Gene-Environment Interaction in Psychological Traits and Disorders

CP07CH15-Dick ARI 9 March 2011 21:44

Gene-EnvironmentInteraction in PsychologicalTraits and DisordersDanielle M. DickDepartment of Psychiatry, Virginia Institute for Psychiatric and Behavioral Genetics,Virginia Commonwealth University, Richmond, Virginia 23298-0126; email: [email protected]

Annu. Rev. Clin. Psychol. 2011. 7:383–409

First published online as a Review in Advance onJanuary 6, 2011

The Annual Review of Clinical Psychology is onlineat clinpsy.annualreviews.org

This article’s doi:10.1146/annurev-clinpsy-032210-104518

Copyright c© 2011 by Annual Reviews.All rights reserved

1548-5943/11/0427-0383$20.00

Keywords

genetics, association, review, GxE, psychopathology

Abstract

There has been an explosion of interest in studying gene-environmentinteractions (GxE) as they relate to the development of psychopathol-ogy. In this article, I review different methodologies to studygene-environment interaction, providing an overview of methods fromanimal and human studies and illustrations of gene-environment inter-actions detected using these various methodologies. Gene-environmentinteraction studies that examine genetic influences as modeled latently(e.g., from family, twin, and adoption studies) are covered, as well asstudies of measured genotypes. Importantly, the explosion of interestin gene-environment interactions has raised a number of challenges,including difficulties with differentiating various types of interactions,power, and the scaling of environmental measures, which haveprofound implications for detecting gene-environment interactions.Taking research on gene-environment interactions to the next level willnecessitate close collaborations between psychologists and geneticistsso that each field can take advantage of the knowledge base of the other.

383

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Gene: unit ofheredity; a stretch ofDNA that codes for aprotein

GxE: gene-environmentinteraction

Contents

INTRODUCTION . . . . . . . . . . . . . . . . . . 384DEFINING

GENE-ENVIRONMENTINTERACTION ANDDIFFERENTIATINGGENE-ENVIRONMENTCORRELATION . . . . . . . . . . . . . . . . . 385

METHODS FOR STUDYINGGENE-ENVIRONMENTINTERACTION. . . . . . . . . . . . . . . . . . 386Animal Research . . . . . . . . . . . . . . . . . . . 386Human Research . . . . . . . . . . . . . . . . . . 387

THE NATURE OFGENE-ENVIRONMENTINTERACTION. . . . . . . . . . . . . . . . . . 393

EPIGENETICS: A POTENTIALBIOLOGICAL MECHANISMFOR GENE-ENVIRONMENTINTERACTION. . . . . . . . . . . . . . . . . . 397

CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 399Knowledge Is Power . . . . . . . . . . . . . . . 399Use What We Already Know from

the Twin and DevelopmentalLiteratures . . . . . . . . . . . . . . . . . . . . . 399

Statistics Are Not a Substitutefor Critical Thinking . . . . . . . . . . . . 401

Play Nice in the Sandbox . . . . . . . . . . . 401

INTRODUCTION

Gene-environment interaction (GxE) has be-come a hot topic of research, with an expo-nential increase in interest in this area in thepast decade. Consider that PubMed lists only24 citations for “gene environment interaction”prior to the year 2000, but nearly four timesthat many in the first half of the year 2010alone! The projected publications on gene-environment interaction for 2008–2010 are ontrack to constitute more than 40% of the totalnumber of publications on gene-environmentinteraction indexed in PubMed. Where does allthis interest stem from? It may, in part, reflect a

merging of interests from fields that were tradi-tionally at odds with one another. Historically,there was a perception that behavior geneticistsfocused on genetic influences on behavior at theexpense of studying environmental influencesand that developmental psychologists focusedon environmental influences and largely ig-nored genetic factors. Although this criticism isnot entirely founded on the part of either field,methodological and ideological differences be-tween these respective fields meant that geneticand environmental influences were tradition-ally studied in isolation. More recently, therehas been recognition on the part of both ofthese fields that both genetic and environmentalinfluences are critical components to develop-mental outcome and that it is far more fruitful toattempt to understand how these factors cometogether to impact psychological outcomesthan to argue about which one is more im-portant. As Kendler and Eaves argued in theirarticle on the joint effect of genes and environ-ments, published more than two decades ago:

It is our conviction that a complete under-standing of the etiology of most psychiatricdisorders will require an understanding ofthe relevant genetic risk factors, the relevantenvironmental risk factors, and the ways inwhich these two risk factors interact. Such un-derstanding will only arise from research inwhich the important environmental variablesare measured in a genetically informative de-sign. Such research will require a synthesis ofresearch traditions within psychiatry that haveoften been at odds with one another in thepast. This interaction between the researchtradition that has focused on the genetic eti-ology of psychiatric illness and that which hasemphasized environmental causation will un-doubtedly be to the benefit of both. (Kendler& Eaves 1986, p. 288)

The PubMed data showing an expo-nential increase in published papers ongene-environment interaction suggest thatthat day has arrived. This has been facilitatedby the rapid advances that have taken place in

384 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


the field of genetics, making the incorporationof genetic components into traditional psycho-logical studies a relatively easy and inexpensiveendeavor. But with this surge of interest ingene-environment interaction, a number ofnew complications have emerged, and thestudy of gene-environment interaction facesnew challenges, including a recent backlashagainst studying gene-environment interaction(Risch et al. 2009). Addressing these chal-lenges will be critical to moving research ongene-environment interaction forward in aproductive way.

In this article, I first review differentstudy designs for detecting gene-environmentinteraction, providing an overview of methodsfrom animal and human studies. I covergene-environment interaction studies thatexamine genetic influences as modeled latentlyas well as studies of measured genotypes. In thestudy of latent gene-environment interaction,specific genotypes are not measured, butrather genetic influence is inferred basedon observed correlations between peoplewho have different degrees of genetic andenvironmental sharing. Thus, latent gene-environment interaction studies examine theaggregate effects of genes rather than any onespecific gene. Molecular genetic studies, incontrast, have generally focused on one specificgene of interest at a time. Relevant examplesof gene-environment interaction across thesedifferent methodologies are provided, thoughthese are meant to be more illustrative thanexhaustive, intended to introduce the reader torelevant studies and findings generated acrossthese various designs. Subsequently I reviewmore conceptual issues surrounding the studyof gene-environment interaction, coveringthe nature of gene-environment interactioneffects as well as the challenges facing thestudy of gene-environment interaction, such asdifficulties with differentiating various types ofinteractions, and how issues such as the scalingof environmental measures can have profoundimplications for studying gene-environmentinteraction. I include an overview of epigenet-ics, a relatively new area of study that provides

Epigenetics:modifications to thegenome that do notinvolve a change innucleotide sequence

a potential biological mechanism by which theenvironment can moderate gene expressionand affect behavior. Finally, I conclude withrecommendations for future directions andhow we can take research on gene-environmentinteraction to the next level.

