B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te /b ra in res

Research Report

Identification of first candidate genes for creativity:A pilot study

Martin Reuter⁎, Sarah Roth, Kati Holve, Jürgen HennigDepartment of Psychology, Justus-Liebig-University of Giessen, Otto-Behaghel-Str. 10F, D-35394 Giessen, Germany

A R T I C L E I N F O

⁎ Corresponding author. Fax: +49 641 9926159E-mail address: martin.reuter@psychol.un

0006-8993/$ – see front matter © 2005 Elsevidoi:10.1016/j.brainres.2005.11.046

A B S T R A C T

Article history:Accepted 16 November 2005Available online 3 January 2006

Studies from behavioral genetics have demonstrated the high heritability of intelligence.However, the endeavor to detect the genes forming the molecular basis of intelligence hasbeen rather unsuccessful until now. Pharmacological studies have demonstrated theinfluence of the dopaminergic (DA) and the serotonergic (5-HT) system on subcomponentsof cognitive functioning, and first studies from molecular genetics have demonstrated thatgenes related to the DA metabolism are associated with mental abilities. However,candidate genes for creativity have not been identified so far. Therefore, the influence ofthe catechol-O-methyltransferase (locus: COMT VAL158MET) gene and the dopamine D2receptor gene (locus: DRD2 TAQ IA) on creativity was tested in addition to a serotonergicgene, TPH1 (locus: TPH-A779C), in a sample of N = 92 healthy Caucasian subjects whilecontrolling for intelligence. Results showed that the DRD2 gene and the TPH gene were bothassociated with total creativity, explaining 9% of the variance, while COMT was not relatedto creativity at all. With respect to the subcomponents, the A1+ allele of DRD2 was related tohigher verbal creativity as compared to the A1− allele, and carriers of the A allele of TPH1showed significantly higher scores in figural and in numeric creativity, indicating that thetwo gene loci discriminate between higher cortical functions according to the organizationof cognitive functions in the respective hemispheres.

© 2005 Elsevier B.V. All rights reserved.

Keywords:CreativityCognitive functionDRD2 TAQ IACOMT VAL158METTPH1 A779C

1. Introduction

Intelligence is one of the most frequently investigatedpsychological constructs due to its tremendous implicationsfor academic achievement and success in daily life (Plominand Thompson, 1993). Actually, it seems that intelligence andacademic achievement share a large common genetic influ-ence (Wainwright et al., 2005). General cognitive ability(intelligence, often indexed by IQ scores) is one of the mosthighly heritable behavioral phenotypes. After the high herita-bility of intelligence had been demonstrated by twin and

.i-giessen.de (M. Reuter).

er B.V. All rights reserved

adoption studies (e.g., Bouchard et al., 1990), the search forthese genes forming the molecular basis of heritability hasstarted (Plomin et al., 2001). Due to the fact that intelligence isa phenotype that is normally distributed in the population, itis evident that not a single gene locus but many genes, socalled quantitative trait loci (QTLs), code for intelligence(Plomin et al., 1994).

In comparison to intelligence, research on creativity israther orphaned. According to Sternberg et al. (1988), creativityresearch has “taken on the role of a prodigal stepbrother toresearch on intelligence”. Obviously, the neglect of creativity

.

191B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

research is mainly due to methodological problems con-cerning reliably and validity of creativity measures and alsodue to disagreements in the definition of creativity. Someresearchers preferred the product, others the process criterionto investigate creativity (Brown, 1989). The process approachassumes that creativity is a trait that is normally distributed inthe general population, whereas the product approach definescreativity in terms of exceptional real-world creative produc-tion, which very few individuals manage to achieve (Stavridouand Furnham, 1996). Therefore, supporters of the process/traitapproach-like Guilford (Guilford, 1956) who followed theassumption that components of divergent thinking, namely,originality, flexibility, and elaboration, are at the core ofcreativity and that these components are also manifest in anormal population, attempted to discover the underlyingprocesses of creativity and tried to develop appropriateinstruments to measure it.

In some intelligence models, creativity and divergentthinking are facets of intelligence (Jäger et al., 1982; Guilford,1967, 1979). Therefore, creativity should also have a stronggenetic basis and should be – in a similar way as intelligence– normally distributed in the population. This assumption isin line with the process/trait approach. However, little isknown about the genetic basis of creativity and its heritabil-ity. A review of ten twin studies of creativity originating fromthe 1970s yielded an average correlation of 0.61 for identicaltwins and 0.50 for fraternal twins (Canter et al., 1973).According to some work, this genetic influence of about20% is primarily due to the correlation between creativity andIQ. When IQ is controlled, identical and fraternal twincorrelations for creativity tests were scarcely different(Nichols, 1978). However, the problem of some of thesestudies was the reliable measurement of creativity and thecomparability across studies.

Furthermore, until now, there is little known about thebiological underpinnings of creativity and the neuroanatom-ical correlates. Findings reporting an association betweencreativity and testosterone levels could not be confirmed byothers (Hassler, 1992; Reuter et al., 2005b). Instead, thepersonality trait of SEEK of the Affective Neurosience PersonalityScales (ANPS; Davis et al., 2003) turned out to be related tocreativity in men as well as in women and to account for morethan 15% of the variance (Reuter et al., 2005b). These resultspersisted even after correcting for intelligence. The SEEKdimension is an interesting trait for creativity researchbecause on the one hand it is conceptualized on a strongbiological basis and on the other hand it explicitly assessesaspects of creativity like eagerness to solve problems andfavoring activities related to exploring new things. Theneuroanatomical structures representing the SEEK systemsare predominantly dopaminergic pathways (Panksepp et al.,1998). Especially mesocortical dopamine (DA) projections tothe forebrain are known to be involved in cognitive function-ing and therefore can be assumed to be involved in creativethinking as well. Therefore, dopaminergic genes are not onlylikely to be the genetic basis of SEEK but also of creativityindicating pleiotropic effects. Pleiotropy refers to the phe-nomenon that a certain gene is influencing different pheno-types at the same time. Indeed, in a recent association study inN = 238 healthy Caucasian subjects of German origin, we have

found an association between the DRD2 TAQ IA gene and SEEK(Reuter et al., submitted for publication).

