1.6) 0.60Emotional problems 1.0 (1.4) 1.8 (2.0) 1.5 (2.0) 1.4 (1.8) 0.27Prosociality 8.1 (2.0) 8.3 (1.7) 9.2 (1.1) 8.4 (1.7) 0.21Peer problems 1.3 (2.0) 0.9 (1.8) 1.3 (1.5) 1.1 (1.8) 0.75Total difficulties 7.2 (4.9) 7.5 (6.0) 7.2 (5.4) 7.3 (5.4) 0.98

Teacher rated SDQ scoresHyperactivity 4.1 (3.0) 3.7 (3.0) 3.0 (3.1) 3.7 (3.0) 0.60Conduct problems 2.2 (2.7) 1.0 (1.6) 1.3 (2.1) 1.5 (2.2) 0.20Emotional problems 1.9 (2.3) 1.2 (1.4) 1.7 (1.8) 1.6 (1.9) 0.45Prosociality 5.6 (3.0) 6.9 (3.0) 6.6 (2.5) 6.3 (2.9) 0.34Peer problems 1.0 (1.7) 1.0 (1.7) 1.5 (2.6) 1.1 (1.9) 0.73Total difficulties 9.2 (7.0) 6.7 (5.8) 7.5 (6.1) 7.3 (5.4) 0.43

Task performanceErrors (%) 4.9 (4.2) 4.3 (6.2) 3.8 (3.9) 4.4 (5.0) 0.83RT (ms) 1402 (231) 1427 (288) 1399 (202) 1412 (248) 0.93

Data are expressed as mean values (±standard deviation). p-values refer to a one-wayANOVA contrasting the different genotypes. n = number of subjects; SDQ = Strengthsand Difficulties Questionnaire; RT = reaction times.

818 A. Mechelli et al. / NeuroImage 49 (2010) 817–822

show behavioural and neuronal deficits in visual processing, particu-larly in the magnocellular system which is glutamate/NMDA-dependent but also in the parvocellular system (see Butler and Javitt,2005 for review). These deficits have been demonstrated usingbehavioural tests (Schechter et al., 2003), as well as in electroence-phalographic (Butler et al., 2005) and functional Magnetic ResonanceImaging (Martínez et al., 2008) studies and are thought to contributeto the impairments in spatial working memory which are welldocumented in schizophrenia (Piskulic et al., 2007; Donohoe et al.,2008). Based on the observations that DTNBP1 is expressed in theoccipital cortex and is implicated in glutamate/NMDA neurotransmis-sion, it has been proposed that genetic variation in DTNBP1 may beassociated with differences in visual processing (Donohoe et al.,2008). Consistent with this hypothesis, Donohoe et al. (2008) havedemonstrated that carriers of the DTNBP1 schizophrenia riskhaplotype that was previously reported in a UK-Irish study (Williamset al., 2004) express early visual processing deficits compared tononcarriers, as indexed by significantly reduced P1 visual evokedpotential. The authors argue that this finding reveals a possiblebiological mechanism by which the DTNBP1 gene may increasevulnerability to the disorder. However, because that investigation wascarried out in patients with schizophrenia, it is currently unknownwhether genetic variation in DTNBP1 also affects visual processing inhealthy participants. This is not a trivial question given the increasingnumber of studies providing evidence for a disease-specific pattern ofgene action (e.g. Frodl et al., 2004; Addington et al., 2007; Mechelli etal., 2008). For example, the impact of the DTNBP1 gene inschizophrenia might depend on the presence of other genetic orenvironmental risk factors andmight not apply to typically developingindividuals. In addition, the effects of DTNBP1 genotype on visualprocessing were identified in mature adults, and it is not knownwhether they are already evident in children. This is anotherimportant issue, since schizophrenia is thought to be a neurodevelop-mental disorder, and risk genes are expected to act through disruptionof brain development (Weinberger, 1987). If gene-related differencesin adult participants reflect alterations that occurred during childhoodand adolescence, a better characterization of these alterations inchildren and adolescents is critical for understanding howgenes affectbrain function to mediate vulnerability to psychopathology (Viding etal., 2006).