DEFININGGENE-ENVIRONMENTINTERACTION ANDDIFFERENTIATINGGENE-ENVIRONMENTCORRELATION

It is important to first address some aspectsof terminology surrounding the study of gene-environment interaction. In lay terms, thephrase gene-environment interaction is oftenused to mean that both genes and environmentsare important. In statistical terms, this does notnecessarily indicate an interaction but couldbe consistent with an additive model, in whichthere are main effects of the environment andmain effects of genes. But in a statistical sense aninteraction is a very specific thing, referring to asituation in which the effect of one variable can-not be understood without taking into accountthe other variable. Their effects are not inde-pendent. When we refer to gene-environmentinteraction in a statistical sense, we are refer-ring to a situation in which the effect of genesdepends on the environment and/or the effectof the environment depends on genotype. Wenote that these two alternative conceptualiza-tions of gene-environment interaction are in-distinguishable statistically. It is this statisticaldefinition of gene-environment interaction thatis the primary focus of this review (except whereotherwise noted).

It is also important to note that geneticand environmental influences are not neces-sarily independent factors. That is to say thatalthough some environmental influences maybe largely random, such as experiencing a nat-ural disaster, many environmental influencesare not entirely random (Kendler et al. 1993).This phenomenon is called gene-environmentcorrelation. Three specific ways by which

www.annualreviews.org • Gene-Environment Interaction 385

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


genes may exert an effect on the environ-ment have been delineated (Plomin et al.1977, Scarr & McCartney 1983): (a) Passivegene-environment correlation refers to thefact that among biologically related relatives(i.e., nonadoptive families), parents providenot only their children’s genotypes but alsotheir rearing environment. Therefore, thechild’s genotype and home environment arecorrelated. (b) Evocative gene-environmentcorrelation refers to the idea that individuals’genotypes influence the responses they receivefrom others. For example, a child who ispredisposed to having an outgoing, cheerfuldisposition might be more likely to receivepositive attention from others than a childwho is predisposed to timidity and tears. Aperson with a grumpy, abrasive temperamentis more likely to evoke unpleasant responsesfrom coworkers and others with whom he/sheinteracts than is a cheerful, friendly person.Thus, evocative gene-environment correlationcan influence the way an individual experiencesthe world. (c) Finally, active gene-environmentcorrelation refers to the fact that an individualactively selects certain environments and takesaway different things from his/her environ-ment, and these processes are influenced by anindividual’s genotype. Therefore, an individualpredisposed to high sensation seeking may bemore prone to attend parties and meet newpeople, thereby actively influencing the envi-ronments he/she experiences. Evidence existsin the literature for each of these processes.The important point is that many sources ofbehavioral influence that we might consider“environmental” are actually under a degreeof genetic influence (Kendler & Baker 2007),so often genetic and environmental influencesdo not represent independent sources of influ-ence. This also makes it difficult to determinewhether the genes or the environment is thecausal agent. If, for example, individuals are ge-netically predisposed toward sensation seeking,and this makes them more likely to spend timein bars (a gene-environment correlation), andthis increases their risk for alcohol problems,are the predisposing sensation-seeking genes

or the bar environment the causal agent? Inactuality, the question is moot—they bothplayed a role; it is much more informative to tryto understand the pathways of risk than to askwhether the genes or the environment was thecritical factor. Though this review focuses ongene-environment interaction, it is importantfor the reader to be aware that this is but oneprocess by which genetic and environmentalinfluences are intertwined. Additionally, gene-environment correlation must be taken intoaccount when studying gene-environment in-teraction, a point that is mentioned again laterin this review. Excellent reviews covering thenature and importance of gene-environmentcorrelation also exist (Kendler 2011).

METHODS FOR STUDYINGGENE-ENVIRONMENTINTERACTION

Animal Research

Perhaps the most straightforward method fordetecting gene-environment interaction isfound in animal experimentation: Differentgenetic strains of animals can be subjected todifferent environments to directly test for gene-environment interaction. The key advantage ofanimal studies is that environmental exposurecan be made random to genotype, eliminatinggene-environment correlation and associatedproblems with interpretation. The most widelycited example of this line of research is Cooperand Zubek’s 1958 experiment, in which ratswere selectively bred to perform differentlyin a maze-running experiment (Cooper &Zubek 1958). Under standard environmentalconditions, one group of rats consistently per-formed with few errors (“maze bright”), whilea second group committed many errors (“mazedull”). These selectively bred rats were thenexposed to various environmental conditions:an enriched condition, in which rats werereared in brightly colored cages with manymoveable objects, or a restricted condition,in which there were no colors or toys. Theenriched condition had no effect on the maze

386 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


bright rats, although it substantially improvedthe performance of the maze dull rats, such thatthere was no difference between the groups.Conversely, the restrictive environment didnot affect the performance of the maze dull rats,but it substantially diminished the performanceof the maze bright rats, again yielding no dif-ference between the groups and demonstratinga powerful gene-environment interaction. Aseries of experiments conducted by Hendersonon inbred strains of mice, in which environ-mental enrichment was manipulated, alsoprovides evidence for gene-environment inter-action on several behavioral tasks (Henderson1970, 1972). These studies laid the foundationfor many future studies, which collectivelydemonstrate that environmental variationcan have considerable differential impact onoutcome depending on the genetic make-up ofthe animal (Wahlsten et al. 2003). However,animal studies are not without their limitations.Gene-environment interaction effects detectedin animal studies are still subject to the problemof scale (Mather & Jinks 1982), as discussed ingreater detail later in this review.

Human Research

Traditional behavior genetic designs.Demonstrating gene-environment interactionin humans has been considerably more difficultwhere ethical constraints require researchersto make use of natural experiments so envi-ronmental exposures are not random. Threetraditional study designs have been used todemonstrate genetic influence on behavior:family studies, adoption studies, and twinstudies. These designs have been used to detectgene-environment interaction also, and each isdiscussed in turn.

Family studies. Demonstration that a behav-ior aggregates in families is the first step in es-tablishing a genetic basis for a disorder (Hewitt& Turner 1995). Decreasing similarity withdecreasing degrees of relatedness lends supportto genetic influence on a behavior (Gottesman1991). This is a necessary, but not sufficient,condition for heritability. Similarity among

Heritability: theproportion of totalphenotypic variancethat can be accountedfor by genetic factors

family members is due both to shared genesand shared environment; family studies cannottease apart these two sources of variance to de-termine whether familiality is due to genetic orcommon environmental causes (Sherman et al.1997). However, family studies provide a pow-erful method for identifying gene-environmentinteraction. By comparing high-risk children,identified as such by the presence of psy-chopathology in their parents, with a controlgroup of low-risk individuals, it is possible totest the effects of environmental characteristicson individuals varying in genetic risk (Cannonet al. 1990). In a high-risk study of Danish chil-dren with schizophrenic mothers and matchedcontrols, institutional rearing was associatedwith an elevated risk of schizophrenia onlyamong those children with a genetic predisposi-tion (Cannon et al. 1990). When these subjectswere further classified on genetic risk as havingone or two affected parents, a significant inter-action emerged between degree of genetic riskand birth complications in predicting ventricleenlargement: The relationship between obstet-ric complications and ventricular enlargementwas greater in the group of individuals with oneaffected parent as compared to controls, andgreater still in the group of individuals with twoaffected parents (Cannon et al. 1993). Anotherstudy also found that among individuals athigh risk for schizophrenia, experiencing ob-stetric complications was related to an earlierhospitalization (Malaspina et al. 1999).