DA-related genes have turned out to be interesting candi-date genes for cognitive functioning. Especially the catechol-O-methyltransferase (COMT) gene has been successfully relatedto prefrontal executive functioning encompassing cognitiveprocesses likeattentionandworkingmemory (Eganet al., 2001;Malhotra et al., 2002; Joober et al., 2002; Goldberg et al., 2003).COMT is an enzyme which plays a crucial role in themetabolism of catecholamines by inactivating them in thesynaptic cleft. A single nucleotide polymorphism (SNP), aG→ A transition in codon 158 of the COMT gene located at theq11 band of human chromosome 22 (GenBank accession no.AY341246), results in 3- to 4-fold difference in COMT enzymeactivity (Lachman et al., 1996) by coding for the synthesis of theamino acid methionine (MET) instead of valine (VAL). Hetero-zygotes (VAL/METgenotype) have intermediate levels of COMTactivity (Lachman et al., 1996; Syvänen et al., 1997). Egan et al.(2001) reported better working memory performance incarriers of the MET/MET genotype. Using an n-back task,Goldberg et al. (2003) also found the best working memoryperformance in carriers of the MET/MET genotype and worstperformance in carriers with the VAL/VAL genotype. In linewith these findings, Malhotra et al. (2002) found that subjectswith only the low-activity MET allele made significantly fewerperseverative errors on theWisconsinCardSortingTest,whichis a task of prefrontal cognition, than did subjectswith the VALallele. This finding could be corroborated in healthy controlsand in schizophrenic patients (Joober et al., 2002). Interestingly,Mattay et al. (2003) reported better results in carriers of theVALallele and a decline in the MET/MET group in a workingmemory task after a pharmacological challenge with the DA-agonistic substance amphetamine. The decline in perfor-mance in the MET/MET group after amphetamine intakesuggests that the association between performance and DAlevels has an inverted ‘U’ shape characteristic, i.e. activation ofthe DA system by working memory load and amphetamine,pushing these subjects beyond their optimal activation level.

Another interesting candidate gene for cognitive ability isthe dopamine D2 receptor gene (DRD2), and one of itspolymorphic regions has been tested in several studies withrespect to cognitive functioning. The DRD2 TAQ I A polymor-phism is a restriction fragment polymorphism (RFLP) onchromosome 11 at q22–q23which is also caused by amutationin a single nucleotide (GenBank accession no. AF050737). Theprevalence of the mutated A1 allele is 28% and that of theA1A1 genotype is about 3% in healthy Caucasians (Noble,2000). Due to the small prevalence of the A1A1 genotype,carriers of the A1 allele are often contrasted with carrierswithout the A1 allele by classifying the A1A1 and the A1A2carriers as A1+ and the carriers of the A2A2 genotype as A1−.Although the DRD2 TAQ IA polymorphism is located in the 3′untranslated region of the DRD2 gene, it seems to havefunctional consequences resulting either from linkage dis-equilibrium with another functional DRD2 variant or frombeing located in an as yet unidentified coding or regulatoryregion downstream of DRD2. In individuals with the A1 allele,a 30–40% reduction in D2 dopamine receptor density could bedemonstrated as compared to those homozygous for the A2allele (Ritchie and Noble, 2003), a result which could be

Table 1 – Genotype and allele frequencies of the COMTVAL158MET, the DRD2 TAQ IA, and the TPH A779Cpolymorphisms

COMT VAL158MET DRD2 TAQ IA TPH A779C

VALVAL 22 (23.9%) A1A1 4 (4.3%) AA 13 (15.1%)VALMET 49 (53.3%) A1A2 28 (30.4%) AC 41 (47.7%)METMET 21 (22.8%) A2A2 60 (65.2%) CC 32 (37.2%)VAL 50.5% A1 19.6% A 39%MET 49.5% A2 80.4% C 61%

192 B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

confirmed by others (Thompson et al., 1997; Pohjalainen et al.,1998; Jonsson et al., 1999; Ritchie and Noble, 1996, 2003).

Berman and Noble (1995) reported an association betweenthe DRD2 TAQ IA gene and visuospatial performance which isa subcomponent of general cognitive ability in a sample ofCaucasian children, and Tsai et al. (2002) found an associationbetween the DRD2 TAQ IA gene and intelligence in an adultfemale Taiwanese sample. However, the results were contra-dictory: the visuospatial performance was significantly re-duced in childrenwith theminor A1 allele, whereas carriers ofthe A1A1 genotype in the study of Tsai et al. had a significantlyhigher IQ than carriers of the A2A2 genotype. The latterfinding was not corroborated by an earlier study in N = 137children, reporting no association between DRD2 TAQ IA andIQ (Petrill et al., 1997).

Furthermore, the serotonergic (5-HT) system seems to playan important role in cognitive functioning. Especially clinicalfindings show the beneficial effects of substances increasingcentral 5-HT activity. For example, Harvey (2003) reported thatthe combined dopamine and serotonin receptor antagonistZiprasidone significantly improves multiple cognitivedomains such as episodic memory, attention/vigilance, exec-utive function, and visuomotor speed in schizophrenicpatients. In the same line, Poyurovsky et al. (2003) showed ina double-blind placebo controlled study in a sample ofschizophrenic patients that the 5-HT2a antagonist mianserinimproved performance in neurocognitive tasks. The resultsare interpreted to demonstrate that the preponderance of 5-HT2a antagonism over dopamine D2 blockade exerted byatypical antipsychotics may contribute to their cognitive-enhancing effects. Cuccaro et al. (1993) reported a significantlynegative correlation betweenwhole blood 5-HT and cognitive–intellectual abilities in a sample of autistic patients and theirfirst-degree relatives. A further elegant method to investigatethe involvement of 5-HT in cognitive functioning, tryptophandepletion, was applied in animal studies and studies onhealthy humans. Results unequivocally showed an impair-ment in cognitive functioning after tryptophan depletion(Mazer et al., 1997; Evers et al., 2005). All these findingsindicate a possible role of the 5-HT system in cognitivefunctioning, although there is abundant evidence for a moresalient role of the DA system. It is possible that some of the 5-HT effects on cognitive functioning are mediated via anindirect pathway by influencing the activity of the DA system.It is known that the 5-HT system exerts an inhibitory effect onDA release (e.g., Porras et al., 2002).