In the present study, we investigated the impact of the DTNBP1gene on visual processing in a cohort of 61 healthy children of 10–12 years of age. Functional Magnetic Resonance Imaging (fMRI) wasused to measure brain responses during a visual matching task whichengaged the occipital cortex. In the absence of consensus evidenceconcerning the most relevant SNP association across studies, weselected rs2619538 for study because it was the best-associated singleSNP from a previous UK based (like the present study) study ofschizophrenia and was also present on all haplotypes showingassociation with the disorder in that study (Williams et al., 2004).The SNP, which lies about 2 kb 5′ to the start of DTNBP1 transcriptNM_183041, was also associated in a UK sample with psychotic mania(Raybould et al., 2005) and is carried on the haplotype that is moststrongly associated with cis-acting variants that influence DTNBP1expression (Bray et al., 2005), on a haplotype showing evidence forassociation in a second UK schizophrenia sample (Li et al., 2005) andon a haplotype earlier reported to be associated with early visualprocessing deficits (Donohoe et al., 2008). It has also been reported tobe individually associated with methamphetamine psychosis (Kishi-moto et al., 2008) and to show a trend for association to severe andpersistent pathology in schizophrenia (Tosato et al., 2007). Thus,while there is no consensus about which markers might be optimal totest, the SNP we selected is associated with haplotypes that arethemselves associated with psychosis, with function as indexed bygene expression, and with a relevant neurocognitive task. We testedthe hypothesis that individuals with the risk allele of the gene would

express altered activation of the occipital cortex relative to thosewithout the risk allele. The rationale was that DTNBP1 genotypewould influence visual processing in the healthy population as well asin schizophrenia, and that this effect would already be evident inchildhood.

Materials and methods

Subjects

A total of 61 boys aged 10–12 years participated in the study. Theywere recruited from the Twins Early Development Study (TEDS)database, as part of an on-going twin neuroimaging project. Only onetwin per pair was included in the present investigation in order toavoid bias due to non-independence of twin data. The sampling frameincluded typically developing children (82% of the analysis sample), aswell as children in the top 10% of the U.K. population for conductproblems (18% of the analysis sample). Crucially for the interpretationof the present data, none of the children included in our sample metdiagnostic criteria for conduct disorder and the mean score onconduct problems or associated hyperactivity was not in the abnormalrange (see Table 1); furthermore, performance accuracy and speed didnot differ (pN0.05) between typically developing children andchildren in the top 10% of the U.K. population for conduct problems.The short version of the Wechsler Abbreviated Scales of Intelligence(WASI) was used to assess IQ (Wechsler, 1999). The Strengths andDifficulties Questionnaire (SDQ) was employed to measure emotionalproblems, conduct problems and hyperactivity, using ratings fromboth parents and teachers (Goodman et al., 2000). All participantswere genotyped for the dysbindin gene (see below), with T and Abeing the low-risk and high-risk alleles, respectively. The samplecomprised 23 childrenwith the TT genotype, 25 with the TA genotypeand 13 with the AA genotype; these frequencies are consistent withthose of previous studies of the U.K. population (Williams et al., 2004).The three genotype groups did not differ in age, IQ, or any of the SDQindicators (one-way ANOVA; pN0.05; Table 1). Genotypic distributionwas similar for typically developing children (19 TT, 20 TA, 11 AA) andchildren in the top 10% of the U.K. for conduct problems (4 TT, 5 TA, 2AA). All participants were Caucasianwith the exception of one subject

819A. Mechelli et al. / NeuroImage 49 (2010) 817–822

in the heterozygote group (TA) who was of Asian ethnicity. Writtenconsent was obtained from all participants and their parents inaccordance with protocols approved by the Local and MulticentreResearch Ethics Committee (LREC, MREC).

Genotyping

DNA was extracted from blood or cheek swabs using standardmethods (Freeman et al., 2003). Genotyping of the single nucleotidepolymorphism rs2619538 was performed by KBioscience (http://www.kbioscience.co.uk), using a competitive allele specific PCRsystem (CASP). The genotyping results were under Hardy–Weinbergequilibrium.

Visual matching task

Each stimulus comprised three images in black and white, withone presented at the bottom of the screen and two presented in its topleft and top right corners, respectively. The task required subjects toclick the left or right button on a key pad, depending on whether theimage at the bottom matched the image presented in the top left orthe top right of the screen. Each set of three images comprised eitherphotographs or geometric shapes presented in blocks of 6; photo-graphs represented either animals or landscapes whereas geometricobjects were either vertical or horizontal ellipses. For the purpose ofthe present investigation, photographs and geometric shapes wereanalysed together, as an initial statistical analysis did not reveal anysignificant interaction between DTNBP1 genotype and image type. Atotal of 48 stimuli were presented, with each stimulus remaining onthe screen for 3 s and an interval between stimuli of 2 s, resulting in astimulus onset asynchrony of 5 s. Two blocks of rest, each lasting 34 s,were also included; during these blocks of rest participants wereinstructed to focus on a fixation cross in the middle of the screen. Thetotal acquisition time was 308 s.