Another creative method has made use ofthe natural experiment of family migration todemonstrate gene-environment interaction:The high rate of schizophrenia among African-Caribbean individuals who emigrated to theUnited Kingdom is presumed to result fromgene-environment interaction. Parents andsiblings of first-generation African-Caribbeanprobands have risks of schizophrenia similarto those for white individuals in the area.However, the siblings of second-generationAfrican-Caribbean probands have markedlyelevated rates of schizophrenia, suggesting thatthe increase in schizophrenia rates is due toan interaction between genetic predispositions


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


and stressful environmental factors encoun-tered by this population (Malaspina et al.1999, Moldin & Gottesman 1997). Althoughfamily studies provide a powerful design fordemonstrating gene-environment interaction,there are limitations to their utility. High-riskstudies are very expensive to conduct becausethey require the examination of individualsover a long period of time. Additionally, alarge number of high-risk individuals must bestudied in order to obtain a sufficient numberof individuals who eventually become affected,due to the low base rate of most mentaldisorders. Because of these limitations, fewexamples of high-risk studies exist.

Adoption studies. Adoption and twin studiesare able to clarify the extent to which similarityamong family members is due to shared genesversus shared environment. In their simplestform, adoption studies involve comparingthe extent to which adoptees resemble theirbiological relatives, with whom they sharegenes but not family environment, with theextent to which adoptees resemble their adop-tive relatives, with whom they share familyenvironment but not genes. Adoption studieshave been pivotal in advancing our under-standing of the etiology of many disordersand drawing attention to the importance ofgenetic factors. For example, Heston’s historicadoption study was critical in dispelling themyth of schizophrenogenic mothers in favor ofa genetic transmission explaining the familialityof schizophrenia (Heston & Denney 1967).Furthermore, adoption studies provide a pow-erful method of detecting gene-environmentinteractions and have been called the humananalogue of strain-by-treatment animal studies(Plomin & Hershberger 1991). The genotypeof adopted children is inferred from theirbiological parents, and the environment ismeasured in the adoptive home. Individualsthought to be at genetic risk for a disorder,but reared in adoptive homes with differentenvironments, are compared to each other andto control adoptees. This methodology hasbeen employed by a number of research groups

to document gene-environment interactionsin a variety of clinical disorders: In a series ofIowa adoption studies, Cadoret and colleaguesdemonstrated that a genetic predisposition toalcohol abuse predicted major depression infemales only among adoptees who also experi-enced a disturbed environment, as defined bypsychopathology, divorce, or legal problemsamong the adoptive parents (Cadoret et al.1996). In another study, depression scoresand manic symptoms were found to be higheramong individuals with a genetic predisposi-tion and a later age of adoption (suggesting amore transient and stressful childhood) thanamong those with only a genetic predisposition(Cadoret et al. 1990). In an adoption studyof Swedish men, mild and severe alcoholabuse were more prevalent only among menwho had both a genetic predisposition andmore disadvantaged adoptive environments(Cloninger et al. 1981). The Finnish AdoptiveFamily Study of Schizophrenia found that highgenetic risk was associated with increased riskof schizophrenic thought disorder only whencombined with communication deviance inthe adoptive family (Wahlberg et al. 1997).Additionally, the adoptees had a greater risk ofpsychological disturbance, defined as neuroti-cism, personality disorders, and psychoticism,when the adoptive family environment was dis-turbed (Tienari et al. 1990). These studies havedemonstrated that genetic predispositions fora number of psychiatric disorders interact withenvironmental influences to manifest disorder.

However, adoption studies suffer froma number of methodological limitations.Adoptive parents and biological parents ofadoptees are often not representative of thegeneral population. Adoptive parents tend tobe socioeconomically advantaged and havelower rates of mental problems, due to theextensive screening procedures conducted byadoption agencies (Kendler 1993). Biologicalparents of adoptees tend to be atypical, as well,but in the opposite way. Additionally, selectiveplacement by adoption agencies is confoundingthe clear-cut separation between genetic andenvironmental effects by matching adoptees

388 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


and adoptive parents on demographics, suchas race and religion. An increasing number ofadoptions are also allowing contact betweenthe biological parents and adoptive children,further confounding the traditional genetic andenvironmental separation that made adoptionstudies useful for genetically informativeresearch. Finally, greater contraceptive useis making adoption increasingly rare (Martinet al. 1997). Accordingly, this research strategyhas become increasingly challenging, thougha number of current adoption studies continueto make important contributions to the field(Leve et al. 2010; McGue et al. 1995, 1996).

Twin studies. Twins provide a number ofways to study gene-environment interaction.One such method is to study monozygotic twinsreared apart (MZA). MZAs provide a uniqueopportunity to study the influence of differentenvironments on identical genotypes. In theSwedish Adoption/Twin Study of Aging, datafrom 99 pairs of MZAs were tested for interac-tions between childhood rearing and adult per-sonality (Bergeman et al. 1988). Several signif-icant interactions emerged. In some cases, theenvironment had a stronger impact on individ-uals genetically predisposed to be low on a giventrait (based on the cotwin’s score). For example,individuals high in extraversion expressed thetrait regardless of the environment; however,individuals predisposed to low extraversion hadeven lower scores in the presence of a control-ling family. In other traits, the environmenthad a greater impact on individuals geneticallypredisposed to be high on the trait: Individualspredisposed to impulsivity were even moreimpulsive in a conflictual family environment;individuals low on impulsivity were not af-fected. Finally, some environments influencedboth individuals who were high and low on agiven trait, but in opposite directions: Familiesthat were more involved masked genetic differ-ences between individuals predisposed towardhigh or low neuroticism, but greater geneticvariation emerged in less controlling families.

The implementation of population-basedtwin studies, inclusion of measured envi-

MZ: monozygotic

DZ: dizygotic

ronments into twin studies, and advances inbiometrical modeling techniques for twin datamade it possible to study gene-environmentinteraction within the framework of the classictwin study. Traditional twin studies involvecomparisons of monozygotic (MZ) and dizy-gotic (DZ) twins reared together. MZ twinsshare all of their genetic variation, whereas DZtwins share on average 50% of their geneticmake-up; however, both types of twins areage-matched siblings sharing their familyenvironments. This allows heritability, or theproportion of variance attributed to additivegenetic effects, to be estimated by (a) doublingthe difference between the correlation foundbetween MZ twins and the correlation foundbetween DZ twins, for quantitative traits, or(b) comparing concordance rates between MZsand DZs, for qualitative disorders (McGue& Bouchard 1998). Biometrical model-fittingmade it possible for researchers to address in-creasingly sophisticated research questions byallowing one to statistically specify predictionsmade by various hypotheses and to comparemodels testing competing hypotheses. Bymodeling data from subjects who vary onexposure to a specified environment, one couldtest whether there is differential expression ofgenetic influences in different environments.