Therefore, the aim of the present study was to investigatethree gene loci on candidate genes for creativity. Two of them,the COMT VAL158MET and the DRD2 TAQ IA SNPs, and theirassociations to cognitive functioning have already beenpresented. The third one is the A779C polymorphism on thetryptophan hydroxylase gene 1 (TPH1) (GenBank accession no.AC005728). TPH is the rate-limiting biosynthetic enzyme in theserotonin pathway and regulates levels of 5-HT by convertingtryptophan into 5-hydroxytryptophan, which is the directprecursor of 5-HT. It is conceivable that variations in the TPHgene could contribute to low activity of the 5-HT system. Giventhe fact that tryptophan depletion has been shown to causereliable alterations in cognitive functioning, it is very likelythat also the TPH gene is related to individual differences in

creativity. The A779C SNP is located on the seventh intron ofthe tryptophan hydroxylase gene-1 on chromosome 11p15.3–p14. Since the A779C SNP is located in a non-coding region andbecause no exon skipping or alternative splicing have beenreported, it is assumed that this polymorphism is not anetiological mutation but may be in linkage disequilibrium (co-occurrence of two DNA markers at a higher frequency thanwould be predicted by random chance) with an as yetunidentified polymorphism in the TPH1 gene (Han et al.,1999; Rotondo et al., 1999). However, numerous positive resultsin association studies indicate that the A779C SNP is indeed agood marker for many phenotypes, although the SNP is notetiological. The TPH1 gene has been found to be related tonicotine addiction (Sullivan et al., 2001; Lerman et al., 2001;Reuter and Hennig, 2005) as well as to aggressive personalitytraits (Reuter and Hennig, 2005; Rujescu et al., 2002; Hennig etal., 2005) and behaviors including suicide which reflects anextreme form of auto-aggression (Nielsen et al., 1994; Mann etal., 1997; Abbar et al., 2001; Souery et al., 2001).

Recently, Walther et al. (2003) identified a second TPHisoform – referred to as TPH2 – in mice, which is predom-inantly expressed in the brain stem, while the classical TPHgene – now called TPH1 – is expressed in the gut, pineal gland,spleen, and thymus. The authors also identified a TPH2homolog on human chromosome 12 (GenBank accession no.AY098914). In the meantime, SNPs on TPH2 have beendetected (e.g., Harvey et al., 2004) but unfortunately aftercompletion of the present study so that this promisingcandidate gene could not be considered.

2. Results

The DRD2 TAQ IA and the COMT VAL158MET polymorphismhad a completion rate of 100%, yet the TPH A779C assay wasless effective, resulting in a drop out of 6 out of 92 subjects(completion rate 93.5%), who could not be genotyped withrespect to this SNP. The genotype and allele frequencies arereported in Table 1. All three gene loci were in Hardy–Weinberg Equilibrium and in linkage equilibrium.

Total creativity was not significantly correlated withmeasures of intelligence (correlations ranging from 0.04 to0.16). Numeric creativity, however, was modestly correlatedwith all measures of intelligence (correlations ranging from0.21 to 0.27) except with verbal crystallized intelligence.

Results of the genetic data showed no effects of the COMTSNP on creativity measures. However, DRD2 TAQ IA wasassociated with verbal creativity (F(2,88) = 4.73, P = 0.027,

Table 2 – ANCOVA results including descriptive statistics

N CR_FIG CR_NUM CR_VERB CR_TOT

COMTVALVAL 22 −0.02 (0.40) 0.00 (0.35) −0.27 (0.28) −0.26 (0.98)VALMET 49 0.02 (0.27) −0.12 (0.23) 0.09 (0.25) −0.03 (0.66)METMET 21 −0.02 (0.42) 0.28 (0.36) 0.08 (0.39) 0.34 (1.02)COMT F(2,87) = 0.01 F(2,87) = 0.43 F(2,87) = 0.35 F(2,89) = 0.09

P = 0.994 P = 0.652 P = 0.708 P = 0.915

DRD2A1A1 4 −0.36 (0.93) 0.26 (0.80) 0.02 (0.86) −0.01 (2.23)A1A2 28 −0.44 (0.36) −0.54 (0.30) −0.61 (0.33) −1.64 (0.85)A2A2 60 0.23 (0.24) 0.23 (0.21) 0.29 (0.22) 0.77 (0.58)DRD2 F(2,87) = 1.27 F(2,87) = 2.28 F(2,87) = 2.59 F(2,87) = 2.73

P = 0.285 P = 0.108 P = 0.081 P = 0.071A1+ 32 0.23 (0.24) 0.23 (0.21) 0.28 (0.22) 0.76 (0.57)A1− 60 −0.43 (0.33) −0.44 (0.28) −0.53 (0.30) −1.43 (79)A1 Allele F(1,88) = 2.57 F(1,88) = 3.70 F(1,88) = 4.73 F(1,88) = 5.03

P = 0.113 P = 0.058 P = 0.032 P = 0.027

TPH1AA 13 0.16 (0.52) 0.58 (0.44) 0.24 (0.49) 0.95 (1.23)AC 41 0.51 (0.29) 0.24 (0.25) 0.36 (0.27) 1.14 (0.69)CC 32 −0.56 (0.33) −0.43 (0.28) −0.41 (0.31) −1.42 (0.78)TPH1 F(2,81) = 3.00 F(2,81) = 2.48 F(2,81) = 1.77 F(2,81) = 3.25

P = 0.055 P = 0.090 P = 0.177 P = 0.044CC 32 −0.56 (0.33) −0.43 (0.28) −0.41 (0.31) −1.43 (0.78)AC, AA 54 0.42 (0.25) 0.324 (0.22) 0.33 (0.24) 1.09 (0.60)A allele F(1,82) = 5.71 F(1,82) = 4.54 F(1,82) = 3.54 F(1,82) = 6.57

P = 0.019 P = 0.036 P = 0.064 P = 0.012

Means (SEM) z scores or composites of z scores.