Image acquisition

Neuroimaging data were acquired using a General Electric Signa3.0 Telsa Excite II MRI scanner (Medical systems, Milwaukee,WI, USA)at the Centre for Neuroimaging Science, Institute of Psychiatry. Inorder to exclude the presence of gross anatomical abnormalities, astructural brain image was acquired from each subject using anisotropic resolution 3D inversion recovery prepared spoiled gradientecho (IR-SPGR). Parameters for the IR-SPGR were TR=8 ms;TE=2.9 ms; TI=450 ms; and excitation FA=20°. The in-planematrix size was 256×192 over a 280×210 mm field of view,reconstructed to 256×256 over 280×280 mm. Two hundred throughplane partitions (each 1.1 mm thick) were collected, with twopartitions being discarded at each end of the imaging volume tominimize wrap-round artefacts, resulting in a scanning time of 6 min.In addition, a total of 160 functional images, each comprising 28 slices,were acquired in a single run with a T2⁎-weighted gradient echo-planar imaging sequence (EPI) and a repetition time (RT) of 2 s (slicethickness 3.5 mm, gap= 0.3 mm; TE= 25 ms, field ofview=220×220, matrix size 64×64). The orientation of the axialslices was parallel to the ACPC line. Stimuli were projected onto ahigh-resolution screen located in front of the participant's head, andwere viewed viamirror attached to the head coil. Theywere presentedusing Visual Basic software and synchronized with pulses generatedby the scanner at the beginning of each scan sequence.

Image analysis

The analysis was performed using SPM5 software (Friston, 2003),running under Matlab 6.5. In brief, all volumes from each subject wererealigned and unwarped to minimize movement-related artefacts,

normalized to a standard template and spatially smoothed with a6 mm full width at half maximum isotropic Gaussian kernel. There area number of studies providing evidence that atlas-transformed brainmorphology is relatively consistent between 7 and 8 year old childrenand adults at a resolution appropriate to functional MagneticResonance Imaging (Burgund et al., 2002; Kang et al., 2003). Thus, astandard MNI-305 template was used for special normalization, as inprevious functional neuroimaging investigations in children (e.g.Bitan et al., 2009). First, the statistical analysis of regional responseswas performed in a subject-specific fashion, by convolving each onsettime with a synthetic haemodynamic response function (HRF). Inorder to minimize performance confounds, incorrect trials in whichthe subject did not respond correctly were discarded, using an event-related model (Mechelli et al., 2003). In order to remove low-frequency drifts, the data were high-pass filtered using a set ofdiscrete cosine basis functions with a cutoff period of 128 s. Aftercalculating the parameter estimates for all voxels using the generallinear model, a contrast image was computed for the comparisonbetween visual stimuli and fixation in each subject independently.Second, the subject-specific contrast images were entered into a full-factorial ANOVA to permit inferences at the second level (Penny andHolmes, 2003). This allowed us to examine the effect of visualstimulation (i.e. visual stimuliNfixation) across all participants, aswell as the impact of DTNBP1 genotype on the effect of visualstimulation. Because the aim of the present investigation was toexamine the impact of DTNBP1 on visual processing, statisticalinferences were constrained to the occipital cortex using a bilateralmask available from the WFU Pickatlas toolbox. This mask includedthe lingual gyrus, the fusifurm gyrus, the inferior occipital gyrus, themiddle occipital gyrus, the superior occipital gyrus and the precuneusand comprised of a total of 15221 voxels. Statistical inferences weremade using a statistical threshold of pb0.05 with family-wise error(FWE) correction for multiple comparisons and an extent threshold of10 voxels. In regions where there was a significant effect of genotype,we used the R2 measure of effect size to assess howmuch of the inter-individual variance in BOLD activation was explained by the geneticvariation.

Results

Behavioural performance

Analysis of response accuracy and reaction times was performedusing SPSS software; the three genotypic groups were compared usinga one-way ANOVA. Variation in DTNBP1 genotype was not associatedwith significant differences in response accuracy (F=0.183 df=2,58p=0.833) or reaction times (F=0.077 df=2,58 p=0.93). Table 1reports the number of errors and the reaction times for each genotypicgroup and for all subjects combined.

Effect of visual processing

As expected, the visual matching task was associated with strongactivation of in the occipital cortex bilaterally (Fig. 1), including thelingual gyrus (left: x=−10 y=−82 z=8 z-score=4.6; right: x=−10 y=−82 z=8 z-score=6.7), the fusiform gyrus (left: x=28y=−68 z=−12 z-score=7.8; right: x=28 y=−68 z=−16 z-score=7.4), the inferior occipital gyrus (left: x=−32 y=−82 z=−14 z-score=6.3; right: x=34 y=−80 z=−16 z-score=6.6), themiddle occipital gyrus (left: x=−32 y=−92 z=−2 z-score=5.7;right: x=34 y=−88 z=2 z-score=7.6), the superior occipitalgyrus (left: x=−24 y=−88 z=20 z-score=7.4; right: x=38 y=−86 z=22 z-score=7.8, the cuneus (left: x=−24 y=−88 z=22z-score=8.5; right: x=18 y=−94 z=18; z-score=7.9) and theprecuneus (left: x=−18 y=−70 z=26 z-score=7.4; right: x=16

Fig. 1. Regions of the occipital cortex activated during the visual matching task relative to fixation (pb0.05 after FWE correction).