Early examples of gene-environment in-teraction in twin models necessitated “group-ing” environments to fit multiple group mod-els. The basic idea was simple: Fit models todata for people in environment 1 and envi-ronment 2 separately and then test whetherthere were significant differences in the im-portance of genetic and environmental fac-tors across the groups using basic structuralequation modeling techniques. In an early ex-ample of gene-environment interaction, datafrom the Australian twin register were usedto test whether the relative importance of ge-netic effects on alcohol consumption varied asa function of marital status, and in fact theydid (Heath et al. 1989). Having a marriage-like relationship reduced the impact of geneticinfluences on drinking: Among the youngersample of twins, genetic liability accounted


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


for but half as much variance in drinkingamong married women (31%) as among un-married women (60%). A parallel effect wasfound among the adult twins: Genetic effectsaccounted for less than 60% of the variance inmarried respondents but more than 76% in un-married respondents (Heath et al. 1989). In anindependent sample of Dutch twins, religiositywas also shown to moderate genetic and envi-ronmental influences on alcohol use initiationin females (with nonsignificant trends in thesame direction for males): In females withouta religious upbringing, genetic influences ac-counted for 40% of the variance in alcohol useinitiation compared to 0% in religiously raisedfemales. Shared environmental influences werefar more important in the religious females(Koopmans et al. 1999). In data from ourpopulation-based Finnish twin sample, we alsofound that regional residency moderates theimpact of genetic and environmental influenceson alcohol use. Genetic effects played a largerrole in longitudinal drinking patterns from lateadolescence to early adulthood among individ-uals residing in urban settings, whereas com-mon environmental effects exerted a greater in-fluence across this age range among individualsin rural settings (Rose et al. 2001). When onehas pairs discordant for exposure, it is also pos-sible to ask about genetic correlation betweentraits displayed in different environments.

One obvious limitation of modeling gene-environment interaction in this way was thatit constrained investigation to environmentsthat fell into natural groupings (e.g., mar-ried/unmarried; urban/rural) or it forcedinvestigators to create groups based on envi-ronments that may actually be more continuousin nature (e.g., religiosity). In the first extensionof this work to quasi-continuous environmentalmoderation, we developed a model that allowedgenetic and environmental influences to varyas a function of a continuous environmentalmoderator and used this model to follow-up onthe urban/rural interaction reported previously(Dick et al. 2001). We believed it likely that theurban/rural moderation effect reflected a com-posite of different processes at work. Accord-

ingly, we expanded the analyses to incorporatemore specific information about neighborhoodenvironments, using government-collectedinformation about the specific municipalities inwhich the twins resided (Dick et al. 2001). Wefound that genetic influences were strongerin environments characterized by higher ratesof migration in and out of the municipality;conversely, shared environmental influencespredominated in local communities charac-terized by little migration. We also foundthat genetic predispositions were stronger incommunities composed of a higher percentageof young adults slightly older than our age-18Finnish twins and in regions where there werehigher alcohol sales. Further, the magnitude ofgenetic moderation observed in these modelsthat allowed for variation as a function of aquasi-continuous environmental moderatorwas striking, with nearly a fivefold differencein the magnitude of genetic effects betweenenvironmental extremes in some cases.

The publication of a paper the followingyear (Purcell 2002) that provided straightfor-ward scripts for continuous gene-environmentinteraction models using the most widelyused program for twin analyses, Mx (Neale2000), led to a surge of papers studying gene-environment interaction in the twin literature.These scripts also offered the advantage of be-ing able to take into account gene-environmentcorrelation in the context of gene-environmentinteraction. This was an important advance be-cause previous examples of gene-environmentinteraction in twin models had been limitedto environments that showed no evidence ofgenetic effects so as to avoid the confoundingof gene-environment interaction with gene-environment correlation. Using these models,we have demonstrated that genetic influenceson adolescent substance use are enhanced inenvironments with lower parental monitoring(Dick et al. 2007c) and in the presence ofsubstance-using friends (Dick et al. 2007b).Similar effects have been demonstrated formore general externalizing behavior: Geneticinfluences on antisocial behavior were higherin the presence of delinquent peers (Button

390 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


et al. 2007) and in environments characterizedby high parental negativity (Feinberg et al.2007), low parental warmth (Feinberg et al.2007), and high paternal punitive discipline(Button et al. 2008). Further, in an exten-sion of the socioregional-moderating effectsobserved on age-18 alcohol use, we found aparallel moderating role of these socioregionalvariables on age-14 behavior problems in girlsin a younger Finnish twin sample. Geneticinfluences assumed greater importance inurban settings, communities with greatermigration, and communities with a higherpercentage of slightly older adolescents.

Other psychological outcomes havealso yielded significant evidence of gene-environment interaction effects in the twinliterature. For example, a moderating ef-fect, parallel to that reported for alcoholconsumption above, has been reported fordepression symptoms (Heath et al. 1998) infemales. A marriage-like relationship reducedthe influence of genetic liability to depressionsymptoms, paralleling the effect found for al-cohol consumption: Genetic factors accountedfor 29% of the variance in depression scoresamong married women, but for 42% of thevariance in young unmarried females and 51%of the variance in older unmarried females(Heath et al. 1998). Life events were also foundto moderate the impact of factors influencingdepression in females (Kendler et al. 1991).Genetic and/or shared environmental influ-ences were significantly more important ininfluencing depression in high-stress than inlow-stress environments, as defined by a me-dian split on a life-event inventory, althoughthere was insufficient power to determinewhether the moderating influence was ongenetic or environmental effects.

More than simply accumulating examplesof moderation of genetic influence by envi-ronmental factors, efforts have been made tointegrate this work into theoretical frameworkssurrounding the etiology of different clinicalconditions. This is critical if science is toadvance beyond individual observations totestable broad theories. A 2005 review paper

by Shanahan and Hofer suggested four pro-cesses by which social context may moderatethe relative importance of genetic effects(Shanahan & Hofer 2005). The environmentmay (a) trigger or (b) compensate for a geneticpredisposition, (c) control the expression of agenetic predisposition, or (d ) enhance a geneticpredisposition (referring to the accentuationof “positive” genetic predispositions). Theseprocesses are not mutually exclusive and canrepresent different ends of a continuum. Forexample, the interaction between geneticsusceptibility and life events may represent asituation whereby the experience of life eventstriggers a genetic susceptibility to depression.Conversely, “protective” environments, suchas marriage-like relationships and low stresslevels, can buffer against or reduce the impact ofgenetic predispositions to depressive problems.Many different processes are likely involved inthe gene-environment interactions observedfor substance use and antisocial behavior.For example, family environment and peersubstance use/delinquency likely constitutea spectrum of risk or protection, and fam-ily/friend environments that are at the “poor”extreme may trigger genetic predispositionstoward substance use and antisocial behavior,whereas positive family and friend relationshipsmay compensate for genetic predispositionstoward substance use and antisocial behavior.Social control also appears to be a particularlyrelevant process in substance use, as it is likelythat being in a marriage-like relationshipand/or being raised with a religious upbringingexert social norms that constrain behavior andthereby reduce genetic predispositions towardsubstance use. Further, the availability of thesubstance also serves as a level of control overthe ability to express genetic predispositions,and accordingly, the degree to which geneticinfluences will be apparent on an outcome atthe population level. In a compelling illustra-tion of this effect, Boardman and colleaguesused twin data from the National Survey ofMidlife Development in the United States andfound a significant reduction in the importanceof genetic influences on people who smoke


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


DNA:deoxyribonucleic acid

Polymorphism: alocation in a gene thatcomes in multipleforms

Allele: naturalvariation in the geneticsequence; can be achange in a singlenucleotide or longerstretches of DNA

regularly following legislation prohibitingsmoking in public places (Boardman et al.2010).