193B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

eta2 = 0.051) and the total creativity score (F(1,88) = 5.03, P = 0.027,eta2 = 0.054), indicating higher creativity scores in carriers ofthe A1 allele. The TPH A779 allele was significantly associatedwith figural (F(1,82) = 5.71, P = 0.019, eta2 = 0.065), numeric(F(1,82) = 4.5 P = 0.036, eta2 = 0.052), and total (F(1,82) = 6.57,P = 0.012, eta2 = 0.074) creativity and also showed a trendtoward an association with verbal creativity (F(1,82) = 3.54,P = 0.064, eta2 = 0.041) (see Table 2). Entering intelligence as acovariate did not change the proportion of explained variancein creativity in any of the analyses. Measures of intelligencewere associated with none of the three gene loci.

3. Discussion

Based on previous findings which have demonstrated the roleof the dopaminergic and the serotonergic system for cognitivefunctioning, the aim of the present study was to test forpossible associations between three polymorphisms related tothese neurotransmitter systems and creativity. Althoughthere exist many studies investigating the molecular basis ofgeneral ability (IQ, e.g., Plomin et al., 1994, 2001), subcompo-nents of intelligence, and cognitive functioning (e.g., Plomin etal., 2001; Egan et al., 2001; Berman and Noble, 1995), there areto our knowledge no reports on candidate genes for creativityavailable so far. Obviously, it is not a lack of importance whichcauses the disregard of creativity research but rather thedifficulty to define properly what creativity really is and mostimportantly how it can be adequately measured (Brown, 1989;

Reuter et al., 2005b). However, it is beyond question that mostnotably divergent thinking and originality, which are at thecore of creativity, are the prerequisites for new inventions,innovative creation, and technical progress.

Association studies in the field of cognitive functioninghave yielded far more evidence for the involvement of the DAsystem than for the 5-HT system. Especially the COMT genehas been reported to be related to basal cognitive processeslike, for example, attention and working memory (Egan et al.,2001; Malhotra et al., 2002; Joober et al., 2002; Goldberg et al.,2003) in healthy subjects as well as in schizophrenic andautistic patients (Bellgrove et al., 2005). Therefore, the COMTgene was also of interest to test for associations withcreativity. In the same line, we investigated another dopami-nergic gene locus, the DRD2 TAQ IA gene, which had beensuccessfully related to visuospatial performance (Berman andNoble, 1995) and intelligence (Tsai et al., 2002) in some but notall studies (Petrill et al., 1997). Inspired by clinical findingsdemonstrating the beneficial effects of substances increasingcentral 5-HT activity, we intended to include also a seroto-nergic marker gene into our studies on creativity. Due tofindings reporting deteriorating effects on cognitive function-ing after tryptophan depletion, associations between the TPHA779C SNP and creativity were also tested.

In the absence of any hint on the involvement of COMT,results show a significant association between total creativityand the DRD2 and the TPH1 gene. Higher creativity scoreswere observed in carriers of the A1 allele of DRD2 TAQ IA andin carriers of the A allele of the TPH1 A779C. A closer look atthe subcomponents of creativity revealed that the DRD2 TAQIA SNP was mainly associated with verbal creativity and theTPH A779C SNP with figural and numeric creativity. From aneuroanatomical perspective, it might be concluded that theDRD2 and the TPH1 gene are specific for hemisphere-specificcognitive processes: DRD2 promoting predominantly creativeprocesses in the left hemisphere (where verbal information ispredominantly processed) and TPH1 stimulating innovativeprocesses in the right hemisphere (where figural and numerictasks are processed). This hemisphere-related hypothesis is ofcourse highly speculative and only one possible explanationfor the effect. The domain-specific contribution of these twogene loci to creativity is corroborated by the results of a two-factorial ANCOVA after listwise deletion (N = 86), showing thatboth SNPs explain independent proportions of variance intotal creativity, accounting together for about 9% of thevariance (DRD2 A1: eta2 = 0.034, TPH A779: eta2 = 0.055, DRD2A1 × TPH A779: eta2 = 0.009).

The negative findings with respect to the COMT gene are incontradiction to previous findings relating the COMT gene tocognitive functioning. Further studies have to find out if theeffects of the COMT gene are restricted to basal cognitiveabilities and do not account for higher cognitive abilitiesrelated to creativity as indicated by our study. The positiveassociation between the A1+ allele of DRD2 TAQ IA andcreativity is in line with previous findings reporting anassociation between the personality dimension of SEEK(encompassing items of creativity self-report) and the A1+allele (Reuter et al., submitted for publication) especiallybecause SEEK has turned out to be associated with creativityin another study (Reuter et al., 2005b).

194 B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

Surprisingly, neither the TPH1 gene nor the DRD2 gene thatwas related to measures of creativity or the COMT geneshowed associations with intelligence. However, it turned outthat the measures of intelligence and creativity used in thepresent study were only moderately correlated. Therefore, thetwo genes seem to account for proportions of variance increativity that are not in common with intelligence.

Although the present pilot study identifies two promisingcandidate genes for creativity, the findings are preliminaryand have to be confirmed in independent samples of greatersize. A conservative Bonferroni correction for multiple testingwould have the effect that the significant results would turnout to be no longer significant. However, the effect sizes weresubstantial, and the absence of smaller P values was aconsequence of the small sample size related to low power.