820 A. Mechelli et al. / NeuroImage 49 (2010) 817–822

y=−66 z=26 z-score=6.6); all effects were significant at b0.05after FWE correction for multiple comparisons.

Effect of DTNBP1 genotype

The main effect of genotype was examined by comparing brainactivation during visual stimulation relative to fixation in the differentgenotype groups. Contrasting participants with either one or two riskalleles (n=38) against those without any risk alleles (n=23)revealed that the former expressed greater activation in three regionswithin the left occipital cortex, focused on the inferior occipital (x=−38 y=−84 z=−14; z-score=4.4 p-value=0.010 after FWEcorrection), lingual (x=−18 y=−86 z=−20; z-score=4.2 p-value=0.017 after FWE correction) and fusiform gyri (x=−30 y=

Fig. 2. Location and parameter estimates of regions showing a significant effect of DTNBP1 genoccipital gyrus (x=−38 y=−84 z=−14); middle: lingual gyrus (x=−18 y=−86 z=−high-risk variant of the DTNBP1 gene respectively.

−72 z=−20; z-score=4.0 p-value=0.047), respectively (Fig. 2).We then contrasted participants with two risk alleles (n=13) againstthose without any risk alleles (n=23), to examine whether therewere additional differences when the two most genetically hetero-geneous groups were compared. This contrast did not reveal anyadditional effects, but confirmed that the risk allele was againassociated with greater activation in the left fusiform (x=−26 y=−76 z=−20; z-score=4.1 p=0.026 after FWE correction), lingual(x=−18 y=−86 z=−20; z-score=4.1 p=0.031 after FWEcorrection) and inferior occipital gyri (x=−38 y=−84 z=−12;z-score=3.4 p-value=0.304 after FWE correction). All effectsremained significant (pb0.05 after FWE) when the statistical analysiswas repeated including a covariate of no interest which modelledtypically developing children and children in the top 10% of the U.K.

otype on activation during visual matching (pb0.05 after FWE correction). Top: inferior20); bottom: fusiform gyrus (x=−30 y=−72 z=−20). T and A refer to the low- and

821A. Mechelli et al. / NeuroImage 49 (2010) 817–822

population for conduct problems separately. Estimate of the R2

measure for the comparison between brain activation during visualstimulation and fixation revealed that DTNBP1 genotype accountedfor 19%, 20% and 14% of the inter-individual variance in brainactivation in the left fusiform gyrus, the left lingual gyrus and theinferior occipital gyrus, respectively. Brain activation in these regionsdid not significantly correlate with performance, either in terms oferrors and reaction times (pN0.05).

Although the focus of the present investigation was on the impactof DTNBP1 in the occipital cortex, we also performed an exploratoryanalysis of the whole brain for completeness, given the evidence forgeneral role of DTNBP1 in brain function including not only perceptualbut also high level processes (Burdick et al., 2006; Fallgatter et al.,2006; Burdick et al., 2007; Donohoe et al., 2007). This secondaryanalysis did not detect any effects of DTNBP1 outside the visual cortex,even when lowering the statistical threshold to p b0.001(uncorrected).

Discussion

The aim of the present investigation was to examine the impact ofvariation in the DTNBP1 gene on brain function during visualprocessing in children. To our knowledge, this is the first functionalneuroimaging study of the effects of DTNBP1 on brain activation. Wetested the hypothesis that the variant that has previously beenassociated with increased risk of schizophrenia would be associatedwith altered activation of the occipital cortex during a visual matchingtask.

A previous investigation (22) measuring visual evoked potentialshad reported that, in patients with schizophrenia, carriers of theDTNBP1 risk haplotype displayed impairments in early stage visualprocessing compared to noncarriers (Donohoe et al., 2008). Ourfindings extend these results by indicating that the risk allele isassociated with increased activation of the occipital cortex during atask that engages visual processing in healthy young children.Examination of the group-specific parameter estimates revealed thatthe lingual, fusiform gyrus and inferior frontal gyri were activated incarriers of the risk allele but showed no activation in noncarriers; thispattern of responses suggests that task performance was associatedwith the recruitment of additional regions in individuals with thehigh-risk allele relative to those without it. Moreover, DTNBP1genotype accounted for a considerable proportion (up to 20%) of theinter-individual variance in activation in these regions.