Molecular analyses. All of the analyses dis-cussed thus far use latent, unmeasured indicesof genetic influence to detect the possiblepresence of gene-environment interaction.This is largely because it was possible to testfor the presence of latent genetic influencein humans (via comparisons of correlationsbetween relatives with different degrees ofgenetic sharing) long before molecular geneticsyielded the techniques necessary to identifyspecific genes influencing complex psycholog-ical disorders. However, recent advances havemade the collection of deoxyribonucleic acid(DNA) and resultant genotyping relativelycheap and straightforward. Additionally, thepublication of high-profile papers broughtgene-environment interaction to the forefrontof mainstream psychology. In a pair of paperspublished in Science in 2002 and 2003, re-spectively, Caspi and colleagues analyzed datafrom a prospective, longitudinal sample from abirth cohort from New Zealand, followed frombirth through adulthood. In the 2002 paper,they reported that a functional polymorphismin the gene encoding the neurotransmitter-metabolizing enzyme monoamine oxidase A(MAOA) moderated the effect of maltreatment:Males who carried the genotype conferringhigh levels of MAOA expression were lesslikely to develop antisocial problems whenexposed to maltreatment (Caspi et al. 2002). Inthe 2003 paper, they reported that a functionalpolymorphism in the promoter region of theserotonin transporter gene (5-HTT) was foundto moderate the influence of stressful life eventson depression. Individuals carrying the shortallele of the 5-HTT promoter polymorphismexhibited more depressive symptoms, diagnos-able depression, and suicidality in relation tostressful life events than did individuals ho-mozygous for the long allele (Caspi et al. 2003).Both studies were significant in demonstratingthat genetic variation can moderate individuals’sensitivity to environmental events. These

studies sparked a multitude of reports thataimed to replicate, or to further extend andexplore, the findings of the original papers,resulting in huge literatures surrounding eachreported gene-environment interaction in theyears since the original publications (e.g., Ed-wards et al. 2009, Enoch et al. 2010, Frazzettoet al. 2007, Kim-Cohen et al. 2006, McDer-mott et al. 2009, Prom-Wormley et al. 2009,Vanyukov et al. 2007, Weder et al. 2009). It isbeyond the scope of this review to detail thesestudies; however, of note was the publicationin 2009 of a highly publicized meta-analysis ofthe interaction between 5-HTT, stressful lifeevents, and risk of depression that concludedthere was “no evidence that the serotonintransporter genotype alone or in interactionwith stressful life events is associated with an el-evated risk of depression in men alone, womenalone, or in both sexes combined” (Risch et al.2009). Further, the authors were critical ofthe rapid embracing of gene-environmentinteraction and the substantial resources thathave been devoted to this research. The paperstimulated considerable backlash against thestudy of gene-environment interactions, andthe pendulum appeared to be swinging backthe other direction. However, a recent reviewby Caspi and colleagues entitled “GeneticSensitivity to the Environment: The Caseof the Serotonin Transporter Gene and ItsImplications for Studying Complex Diseasesand Traits” highlighted the fact that evidencefor involvement of 5-HTT in stress sensitivitycomes from at least four different types ofstudies, including observational studies inhumans, experimental neuroscience studies,studies in nonhuman primates, and studiesof 5-HTT mutations in rodents (Caspi et al.2010). Further, the authors made the distinc-tion between different cultures of evaluatinggene-environment interactions: a purely statis-tical (theory-free) approach that relies whollyon meta-analysis (e.g., such as that taken byRisch et al. 2009) versus a construct-validity(theory-guided) approach that looks for anomological network of convergent evidence,such as the approach that they took.

392 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


It is likely that this distinction also reflectsdifferences in training and emphasis across dif-ferent fields. The most cutting-edge geneticstrategies at any given point, though they havechanged drastically and rapidly over the pastseveral decades, have generally involved athe-oretical methods for gene identification (Nealeet al. 2008). This was true of early linkage analy-ses, where ∼400 to 1,000 markers were scannedacross the genome to search for chromoso-mal regions that were shared by affected fam-ily members, suggesting there may be a genein that region that harbored risk for the par-ticular outcome under study. This allowed ge-neticists to search for genes without having toknow anything about the underlying biology,with the ideas that the identification of riskgenes would be informative as to etiologicalprocesses and that our understanding of the bi-ology of most psychiatric conditions is limited.Although it is now recognized that linkage stud-ies were underpowered to detect genes of smalleffect, such as those now thought to be operat-ing in psychiatric conditions, this atheoreticalapproach was retained in the next generation ofgene-finding methods that replaced linkage—the implementation of genome-wide associa-tion studies (GWAS) (Cardon 2006). GWASalso have the general framework of scanningmarkers located across the entire genome inan effort to detect association between geneticmarkers and disease status; however, in GWASover a million markers (or more, on the newestgenetic platforms) are analyzed. The next tech-nique on the horizon is sequencing, in whichentire stretches of DNA are sequenced to knowthe exact base pair sequence for a given region(McKenna et al. 2010). From linkage to se-quencing, common across all these techniquesis an atheoretical framework for finding genesthat necessarily involves conducting very largenumbers of tests. Accordingly, there has beengreat emphasis in the field of genetics on correc-tion for multiple testing (van den Oord 2007).In addition, the estimated magnitude of effectsize of genetic variants thought to influencecomplex behavioral outcomes has been con-tinually shifted downward as studies that were

GWAS: genome-wideassociation study

ORs: odds ratios

sufficiently powered to detect effect sizes pre-viously thought to be reasonable have failed togenerate positive findings (Manolio et al. 2009).GWAS have led the field to believe that genesinfluencing complex behavioral outcomes likelyhave odds ratios (ORs) on the order of magni-tude of 1.1. This has led to a need for incredi-bly large sample sizes, requiring meta-analyticGWAS efforts with several tens of thousandsof subjects (Landi et al. 2009, Lindgren et al.2009).

It is important to note there has beenincreasing attention to the topic of gene-environment interaction from geneticists(Engelman et al. 2009). This likely reflects,in part, frustration and difficulty with identi-fying genes that impact complex psychiatricoutcomes. Several hypotheses have been putforth as possible explanations for the failure torobustly detect genes involved in psychiatricoutcomes, including a genetic model involvingfar more genes, each of very small effect, thanwas previously recognized, and failure to payadequate attention to rare variants, copy num-ber variants, and gene-environment interaction(Manolio et al. 2009). Accordingly, gene-environment interaction is being discussed farmore in the area of gene finding than in yearspast; however, these discussions often involveatheoretical approaches and center on methodsto adequately detect gene-environment inter-action in the presence of extensive multiple test-ing (Gauderman 2002, Gauderman et al. 2010).The papers by Risch et al. (2009) and Caspiet al. (2010) on the interaction between 5-HTT,life stress, and depression highlight the concep-tual, theoretical, and practical differences thatcontinue to exist between the fields of geneticsand psychology surrounding the identificationof gene-environment interaction effects.