Our sample exclusively consisted of university students,with an average IQ of 115 (as measured by the CFT3) whichconfines the results to subjects with an IQ above average.Although intelligence did not account for the associationsfound between the two gene loci and creativity, intelligencemight be a crucial factor. According to Guilford's thresholdmodel of creativity (Guilford, 1956), intelligence is a neces-sary but not sufficient prerequisite for creativity, andtherefore the probability of creative thinking to occur ishigher in university students than in an unselected sample.Moreover, the relation between the TPH1 gene and creativitycannot explain the exact genetic mechanisms underlyingthis association because the TPH A779C SNP is located on anon-coding region of the TPH1 gene that only expresses TPHin peripheral tissues. Therefore, the A779C SNP can only beconsidered as a marker of creativity likely to be in linkagewith an as yet unidentified functional gene locus than to beinvolved in the expression of creative behavior by itself. Atthe time when the present study was designed, the humanTPH2 gene involved in the metabolism of central 5-HT wasnot identified (Walther et al., 2003). Given the encouragingpharmacological studies reporting changes in cognitiveperformance after TPH depletion and our own findingsrelating TPH1 to creativity, the TPH2 gene is an interestingcandidate gene for creativity research. However, presentresults do not only stress the importance of the DA systembut also of the 5-HT system for higher cognitive functions. Ithas to be elucidated in further research if the effect of 5-HTis exerted directly or via an interaction with the DA system.The absence of an interaction effect DRD2 TAQ IA × TPH1A779C does not exclude an interaction of the DA and the 5-HT systems at another metabolic site. The use of ahaplotype approach in further studies will help to identifyadditional functional gene regions related to creativity.

The few studies on the heritability of creativity are ratherold and showed a rather small heritability of creativity (about20%) (Canter et al., 1973). Even it was questioned if creativity isheritable at all when intelligence is controlled (Nichols, 1978).We cannot deny this assumption until our data get supportfrom other studies. However, our data show that the BIS scalesare measures of creativity that have a very good reliabilityα = 0.85, are very modestly correlated with intelligence, andcan be associated with gene markers that were not related tomeasures of intelligence. The probability to have identifiedfalse positive candidate genes for creativity cannot be ruled

out based on the data of a pilot study in which the smallsample size increases the chance for a stratification bias. But,themeasures of creativity as well as the genemarkers seem tobe stimulating for further research indebted to contribute tothe exploration of the biological basis of creativity.

4. Experimental procedures

4.1. Sample

The sample consisted of N = 92 healthy unrelated Caucasianuniversity students of German origin (17 males and 75 females,age: M = 22.60, SD = 4.06) who participated on a voluntary basis inthe study after having given informed consent. All subjects wereundergraduate students who were not familiar with the achieve-ment tests applied. The imbalance between numbers ofmales andfemales was caused by a respective imbalance among psychologystudents at the university. The study was in accordance with theethical guidelines of theWorldMedical Association (Declaration ofHelsinki).

4.2. Measures of creativity

Creativity was assessed by the test battery “inventiveness” of the“Berlin Intelligence Structure Test” (BIS, Jäger et al., 1997). Asmentioned above, the BIS defines creativity explicitly as asubcomponent of intelligence. The inventiveness tests measurethe components figural, verbal, and numeric creativity andchallenges the flexible production of ideas, the power of imagi-nation, and the skill to consider many possible ways and solutionsfor solving a problem. Two tests for each component of creativitywere administered: ZF and ZK for figural creativity, MA and AM forverbal creativity, and TN and ZR for numeric creativity. Given thepsychometric differences between the subscales, z scores werecomputed for the calculation of composite scales of figural, verbal,numeric, and total creativity.

ZF: Completing a set of symbols: create as many possible realobjects from a single simple figure (circles, triangles etc.).ZK: Combining symbols: compose as many possible differentintegrated figures from four single figures as you can.MA: Masselon: Form as many different sentences as you canwhich contain three given nouns (resembles Guilford's DSS-test).AM: Possibilities for using objects: name as many possibilitiesas you can how a given object can be used (e.g., a brick…can beused to build a house, to build a shelf, or to demonstratekarate).TN: Telephone numbers: create as many phone numbers withfour digits as you can which each follows a different rule oforder or construction (e.g., 8282 or 1771).ZR: Number-quiz: invent patterns of numbers which follow asystematic principle and overlap in a way that always twopatterns share one common number.

The reliability of the total creativity scale, consisting of all sixsubtests, was α = 0.85.

4.3. Measures of intelligence

Since it is known that intelligence and creativity are highlycorrelated (up to r = 0.50, Cropley, 1966), it was necessary toexclude that possible significant associations between creativityand polymorphismsmightmerely reflect a relation to intelligence.

195B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

Therefore, measurement of intelligence was also requested.Given that the score of the inventiveness scale is included intothe calculation of the IQ in Jäger's BIS, it was decided to chooseother intelligence tests than the BIS to avoid a pars pro totocorrelation, that is, because the inventiveness scales measuringcreativity are part of the BIS, it is very likely that thesecreativity scales are highly correlated with the other subcom-ponents of the BIS measuring intelligence. Therefore, fluidintelligence was assessed by Cattell's CFT-3 (Cattell et al.,1971) and crystallized intelligence by the subtest knowledge ofthe Structure-of-Intelligence-Test (Intelligenz-Struktur-Test, IST2000R; Amthauer et al., 2000). The knowledge tests of the IST2000R assess figural, numeric, and verbal crystallized intelli-gence and are therefore constructed in analogy to the inven-tiveness battery of the BIS which also comprises all threecognitive components. The CFT-3, on the other hand, does notcontain verbal test material. Nevertheless, the CFT-3 waspreferred to the figural test batteries of the IST 2000R measuringverbal, numeric, and figural fluid intelligence because of itslower time consumption.