Our data thus suggest that the influence of genetic variation inDTNBP1 on visual processing is not specific to patients withschizophrenia but is also evident in healthy participants. This is nota trivial observation, given the increasing number of studies providingevidence for a disease-specific pattern of gene action (e.g. Frodl et al.,2004; Addington et al., 2007; Mechelli et al., 2008). The finding thateffects of the DTNBP1 gene are already evident in children as young as10–12 years of age is important, as previous reports of an associationbetween the DTNBP1 genotype and psychotic illness have been basedon adult participants (Straub et al., 2002; Schwab et al., 2003; Oord etal., 2003; Van Den Bogaert et al., 2003; Numakawa et al., 2004; Tochigiet al., 2006; Kirov et al., 2004; Williams et al., 2004; Li et al., 2005;Tang et al., 2003). However much of adult psychopathology may berooted early in life, first emerging during childhood and adolescence(Kim-Cohen et al., 2003), and our data are consistent with the notionthat risk genes for schizophrenia act through disruption of neurode-velopmental processes from an early age (Gornick et al., 2005;Weinberger, 1987). The observation of a significant effect of DTNBP1genotype on the visual cortex of healthy participants suggests thatbehavioural deficits in visual processing in schizophrenia might beexplained, at least in part, by genetic vulnerability rather than thedisorder itself; however, in the present study, response accuracy orreaction times did not vary as a function of genotype and were not

associated with brain activation in the occipital regions modulated bygenotype (pN0.05).

Although the present investigation focused on visual processing,the DTNBP1 gene is expressed throughout the cortex and is likely toaffect several other cognitive functions; for instance, recent studieshave reported that the risk variant is associated with poorerperformance on tests of general cognitive function in healthyparticipants (Burdick et al., 2006; Fallgatter et al., 2006), as well aspatients with schizophrenia (Burdick et al., 2006; Burdick et al., 2007;Donohoe et al., 2007). Taken collectively, these studies provide supportfor a general role of DTNBP1 in brain function including perceptual aswell as high level processes, consistent with its putative role in theformation and pruning of synaptic connections throughout the brain(Donohoe et al., 2008). Nevertheless, our secondary analysis of theimpact of DTNBP1 across thewhole brain did not reveal any significanteffects outside the occipital cortex; this aspect of our findings may beexplained by the nature of our experimental task which requiredlimited cognitive resources for successful performance.

We note that, while Donohoe et al. reported reduced P100 amplitudeassociatedwith the high-risk haplotype, in the present investigationwefound increased activations in carriers of the risk allele. There are threepossible explanations for this apparent contradiction. First, Donohoe etal. examined patients with schizophrenia whereas our sample com-prised of clinically healthy subjects; it is possible that the effects of ourpolymorphism of interest differ for patients and controls (Frodl et al.,2004; Addington et al., 2007;Mechelli et al., 2008). Second, Donohoe etal. examined mature adults whereas participants in our investigationwere 10–12 year old; in light of growing evidence thatmaturation of thevisual system continues until 20 years of age (Breceli, 2003), it will beimportant to examine the extent towhich the effects in our child samplehold in the adult population. Third, careful examination of the literaturesuggests that the assumption that positive ERPs amplitude shouldcorrelate with positive BOLD response is problematic; although theBOLD signal has been shown to correlate with local field potentials(Logothetis et al., 2001), the relationship between more macrorecordings and the BOLD signal is not as clear. For instance, in a recentinvestigation, the P100 mapped to a negative BOLD response in theoccipital cortex (Whittingstall et al., 2007).

The present investigation has a number of potential limitationswhich require careful consideration. A first potential limitation is thatwe genotyped our subjects for rs2619538 for the reasons outlined inthe introduction. However, within the DTNBP1 gene, other SNPs havealso been associated with schizophrenia, and there is no agreement asto which is the most significant marker; the identification of differentSNPs in different studies may be explained by the presence of eitherallelic heterogeneity, or alternatively, of population differences withinthe LD structure across the gene (Bray et al., 2005). The function of thedifferent SNPs is still unclear, and it is therefore difficult to makepredictions in terms of their functional impact in the brain. A secondlimitation is that patients with schizophrenia appear to showdysfunction of the magnocellular system (which conducts low-resolution visual information and is involved in the processing ofoverall stimulus organization) to a greater extent than of theparvocellular system (which conducts high-resolution visual informa-tion and in involved in processing of fine-grained stimulus configura-tion) (Butler and Javitt, 2005). However, the experimental paradigmemployed in the present investigation was not designed to examinethese two systems independently. Future studies could use a moresensitive experimental paradigm in order to establish the extent towhich effects of DTNBP1 genotype are specific to the magnocellularand parvocellular visual systems. Finally, as our sample was restrictedto males, it is unclear whether the results can be generalized to thefemale population. Given that there are gender differences in the bothincidence and course of schizophrenia, in future studies it would beuseful to assess the extent to which the effects of DTNBP1 genotypeare related to gender.