THE NATURE OFGENE-ENVIRONMENTINTERACTION

An important consideration in the studyof gene-environment interaction is the na-ture, or shape, of the interaction that one


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Phenotype: theobserved outcomeunder study; can bethe manifestation ofboth genetic and/orenvironmental factors

hypothesizes. Figure 1 illustrates the twoprimary types of interactions. One type ofinteraction is the fan-shaped interaction(Figure 1a). In this type of interaction,the influence of genotype is greater in oneenvironmental context than in another. Thisis the kind of interaction that is hypothesizedby a diathesis-stress framework, wherebygenetic influences become more apparent, i.e.,are more strongly related to outcome, in thepresence of negative environmental conditions.There is a reduced (or no) association of geno-type with outcome in the absence of exposureto particular environmental conditions. Theliterature surrounding depression and lifeevents would be an example of a hypothesizedfan-shaped interaction: When life stressors areencountered, genetically vulnerable individualsare more prone to developing depression,whereas in the absence of life stressors, theseindividuals may be no more likely to developdepression. In essence, it is only when adverseenvironmental conditions are experienced thatthe genes “come on-line.” Gene-environmentinteractions in the area of adolescent substanceuse are also hypothesized to be fan-shaped,where some environmental conditions willallow greater opportunity to express geneticpredispositions (allowing for more variation bygenotype as in the right side of Figure 1a), andother environments will exert social controlin such a way as to curb genetic expression(Shanahan & Hofer 2005), leading to re-duced genetic variance (as on the left side ofFigure 1a). Twin analyses yielding evidence ofgenetic influences being more or less importantin different environmental contexts are gen-erally suggestive of fan-shaped interactions.Changes in the overall heritability do notnecessarily dictate that any one specific suscep-tibility gene will operate in a parallel manner;however, a change in heritability suggeststhat at least a good portion of the involvedgenes (assuming many genes of approximatelyequal and small effect) must be operating inthat manner for a difference in heritability byenvironment to be detectable. The diathesis-stress model has largely been the dominant

model in psychiatry. Gene-finding effortshave focused on the search for vulnerabilitygenes, and gene-environment interaction hasbeen discussed in the context of these geneticeffects becoming more or less important underparticular environmental conditions.

More recently, an alternative framework hasbeen proposed by Belsky and colleagues—thedifferential susceptibility hypothesis—in whichthe same individuals who are most adverselyaffected by negative environments may also bethose who are most likely to benefit from pos-itive environments. Rather than searching for“vulnerability genes” influencing psychiatricand behavioral outcomes, they propose the ideaof “plasticity genes,” or genes involved in re-sponsivity to environmental conditions (Belskyet al. 2009). Belsky and colleagues reviewed theliteratures surrounding gene-environment in-teractions associated with three widely studiedcandidate genes, MAOA, 5-HTT, and DRD4,and suggested that the results provide evidencefor differential susceptibility associated withthese genes (Belsky et al. 2009). Their hypothe-sis is closely related to the concept of biologicalsensitivity to context (Ellis & Boyce 2008). Theidea of biological sensitivity to context has itsroots in evolutionary developmental biology,whereby selection pressures should favor geno-types that support a range of phenotypes inresponse to environmental conditions becausethis flexibility would be beneficial from theperspective of survival of the species. However,biological sensitivity to context has the poten-tial for both positive effects under more highlysupportive environmental conditions and neg-ative effects in the presence of more negativeenvironmental conditions. This theory hasbeen most fully developed and discussed inthe context of stress reactivity (Boyce & Ellis2005), where it has been demonstrated thathighly reactive children show disproportionaterates of morbidity when raised in adverseenvironments, but particularly low rates whenraised in low-stress, highly supportive environ-ments (Ellis et al. 2005). In these studies, highreactivity was defined by response to differentlaboratory challenges, and the authors noted

394 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


that the underlying cellular mechanisms thatwould produce such responses are currentlyunknown, though genetic factors are likely toplay a role (Ellis & Boyce 2008).

Although fan-shaped and crossover interac-tions are theoretically different, in practice, theycan be quite difficult to differentiate. In lookingat Figures 1a and b, one can imagine several“variations on the theme” for both fan-shapedand crossover interactions. In general for a fan-shaped interaction, a main effect of genotypewill be present as well as a main effect of the en-vironment. In the fan-shaped interaction shownin Figure 1a, there is a main effect of genotypeat both environmental extremes; it is simplyfar stronger in environment 2 (far right side ofthe graph) as compared to environment −2 (farleft side). But one could imagine a fan-shapedinteraction where there was no genotypic ef-fect at one extreme (e.g., the lines converge tothe same phenotypic mean at environment −2).Further, fan-shaped interactions can differ inthe slope of the lines for each genotype, whichindicate how much the environment is modify-ing genetic effects. In the crossover interactionshown in Figure 1b, the lines cross at envi-ronment 0 (i.e., in the middle). But crossoverinteractions can vary in the location of thecrossover. It is possible that crossing over onlyoccurs at the environmental extreme. As pre-viously noted, the crossing over of the geno-typic groups in the Caspi et al. publications ofthe interactions between the 5-HTT gene, lifeevents, and depression (Caspi et al. 2003) andbetween MAOA, maltreatment, and antisocialbehavior (Caspi et al. 2002) occurred at the ex-treme low ends of the environmental measures,and the degree of crossing over was quite mod-est. Rather, the shape of the interactions (andthe way the interactions were conceptualizedin the papers) was largely fan-shaped, wherebycertain genotypic groups showed stronger as-sociations with outcome as a function of theenvironmental stressor. Also, in both cases, thegenetic variance was far greater under one envi-ronmental extreme than the other, rather thanbeing approximately equivalent at both ends ofthe distribution (but with genotypic effects in

opposite directions) as is the case in thecrossover effect illustrated in Figure 1b. In gen-eral, it is assumed that main effects of genotypewill not be detected in crossover interactions,but this will actually depend on the frequencyof the different levels of the environment. Thisis also true of fan-shaped interactions, but toa lesser degree. Note that the crossover de-picted in Figure 1b does indicate a main ef-fect of the environment. However, here toosubstantial variation can be imagined. For ex-ample, if the crossover took the shape seen inFigure 1c, assuming approximately equal dis-tributions of environmental levels, you wouldnot find a main effect for genotype or environ-ment. Interactions of this sort are assumed tobe relatively rare.

Evaluating the relative importance, or fre-quency of existence, of each type of interactionis complicated by the fact that there is far morepower to detect crossover interactions thanfan-shaped interactions. Knowing that most ofour genetic studies are likely underpowered, wewould expect a preponderance of crossover ef-fects to be detected as compared to fan-shapedeffects purely as a statistical artifact. Further,even when a crossover effect is observed,power considerations can make it difficult todetermine if it is “real.” For example, an inter-action observed in our data between the geneCHRM2, parental monitoring, and adolescentexternalizing behavior yielded consistent evi-dence for a gene-environment interaction, witha crossing of the observed regression lines (asin Figure 1b). However, the mean differencesby genotype were not significant at eitherend of the environmental continuum, so it isunclear whether the crossover reflected truedifferential susceptibility or simply overfittingof the data across the environmental levelscontaining the majority of the observations,which contributed to a crossing over of theregression lines at one environmental extreme(Dick et al. 2011). Larger studies would havegreater power to make these differentiations;however, there is the unfortunate paradox thatthe samples with the greatest depth of pheno-typic information, allowing for more complex


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


tests about risk associated with particulargenes, usually have much smaller sample sizesdue to the trade-off necessary to collect the richphenotypic information. This is an importantissue for gene-environment interaction studiesin general: Most have been underpowered,and this raises concerns about the likelihoodthat detected effects are true positives. Thereare several freely available programs to esti-mate power (Gauderman 2002, Purcell et al.2003), and it is critical that papers reportinggene-environment interaction effects (or a lackthereof) include information about the powerof their sample in order to interpret the results.