4.4. Genetic analyses

DNAwas extracted from buccal cells to avoid a selective exclusionof subjects with blood and injection phobias. Purification ofgenomic DNA was performed with a standard commercialextraction kit (High Pure PCR Template Preparation Kit; RocheDiagnostics, Mannheim, Germany). Genotyping of the three singlenucleotide polymorphisms (SNPs) was performed by real-timePCR using fluorescencemelting curve detection analysis bymeansof the Light Cycler System (Roche Diagnostics, Mannheim,Germany). Details of the PCR protocols were described elsewhere(Reuter and Hennig, 2005; Reuter et al., 2005a). The primers andhybridization probes used (TIB MOLBIOL, Berlin, Germany) were asfollows:

For COMT VAL158MET:forward primer: 5′-GGGCCTACTGTGGCTACTCA-3′;reverse primer: 5′-GGCCCTTTTTCCAGGTCTG-3′;anchor hybridization probe: 5′-LCRed640-TGTGCATGC-

CTGACCCGTTGTCA-phosphate-3′;sensor hybridization probe: 5′-ATTTCGCTGGCATGAAGGA-

CAAG-fluorescein-3′.For DRD2 TAQ IA:

forward primer: 5′-CGGCTGGCCAAGTTGTCTAA-3′;reverse primer: 5′-CAAATGTCCACGCCCGCA-3′;anchor hybridization probe: 5′-LCRed640-TGAGGAT-

GGCTGTGTTGCCCTT-phosphate-3′;sensor hybridization probe: 5′-CTGCCTCGACCAGCACT-

fluorescein-3′.

For TPH A779C:forward primer: 5′-CTTATATGTGTGAGTCTGAGTGG-3′;reverse primer: 5′-GGACATGACCTAAGAGTTCATGGCA-3′;anchor hybridization probe: 5′-LCRed640-CACGCTG-

CAGTGCTTAACATACGTTTATAA-phosphate-3′;sensor hybridization probe: 5′-CTGAAAGAGAGGTA-CAAGTT-fluorescein-3′.

4.5. Statistical methods

The six subscales of the test battery inventiveness were transformedinto z scores and the two corresponding subscales for each creativitydimension (CR_FIG = figural creativity, CR_NUM = numeric creativity,CR_VERB = verbal creativity) were added up and served as dependentvariables. The sum of the z scores of all six subscales resulted in the

total creativity score (CR_TOT). The influence of the gene loci oncreativity was tested by ANCOVA with fluid intelligence (CFT-3) as acovariate to test if effects of the gene loci on creativity were due toconfounding by intelligence. For each of the four dependent variablesfigural, numeric, verbal, and total creativity, separate ANCOVAanalyses were computed. Moreover, the components of creativitywere additionally controlled for crystallized intelligence by therespective subtests of the IST 2000R, for example, verbal crystallizedintelligence together with fluid intelligence was entered as covariatesinto an analysis when the effect of a certain gene locus on verbalcreativity was tested etc. Sex × gene interactions could not be testeddue to the small proportion of males in the sample.

R E F E R E N C E S

Abbar, M., Courtet, P., Bellivier, F., Leboyer, M., Boulenger, J.P.,Castelhau, D., Ferreira, M., Lambercy, C., Mouthon, D.,Paoloni-Giacobino, A., Vessaz, M., Malafosse, A., Buresi, C.,2001. Suicide attempts and the tryptophan hydroxylase gene.Mol. Psychiatry 6, 268–273.

Amthauer, R., Brocke, B., Liepmann, D., Beauducel, A., 2001.Intelligenz-Struktur-Test 2000R. Hogrefe, Göttingen.

Bellgrove, M.A., Domschke, K., Hawi, Z., Kirley, A., Mullins, C.,Robertson, I.H., Gill, M., 2005. The methionine allele of theCOMT polymorphism impairs prefrontal cognition inchildren and adolescents with ADHD. Exp. Brain Res. 163,352–360.

Berman, S.M., Noble, E.P., 1995. Reduced visuospatial performancein children with the D2 dopamine receptor A1 allele. Behav.Genet. 25, 45–58.

Bouchard Jr., T.J., Lykken, D.T., McGue, M., Segal, N.L., Tellegen, A.,1990. Sources of human psychological differences: theMinnesota Study of Twins Reared Apart. Science 250, 223–228.

Brown, R.T., 1989. Creativity: what are we tomeasure? In: Glover, J.A., Ronning, R.R., Reynolds, C.R. (Eds.), Handbook of Creativity.Cambridge Univ. Press, Cambridge.

Canter, S., 1973. Personality traits in twins. In: Claridge, G., Canter,S., Hume, W.I. (Eds.), Personality Differences and BiologicalVariations. Pergamon Press, New York, pp. 21–51.

Cattell, R.B., Weiss, R., 1971. Grundintelligenztest CFT3 Skala 3.Westermann, Braunschweig.

Cropley, A.J., 1966. Creativity and intelligence. Br. J. Educ. Psychol.36, 259–266.

Cuccaro, M.L., Wright, H.H., Abramson, R.K., Marsteller, F.A.,Valentine, J., 1993. Whole-blood serotonin and cognitivefunctioning in autistic individuals and their first-degreerelatives. J. Neuropsychiatry Clin. Neurosci. 5, 94–101.

Davis, K.L., Panksepp, J., Normansell, L., 2003. The affectiveneuroscience personality scales: normative data andimplications. Neuro-Psychoanalysis 5, 57–69.

Egan, M.F., Goldberg, T.E., Kolachana, B.S., Callicott, J.H., Mazzanti,C.M., Straub, R.E., Goldman, D., Weinberger, D.R., 2001. Effect ofCOMT Val108/158 Met genotype on frontal lobe function andrisk for schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 98,6917–6922.

Evers, E.A., Tillie, D.E., van der Veen, F.M., Lieben, C.K., Jolles, J.,Deutz, N.E., Schmitt, J.A., 2005. Effects of a novel method ofacute tryptophan depletion on plasma tryptophan andcognitive performance in healthy volunteers.Psychopharmacology 178, 92–99.

Goldberg, T.E., Egan, M.F., Gscheidle, T., Coppola, R., Weickert, T.,Kolachana, B.S., Goldman, D., Weinberger, D.R., 2003. Executivesubprocesses in working memory: relationship tocatechol-O-methyltransferase Val158Met genotype andschizophrenia. Arch. Gen. Psychiatry 60, 889–896.

Guilford, J.P., 1956. The structure of intellect. Psychol. Bull. 53,267–293.

196 B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

Guilford, J.P., 1967. The Nature of Human Intelligence.MacGraw-Hill, New York.

Guilford, J.P., 1979. Some incubated thoughts on incubation. J.Creat. Behav. 13, 1–8.