822 A. Mechelli et al. / NeuroImage 49 (2010) 817–822

In conclusion, this is the first investigation to examine the effects ofgenetic variation in DTNBP1 on visual processing in children. Wereport that the high- and low-risk genotypes are associated withdifferences in occipital activation during a visual matching task inchildren as young as 10–12 years of age. Our investigation providessupport to the idea that genetic variation in DTNBP1 affects brainfunction to moderate vulnerability to psychopathology from child-hood, before the possible manifestation of any symptoms in lateadolescence and adulthood.

Acknowledgments

The authors want to thank the families of the Twins EarlyDevelopment Study, Prof. Robert Plomin, Dr. Alice Jones and Mrs.Patricia Busfield for their generous help with this study. This work wassupported by a grant from the Medical Research Council (G0401170)to EV.

References

Addington, A.M., Gornick, M.C., Shaw, P., Seal, J., Gogtay, N., Greenstein, D., et al., 2007.Neuregulin 1 (8p12) and childhood-onset schizophrenia: susceptibility haplotypesfor diagnosis and brain developmental trajectories. Mol. Psychiatry 12 (2),195–205.

Bitan, T., Cheon, J., Lu, D., Burman, D.D., Booth, J.R., 2009. Developmental increase in top-down and bottom-up processing in a phonological task: an effective connectivity,fMRI study. J. Cogn. Neurosci. 21 (6), 1135–1145.

Bray, N.J., Preece, A., Williams, N.M., Moskvina, V., Buckland, P.R., Owen,M.J., et al., 2005.Haplotypes at the dystrobrevin binding protein 1 (DTNBP1) gene locusmediate riskfor schizophrenia through reduced DTNBP1 expression. Hum. Mol. Genet. 14 (14),1947–1954.

Breceli, J., 2003. From immature to mature pattern ERG and VEP. Doc. Ophthalmol. 107(3), 215–224 Review.

Burdick, K.E., Lencz, T., Funke, B., Finn, C.T., Szeszko, P.R., Kane, J.M., et al., 2006. Geneticvariation in DTNBP1 influences general cognitive ability. Hum. Mol. Genet. 15 (10),1563–1568.

Burdick, K.E., Goldberg, T.E., Funke, B., Bates, J.A., Lencz, T., Kucherlapati, R., et al., 2007.DTNBP1 genotype influences cognitive decline in schizophrenia. Schizophr. Res. 89(1–3), 169–172.

Burgund, E.D., Kang, H.C., Kelly, J.E., Buckner, R.L., Snyder, A.Z., Petersen, S.E., et al., 2002.The feasibility of a common stereotactic space for children and adults in fMRIstudies of development. Neuroimage 17 (1), 184–200.

Butler, P.D., Javitt, D.C., 2005. Early-stage visual processing deficits in schizophrenia.Curr. Opin. Psychiatry 18 (2), 151–157.

Butler, P.D., Zemon, V., Schechter, I., Saperstein, A.M., Hoptman, M.J., Lim, K.O., et al.,2005. Early-stage visual processing and cortical amplification deficits in schizo-phrenia. Arch. Gen. Psychiatry 62 (5), 495–504.

Craddock, N., O'Donovan, M.C., Owen, M.J., 2005. The genetics of schizophrenia andbipolar disorder: dissecting psychosis. J. Med. Genet. 42 (3), 193–204.

Donohoe, G., Morris, D.W., Clarke, S., McGhee, K.A., Schwaiger, S., Nangle, J.M., et al.,2007. Variance in neurocognitive performance is associated with dysbindin-1 inschizophrenia: a preliminary study. Neuropsychologia 45 (2), 454–458.

Donohoe, G., Morris, D.W., De Sanctis, P., Magno, E., Montesi, J.L., Garavan, H.P., et al.,2008. Early visual processing deficits in dysbindin-associated schizophrenia. BiolPsychiatry 63 (5), 484–489.

Fallgatter, A.J., Herrmann, M.J., Hohoff, C., Ehlis, A.C., Jarczok, T.A., Freitag, C.M., et al.,2006. DTNBP1 (dysbindin) gene variants modulate prefrontal brain function inhealthy individuals. Neuropsychopharmacology 31 (9), 2002–2010.