Another widely contested issue is whethergene-environment interactions should be ex-amined only when main effects of genotype aredetected. Perhaps not surprisingly, this is theapproach most commonly advocated by sta-tistical geneticists (Risch et al. 2009) and thatwas recommended by the Psychiatric GWASConsortium (Psychiatr. GWAS Consort. Steer.Comm. 2008). However, this strategy couldpreclude the detection of crossover interactioneffects as well as gene-environment interactionsthat occur in the presence of relatively low-frequency environments. In addition, if geneticeffects are conditional on environmental expo-sure, main effects of genotype could vary acrosssamples, that is to say, a genetic effect could bedetected in one sample and fail to replicate inanother if the samples differ on environmentalexposure.

Another issue with the detection and in-terpretation of gene-environment interactioneffects involves the range of environmentsbeing studied. For example, if we assume thatthe five levels of the environment shown inFigure 1b represent the true full range ofenvironments that exist, if a particular studyonly included individuals from environments0–2, it would conclude that there is a fan-shaped gene-environment interaction. Belskyand colleagues (2009) have suggested this maybe particularly problematic in the psychiatricliterature because only in rare exceptions(Bakermans-Kranenburg & van Ijzendoorn2006, Taylor et al. 2006) has the environment

included both positive and negative ends ofthe spectrum. Rather, the absence of envi-ronmental stressors has usually constitutedthe “low” end of the environment, e.g., theabsence of life stressors (Caspi et al. 2003) orthe absence of maltreatment (Caspi et al. 2002).This could lead individuals to conclude thereis a fan-shaped interaction because they areessentially failing to measure, with referenceto Figure 1b, environments −2 and −1, whichrepresent the positive end of the environmentalcontinuum. In looking at Figures 1a and b,one can imagine a number of other incorrectconclusions that could be drawn about thenature of gene-environment interaction effectsas a result of restricted range of environmentalmeasures. For example, in Figure 1b, mea-surement of individuals from environments −2to 0 would lead one to conclude that geneticeffects play a stronger role at lower levelsof environmental exposure. Measurement ofindividuals from environments 0 to 2 wouldlead one to conclude that genetic effects play astronger role at higher levels of exposure to thesame environmental variable. In Figure 1a,if measurement of individuals was limitedto environments −2 and −1, depending onsample size, there may be inadequate power todetect deviation from a purely additive geneticmodel, e.g., the slope of the genotypic linesmay not be significantly different.

It is also important to note that not only arethere several scenarios that would lead one tomake incorrect conclusions about the natureof a gene-environment interaction effect,there are also scenarios that would lead one toconclude that a gene-environment interactionexists when it actually does not. Several of theseare detailed in a sobering paper by my colleagueLindon Eaves, in which significant evidencefor gene-environment interaction was detectedquite frequently using standard regressionmethods, when the simulated data reflectedstrictly additive models (Eaves 2006). This wasparticularly problematic when using logisticregression where a dichotomous diagnosis wasthe outcome. The problem was further exag-gerated when selected samples were analyzed.

396 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


An additional complication with evaluatinggene-environment interactions in psychologyis that often our environmental measures don’thave absolute scales of measurement. Forexample, what is the “real” metric for measur-ing a construct like parent-child bonding, ormaltreatment, or stress? This becomes criticalbecause fan-shaped interactions are very sen-sitive to scaling. Often a transformation of thescale scores will make the interaction disappear.What does it mean if the raw variable shows aninteraction but the log transformation of thescale scores does not? Is the interaction real?Is one metric for measuring the environmenta better reflection of the “real” nature ofthe environment than another? Many of theenvironments of interest to psychologists donot have true metrics, such as those that existfor measures such as height, weight, or otherphysiological variables. This is an issue forthe study of gene-environment interaction.It becomes even more problematic when youconsider that logistic regression is the methodcommonly used to test for gene-environmentinteractions with dichotomous disease statusoutcomes. Logistic regression involves alogarithmic transformation of the probabilityof being affected. By definition, this changesthe nature of the relationship between thevariables being modeled. This compoundsproblems associated with gene-environmentinteractions being scale dependent.

EPIGENETICS: A POTENTIALBIOLOGICAL MECHANISMFOR GENE-ENVIRONMENTINTERACTION

An enduring question remains in the study ofgene-environment interaction: how does theenvironment “get under the skin”? Stated in an-other way, what are the biological processes bywhich exposure to environmental events couldaffect outcome? Epigenetics is one candidatemechanism. Excellent recent reviews on thistopic exist (Meaney 2010, Zhang & Meaney2010), and I provide a brief overview here. Itis important to note, however, that although

epigenetics is increasingly discussed in the con-text of gene-environment interaction, it doesnot relate directly to gene-environment inter-action in the statistical sense, as differentiatedpreviously in this review. That is to say that epi-genetic processes likely tell us something aboutthe biological mechanisms by which the envi-ronment can affect gene expression and impactbehavior, but they are not informative in termsof distinguishing between additive versus inter-active environmental effects.

Although variability exists in defining theterm, epigenetics generally refers to modifi-cations to the genome that do not involve achange in nucleotide sequence. To understandthis concept, let us review a bit about basicgenetics. The expression of a gene is influencedby transcription factors, which bind to specificsequences of DNA. It is through the binding oftranscription factors that genes can be turnedon or off. Epigenetic mechanisms involvechanges to how readily transcription factorscan access the DNA. Several different typesof epigenetic changes are known to exist thatinvolve different types of chemical changesthat can regulate DNA transcription. Oneepigenetic process that affects transcriptionbinding is DNA methylation. DNA methyla-tion involves the addition of a methyl group(CH3) onto a cytosine (one of the four basepairs that make up DNA). This leads to genesilencing because methylated DNA hindersthe binding of transcription factors. A secondmajor regulatory mechanism is related tothe configuration of DNA. DNA is wrappedaround clusters of histone proteins to form nu-cleosomes. Together the nucleosomes of DNAand histone are organized into chromatin.When the chromatin is tightly condensed, itis difficult for transcription factors to reach theDNA, and the gene is silenced. In contrast,when the chromatin is opened, the genecan be activated and expressed. Accordingly,modifications to the histone proteins thatform the core of the nucleosome can affectthe initiation of transcription by affecting howreadily transcription factors can access theDNA and bind to their appropriate sequence.