Han, L., Nielsen, D.A., Rosenthal, N.E., Jefferson, K., Kaye, W.,Murph, D., Altemus, M., Humphries, J., Cassano, G., Rotondo, A.,Virkkunen, M., Linnoila, M., Goldman, D., 1999. No codingvariant of the tryptophan hydroxylase gene detected inseasonal affective disorder, obsessive-compulsive disorder,anorexia nervosa, and alcoholism. Biol. Psychiatry 45, 615–619.

Harvey, P.D., 2003. Ziprasidone and cognition: the evolving story.J. Clin. Psychiatry 64 (Suppl. 19), 33–39.

Harvey, M., Shink, E., Tremblay, M., Gagne, B., Raymond, C., Labbe,M., Walther, D.J., Bader, M., Barden, N., 2004. Support for theinvolvement of TPH2 gene in affective disorders. Mol.Psychiatry 9, 980–981.

Hassler, M., 1992. Creative musical behavior and sex hormones:musical talent and spatial ability in the two sexes.Psychoneuroendocrinology 17, 55–70.

Hennig, J., Reuter, M., Netter, P., Burk, C., Landt, O., 2005. Two types ofaggression are differentially related to serotonergic activity andthe A779C TPH polymorphism. Behav. Neurosci. 119, 16–25.

Jäger, A.O., 1982. Berliner Intelligenzstruktur-Test (BIS-Test).Hogrefe, Göttingen.

Jäger, A.O., Süß, H.M., Beauducel, A., 1997. BerlinerIntelligenzstruktur-Test (BIS-Test): Form 4. Hogrefe, Göttingen.

Jonsson, E.G., Nothen, M.M., Grunhage, F., Farde, L., Nakashima, Y.,Propping, P., Sedvall, G.C., 1999. Polymorphisms in thedopamine D2 receptor gene and their relationships to striataldopamine receptor density of healthy volunteers. Mol.Psychiatry 4, 290–296.

Joober, R., Gauthier, J., Lal, S., Bloom, D., Lalonde, P., Rouleau, G.,Benkelfat, C., Labelle, A., 2002. Catechol-O-methyltransferaseVal-108/158-Met gene variants associatedwith performance onthe Wisconsin Card Sorting Test. Arch. Gen. Psychiatry 59,662–663.

Lachman, H.M., Papolos, D.F., Saito, T., Yu, Y.M., Szumlanski, C.L.,Weinshilboum, R.M., 1996. Humancatechol-O-methyltransferase pharmacogenetics: descriptionof a functional polymorphism and its potential application toneuropsychiatric disorders. Pharmacogenetics 6, 243–250.

Lerman, C., Caporaso, N.E., Bush, A., Zheng, Y.L., Audrain, J., Main,D., Shields, P.G., 2001. Tryptophan hydroxylase gene variantand smoking behavior. Am. J. Med. Genet. B105, 518–520.

Malhotra, A.K., Kestler, L.J., Mazzanti, C., Bates, J.A., Goldberg, T.,Goldman, D., 2002. A functional polymorphism in the COMTgene and performance on a test of prefrontal cognition. Am. J.Psychiatry 159, 652–654.

Mann, J.J., Malone, K.M., Nielsen, D.A., Goldman, D., Erdos, J.,Gelernter, J., 1997. Possible association of a polymorphism ofthe tryptophan hydroxylase gene with suicidal behavior indepressed patients. Am. J. Psychiatry 154, 1451–1453.

Mattay, V.S., Goldberg, T.E., Fera, F., Hariri, A.R., Tessitore, A., Egan,M.F., Kolachana, B., Callicott, J.H., Weinberger, D.R., 2003.Catechol O-methyltransferase val158-met genotype andindividual variation in the brain response to amphetamine.Proc. Natl. Acad. Sci. U. S. A. 100, 6186–6191.

Mazer, C., Muneyyirci, J., Taheny, K., Raio, N., Borella, A.,Whitaker-Azmitia, P., 1997. Serotonin depletion duringsynaptogenesis leads to decreased synaptic density andlearning deficits in the adult rat: a possible model ofneurodevelopmental disorders with cognitive deficits. BrainRes. 760, 68–73.

Nichols, R.C., 1978. Twin studies of ability, personality andinterests. Homo 29, 158–173.

Nielsen, D.A., Goldman, D., Virkkunen, M., Tokola, R., Rawlings, R.,Linnoila, M., 1994. Suicidality and 5-hydroxyindoleacetic acidconcentration associated with a tryptophan hydroxylasepolymorphism. Arch. Gen. Psychiatry 51, 34–38.

Noble, E.P., 2000. Addiction and its reward process throughpolymorphisms of the D2 dopamine receptor gene: a review.Eur. Psychiatr. 15, 79–89.

Panksepp, J., 1998. Affective neuroscience. The Foundations ofHuman and Animal Emotions. Oxford Univ. Press, London.

Petrill, S.A., Plomin, R., McClearn, G.E., Smith, D.L., Vignetti, S.,Chorney, M.J., Chorney, K., Thompson, L.A., Detterman, D.K.,Benbow, C., Lubinski, D., Daniels, J., Owen, M., McGuffin, P., 1997.No association between general cognitive ability and the A1 alleleof the D2 dopamine receptor gene. Behav. Genet. 27, 29–31.

Plomin, R., Thompson, L.A., 1993. Genetics and high cognitiveability. Ciba Found. Symp. 178, 67–79.

Plomin, R., McClearn, G.E., Smith, D.L., Vignetti, S., Chorney, M.J.,Chorney, K., Venditti, C.P., Kasarda, S., Thompson, L.A.,Detterman, D.K., et al., 1994. DNAmarkers associatedwith highversus low IQ: the IQ quantitative trait loci (QTL) project. Behav.Genet. 24, 107–118.

Plomin, R., Hill, L., Craig, I.W., McGuffin, P., Purcell, S., Sham, P.,Lubinski, D., Thompson, L.A., Fisher, P.J., Turic, D., Owen, M.J.,2001. A genome-wide scan of 1842 DNA markers for allelicassociations with general cognitive ability: a five-stage designusing DNA pooling and extreme selected groups. Behav. Genet.31, 497–509.