Freeman, B., Smith, N., Curtis, C., Huckett, L., Mill, J., Craig, I.W., 2003. DNA from buccalswabs recruited by mail: evaluation of storage effects on long-term stability andsuitability for multiplex polymerase chain reaction genotyping. Behav. Genet. 33,67–72.

Friston, K.J., 2003. Introduction: experimental design and statistical parametricmapping. In: Frackowiak, R.S.J., Friston, K.J., Frith, C., Dolan, R., Friston, K.J., Price,C.J., Zeki, S., Ashburner, J., Penny, W.D. (Eds.), Human Brain Function. InAcademicPress, pp. 599–632.

Frodl, T., Meisenzahl, E.M., Zill, P., Baghai, T., Rujescu, D., Leinsinger, G., et al., 2004.Reduced hippocampal volumes associated with the long variant of the serotonintransporter polymorphism in major depression. Arch. Gen. Psychiatry 61 (2),177–183.

Goodman, R., Ford, T., Simmons, H., Gatward, R., Meltzer, H., 2000. Using the Strengthsand Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders in acommunity sample. Br. J. Psychiatry 177, 534–539.

Gornick, M.C., Addington, A.M., Sporn, A., Gogtay, N., Greenstein, D., Lenane, M., et al.,2005. Dysbindin (DTNBP1, 6p22.3) is associated with childhood-onset psychosisand endophenotypesmeasured by the Premorbid Adjustment Scale (PAS). J. Autism.Dev. Disord. 35 (6), 831–838.

Harrison, P., Weinberger, D., 2005. Schizophrenia genes, gene expression, andneuropathology: on the matter of their convergence. Mol. Psychiatry 10 (1), 40–68.

Kang, H.C., Burgund, E.D., Lugar, H.M., Petersen, S.E., Schlaggar, B.L., 2003. Comparison offunctional activation foci in children and adults using a common stereotactic space.Neuroimage 19 (1), 16–28.

Kim-Cohen, J., Caspi, A., Moffitt, T.E., Harrington, H., Milne, B.J., Poulton, R., 2003. Priorjuvenile diagnoses in adults with mental disorder: developmental follow-back of aprospective-longitudinal cohort. Arch. Gen. Psychiatry 60 (7), 709–717.

Kirov, G., Ivanov, D., Williams, N.M., Preece, A., Nikolov, I., Milev, R., et al., 2004. Strongevidence for association between the dystrobrevin binding protein 1 gene(DTNBP1) and schizophrenia in 488 parent–offspring trios from Bulgaria. Biol.Psychiatry 55 (10), 971–975.

Kishimoto, M., Ujike, H., Motohashi, Y., Tanaka, Y., Okahisa, Y., Kotaka, T., et al., 2008. Thedysbindin gene (DTNBP1) is associated with methamphetamine psychosis. Biol.Psychiatry 63 (2), 191–196.

Li, T., Zhang, F., Liu, X., Sun, X., Sham, P.C., Crombie, C., et al., 2005. Identifying potentialrisk haplotypes for schizophrenia at the DTNBP1 locus in Han Chinese and Scottishpopulations. Mol. Psychiatry 10, 1037–1044.

Logothetis, N.K., Pauls, J., Augath, M., Trinath, T., Oeltermann, A., 2001. Neurophysiolo-gical investigation of the basis of the fMRI signal. Nature 412 (6843), 128–130.

Martínez, A., Hillyard, S.A., Dias, E.C., Hagler Jr., D.J., Butler, P.D., Guilfoyle, D.N., et al.,2008. Magnocellular pathway impairment in schizophrenia: evidence fromfunctional magnetic resonance imaging. J. Neurosci. 28 (30), 7492–7500.

Mechelli, A., Henson, R.N., Price, C.J., Friston, K.J., 2003. Comparing event-related andepoch analysis in blocked design fMRI. Neuroimage 18 (3), 806–810.

Mechelli, A., Prata, D.P., Fu, C.H., Picchioni, M., Kane, F., Kalidindi, S., et al., 2008. Theeffects of neuregulin1 on brain function in controls and patients with schizophreniaand bipolar disorder. Neuroimage 42 (2), 817–826.

Numakawa, T., Yagasaki, Y., Ishimoto, T., Okada, T., Suzuki, T., Iwata, N., et al., 2004.Evidence of novel neuronal functions of dysbindin, a susceptibility gene forschizophrenia. Hum. Mol. Genet. 13 (21), 2699–2708.

Oord, E., Sullivan, P., Jiang, Y., Walsh, D., O'Neill, F., Kendler, K., et al., 2003. Identificationof a high-risk haplotype for thedystrobrevin binding protein 1 (DTNBP1) gene in theIrish study of high-density schizophrenia families. Mol. Psychiatry 8, 499–510.