Pohjalainen, T., Rinne, J.O., Nagren, K., Lehikoinen, P., Anttila, K.,Syvalahti, E.K., Hietala, J., 1998. The A1 allele of the human D2dopamine receptor gene predicts lowD2 receptor availability inhealthy volunteers. Mol. Psychiatry 3, 256–260.

Porras, G., Di Matteo, V., Fracasso, C., Lucas, G., De Deurwaerdere,P., Caccia, S., Esposito, E., Spampinato, U., 2002. 5-HT2A and5-HT2C/2B receptor subtypes modulate dopamine releaseinduced in vivo by amphetamine and morphine in both the ratnucleus accumbens and striatum. Neuropsychopharmacology26, 311–324.

Poyurovsky, M., Koren, D., Gonopolsky, I., Schneidman, M., Fuchs,C., Weizman, A., Weizman, R., 2003. Effect of the 5-HT2antagonist mianserin on cognitive dysfunction in chronicschizophrenia patients: an add-on, double-blindplacebo-controlled study. Eur. Neuropsychopharmacol. 13,123–128.

Reuter, M., Hennig, J., 2005. Pleiotropic effect of the TPH A779Cpolymorphism on nicotine dependence and personality. Am. J.Med. Genet., B Neuropsychiatr. Genet. 134, 20–24.

Reuter, M., Schmitz, A., Corr, P., Hennig, J., 2005a. Moleculargenetics support Gray's personality theory: the interaction ofCOMT and DRD2 polymorphisms predicts the behavioralapproach system. Int. J. Neuropsychopharmacol. (Ahead ofprint already published online).

Reuter, M., Panksepp, J., Schnabel, N., Kellerhoff, N., Kempel, P.,Hennig, J., 2005b. Personality and biological markers ofcreativity. Eur. J. Pers. 19, 83–95.

Reuter, M., Hennig, J., Panksepp, J., Initial evidence for moleculargenetic validation of the Affective Neuroscience PersonalityTheory, submitted.

Ritchie, T., Noble, E.P., 1996. [3H]naloxone binding in the humanbrain: alcoholism and the TaqI A D2 dopamine receptorpolymorphism. Brain Res. 718, 193–197.

Ritchie, T., Noble, E.P., 2003. Association of seven polymorphismsof the D2 dopamine receptor gene with brain receptor-bindingcharacteristics. Neurochem. Res. 28, 73–82.

Rotondo, A., Schuebel, K., Bergen, A., Aragon, R., Virkkunen, M.,Linnoila, M., Goldman, D., Nielsen, D., 1999. Identification offour variants in the tryptophan hydroxylase promoter andassociation to behavior. Mol. Psychiatry 4, 360–368.

Rujescu, D., Giegling, I., Bondy, B., Gietl, A., Zill, P., Moller, H.J.,2002. Association of anger-related traits with SNPs in the TPHgene. Mol. Psychiatry 7, 1023–1029.

Souery, D., Van Gestel, S., Massat, I., Blairy, S., Adolfsson, R.,Blackwood, D., Del-Favero, J., Dikeos, D., Jakovljevic, M.,Kaneva, R., Lattuada, E., Lerer, B., Lilli, R., Milanova, V., Muir,

197B R A I N R E S E A R C H 1 0 6 9 ( 2 0 0 6 ) 1 9 0 – 1 9 7

W., Nothen, M., Oruc, L., Papadimitriou, G., Propping, P.,Schulze, T., Serretti, A., Shapira, B., Smeraldi, E., Stefanis, C.,Thomson, M., Van Broeckhoven, C., Mendlewicz, J., 2001.Tryptophan hydroxylase polymorphism and suicidality inunipolar and bipolar affective disorders: a multicenterassociation study. Biol. Psychiatry 49, 405–409.

Stavridou, A., Furnham, A., 1996. The relationship betweenpsychoticism, trait creativity and the attentional mechanismof cognitive inhibition. Pers. Individ. Differ. 21, 143–153.

Sternberg, R.J. (Ed.), 1988. The Nature of Creativity: ContemporaryPsychological Perspectives. Cambridge Univ. Press, Cambridge.

Sullivan, P.F., Jiang, Y., Neale, M.C., Kendler, K.S., Straub, R.E., 2001.Association of the tryptophan hydroxylase gene with smokinginitiation but not progression to nicotine dependence. Am. J.Med. Genet. B105, 479–484.

Syvänen, A.C., Tilgmann, C., Rinne, J., Ulmanen, I., 1997. Geneticpolymorphism of catechol-O-methyltransferase (COMT):correlation of genotype with individual variation of S-COMTactivity and comparison of the allele frequencies in the normal

population and parkinsonian patients in Finland.Pharmacogenetics 7, 65–71.

Thompson, J., Thomas, N., Singleton, A., Piggot, M., Lloyd, S.,Perry, E.K., Morris, C.M., Perry, R.H., Ferrier, I.N., Court, J.A.,1997. D2 dopamine receptor gene (DRD2) TaqI Apolymorphism: Reduced dopamine D2 receptor binding in thehuman striatum associated with the A1 allele.Pharmacogenetics 7, 479–484.

Tsai, S.J., Yu, Y.W., Lin, C.H., Chen, T.J., Chen, S.P., Hong, C.J., 2002.Dopamine D2 receptor and N-methyl-D-aspartate receptor 2Bsubunit genetic variants and intelligence. Neuropsychobiology45, 128–130.

Wainwright, M.A., Wright, M.J., Geffen, G.M., Luciano, M., Martin,N.G., 2005. The genetic basis of academic achievement on theQueensland Core Skills Test and its shared genetic variancewith IQ. Behav. Genet. 35, 133–145.

Walther, D.J., Peter, J.U., Bashammakh, S., Hortnagl, H., Voits, M.,Fink, H., Bader, M., 2003. Synthesis of serotonin by a secondtryptophan hydroxylase isoform. Science 299, 76.