Owen, M.J., Craddock, N., O'Donovan, M.C., 2005. Schizophrenia: genes at last? TrendsGenet. 21 (9), 518–525.

Owen, M.J., Craddock, N., Jablensky, A., 2007. The genetic deconstruction of psychosis.Schizophr. Bull. 33 (4), 905–911.

Penny, W.D., Holmes, A., 2003. Random effects analysis. In: Frackowiak, R.S.J., Friston, K.J.,Frith, C., Dolan, R., Friston, K.J., Price, C.J., Zeki, S., Ashburner, J., Penny, W.D. (Eds.),Human Brain Function. Academic Press, pp. 843–850.

Piskulic, D., Olver, J.S., Norman, T.R., Maruff, P., 2007. Behavioural studies of spatialworking memory dysfunction in schizophrenia: a quantitative literature review.Psychiatry Res. 150 (2), 111–121 Review.

Raybould, R., Green, E.K., MacGregor, S., Gordon-Smith, K., Heron, J., Hyde, S., et al.,2005. Bipolar disorder and polymorphisms in the dysbindin gene (DTNBP1). Biol.Psychiatry 57 (7), 696–701.

Ross, C., Margolis, R., Reading, S., Pletnikov, M., Coyle, J., 2006. Neurobiology ofSchizophrenia. Neuron 52 (1), 139–153.

Schechter, I., Butler, P.D., Silipo, G., Zemon, V., Javitt, D.C., 2003. Magnocellular andparvocellular contributions to backward masking dysfunction in schizophrenia.Schizophr. Res. 64 (2–3), 91–101.

Schwab, S.G., Knapp, M., Mondabon, S., Hallmayer, J., Borrmann-Hassenbach, M., Albus,M., et al., 2003. Support for association of schizophrenia with genetic variation inthe 6p22.3 gene, dysbindin, in sib-pair families with linkage and in an additionalsample of triad families. Am. J. Hum. Genet. 72 (1), 185–190.

Straub, R., Jiang, Y., MacLean, C., Ma, Y., Webb, B., Myakishev, M., et al., 2002. Geneticvariation in 6p22.3 gene DTNBP1, the human ortholog of themouse dysbindin gene,is associated with schizophrenia. Am. J. Hum. Genet. 71 (2), 337–348.

Talbot, K., Eidem, W., Tinsley, C., Benson, M., Thompson, E., Smith, R., et al., 2004.Dysbindin-1 is reduced in intrinsic, glutamatergic terminals of the hippocampalformation in schizophrenia. J. Clin. Invest. 113 (9), 1353.

Tang, J., Zhou, J., Li, X., Shi, Y., Gu, N., Feng, G., et al., 2003. Family-based association studyof DTNBP1 in 6p22.3 and schizophrenia. Mol. Psychiatry 8 (12), 717–718.

Tochigi, M., Zhang, X., Ohashi, J., Hibino, H., Otowa, T., Rogers, M., et al., 2006.Association study of the dysbindin (DTNBP1) gene in schizophrenia from theJapanese population. Neurosci. Res. 56 (2), 154–158.

Tosato, S., Ruggeri, M., Bonetto, C., Bertani, M., Marrella, G., Lasalvia, A., et al., 2007.Association study of dysbindin gene with clinical and outcome measures in arepresentative cohort of Italian schizophrenic patients. Am. J. Med. Genet. BNeuropsychiatr. Genet. 144B (5), 647–659.

Van Den Bogaert, A., Schumacher, J., Schulze, T.G., Otte, A.C., Ohlraun, S., Kovalenko, S., etal., 2003. The DTNPB1 (dysbindin) gene contributes to schizophrenia, dependingon family history of the disease. Am. J. Hum. Genet. 73 (6), 1438–1443.

Viding, E., Williamson, D.E., Hariri, A.R., 2006. Developmental imaging genetics:challenges and promises for translational research. Dev. Psychopathol. 18 (3),877–892 Review.

Wechsler, D., 1999. Wechsler Abbreviated Scale of Intelligence. PsychologicalCorporation.

Weinberger, D.R., 1987. Implications of normal brain development for the pathogenesisof schizophrenia. Arch. Gen. Psychiatry 44 (7), 660–669.

Whittingstall, K., Stroink, G., Schmidt, M., 2007. Evaluating the spatial relationship ofevent-related potential and functional MRI sources in the primary visual cortex.Hum. Brain Mapp. 28 (2), 134–142.

Williams, N.M., Preece, A., Morris, D.W., Spurlock, G., Bray, N.J., Stephens, M., et al., 2004.Identification in 2 independent samples of a novel schizophrenia risk haplotype ofthe dystrobrevin binding protein gene (DTNBP1). Arch. Gen. Psychiatry 61 (4),336–344.