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Chromosome: asingle piece of coiledDNA containing manygenes, regulatoryelements, and othernucleotide sequences

Epigenetic modifications of the genomehave long been known to exist. For example,all cells in the body share the same DNA; ac-cordingly, there must be a mechanism wherebydifferent genes are active in liver cells than,for example, brain cells. The process of cellspecialization involves silencing certain por-tions of the genome in a manner specific toeach cell. DNA methylation is a mechanismknown to be involved in cell specialization.Another well-known example of DNA methy-lation involves X-inactivation in females. Be-cause females carry two copies of the X chromo-some, one must be inactivated. The silencing ofone copy of the X chromosome involves DNAmethylation. Genomic imprinting is anotherlong-established principle known to involveDNA methylation. In genomic imprinting theexpression of specific genes is determined bythe parent of origin. For example, the copy ofthe gene inherited from the mother is silenced,while the copy inherited from the father is ac-tive (or vice versa). The silent copy is inactivethrough processes involving DNA methylation.These changes all involve epigenetic processesparallel to those currently attracting so muchattention. However, the difference is that theseknown epigenetic modifications (cell specializa-tion, X inactivation, genomic imprinting) all oc-cur early in development and are stable. Thediscovery that epigenetic modifications con-tinue to occur across development, and can bereversible and more dynamic, has representeda major paradigm shift in our understanding ofenvironmental regulation of gene expression.

Animal studies have yielded compelling ev-idence that early environmental manipulationscan be associated with long-term effects thatpersist into adulthood. For example, maternallicking and grooming in rats is known tohave long-term influences on stress responseand cognitive performance in their offspring(Champagne et al. 2008, Meaney 2010). Fur-ther, a series of studies conducted in macaquemonkeys demonstrates that early rearingconditions can result in long-term increasedaggression, more reactive stress response,altered neurotransmitter functioning, and

structural brain changes (Stevens et al. 2009).These findings parallel research in humansthat suggests that early life experiences canhave long-term effects on child development(Loman & Gunnar 2010). Elegant work inanimal models suggests that epigenetic changesmay be involved in these associations (Meaney2010, Zhang & Meaney 2010).

Evaluating epigenetic changes in humansis more difficult because epigenetic marks canbe tissue specific. Access to human brain tissueis limited to postmortem studies of donatedbrains, which are generally unique and unrep-resentative samples and must be interpreted inthe context of those limitations. Nonetheless, arecent study of human brain samples from theQuebec Suicide Brain Bank found evidenceof increased DNA methylation of the exon1F promoter in hippocampal samples fromsuicide victims compared with controls—butonly if suicide was accompanied with a historyof childhood maltreatment (McGowan et al.2009). Importantly, this paralleled epigeneticchanges originally observed in rat brain inthe ortholog of this locus. Another line ofevidence suggesting epigenetic changes thatmay be relevant in humans is the observationof increasing discordance in epigenetic marksin MZ twins across time. This is significantbecause MZ twins have identical genotypes,and therefore, differences between them are at-tributed to environmental influences. In a studyby Fraga and colleagues (2005), MZ twins werefound to be epigenetically indistinguishableduring the early years of life, but older MZtwins exhibited remarkable differences in theirepigenetic profiles. These findings suggestthat epigenetic changes may be a mechanismby which environmental influences contributeto the differences in outcome observed fora variety of psychological traits of interestbetween genetically identical individuals.

The above studies complement a growingliterature demonstrating differences in gene ex-pression in humans as a function of environ-mental experience. One of the first studies toanalyze the relationship between social factorsand human gene expression compared healthy

398 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


older adults who differed in the extent to whichthey felt socially connected to others (Cole et al.2007). Using expression profiles obtained fromblood cells, a number of genes were identifiedthat showed systematically different levels of ex-pression in people who reported feeling lonelyand distant from others. Interestingly, these ef-fects were concentrated among genes that areinvolved in immune response. The results pro-vide a biological mechanism that could explainwhy socially isolated individuals show height-ened vulnerability to diseases and illnesses re-lated to immune function. Importantly, theydemonstrate that our social worlds can exertbiologically significant effects on gene expres-sion in humans (for a more extensive review, seeCole 2009).

CONCLUSIONS

This review has attempted to provide anoverview of the study of gene-environmentinteraction, starting with early animal studiesdocumenting gene-environment interaction, todemonstrations of similar effects in family,adoption, and twin studies. Advances in twinmodeling and the relative ease with which gene-environment interaction can now be modeledhas led to a significant increase in the numberof twin studies documenting changing impor-tance of genetic influence across environmentalcontexts. There is now widespread documen-tation of gene-environment interaction effectsacross many clinical disorders (Thapar et al.2007). These findings have led to more inte-grated etiological models of the development ofclinical outcomes. Further, since it is now rela-tively straightforward and inexpensive to collectDNA and conduct genotyping, there has beena surge of studies testing for gene-environmentinteraction with specific candidate genes. Psy-chologists have embraced the incorporation ofgenetic components into their studies, and ge-neticists who focus on gene finding are nowpaying attention to the environment in an un-precedented way. However, now that the ini-tial excitement surrounding gene-environmentinteraction has begun to wear off, a number

of challenges involved in the study of gene-environment interaction are being recognized.These include difficulties with interpreting in-teraction effects (or the lack thereof), due to is-sues surrounding the measurement and scalingof the environment, and statistical concerns sur-rounding modeling gene-environment interac-tions and the nature of their effects.

So where do we go from here? Individu-als who jumped on the gene-environment in-teraction bandwagon are now discovering thatstudying this process is harder than it first ap-peared. But there is good reason to believe thatgene-environment interaction is a very impor-tant process in the development of clinical dis-orders. So rather than abandon ship, I wouldsuggest that as a field, we just need to proceedwith more caution. I close with some thoughtsabout how we can move forward in this way.

Knowledge Is Power

As every student enrolled in Psychology 101 istaught, simply making an individual aware ofsomething (such as their being in a psychologystudy) can change behavior. More widespreadrecognition of the issues surrounding thestudy of gene-environment interactions, asdelineated above, will hopefully lead to morethoughtful and careful evaluations of hypoth-esized gene-environment interaction effects.Studying gene-environment interactions is notas simple as plugging a genotype and environ-ment into a regression equation and seeing ifthe interaction term yields p < 0.05. Scientistsconducting gene-environment interactionresearch and reviewers of papers evaluatinggene-environment interaction effects needto be keenly aware of these issues so thatdue diligence can be carried out to evaluateinteraction effects that are detected (or not) ina given sample.

Use What We Already Know from theTwin and Developmental Literatures

It has been suggested that twin studies areno longer necessary in the era of molecular


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


genetics. The argument is that there is nolonger any need to infer genetic influencenow that we can directly measure genotypes.However, studies of latent genetic influence(as inferred from family, adoption, and twinstudies) and studies of measured genotypesactually yield very different information andhave complementary strengths and weaknesses,as has been nicely reviewed by my colleagueKenneth Kendler in previous papers (Kendler2005, 2010). Information about aggregate ge-netic risk, as yielded by twin studies, gives us anidea of the big picture. It is essentially a satellitepicture, providing an overview of the generallandscape. On the other hand, molecular genet-ics offers a level of detail about the underlyingbiology that twin studies cannot. The corre-sponding metaphor would be the photographeron the ground who is taking pictures of the in-dividual rocks and trees. But lost in that level ofdetail is information about the overall picture.

Perhaps because of the aggregate natureof genetic influence as studied in twin de-signs, findings from twin studies have a far bet-ter record of replication than findings fromspecific candidate gene association studies.Heritabilities for clinical disorders have beenremarkably consistent across populations, andgene-environment interaction effects have alsoyielded consistent results (e.g., (Legrand et al.2007, Rose et al. 2001). This is not the casewith gene finding efforts, where replicationshave been notoriously difficult, both for maineffects (e.g., Lind et al. 2009, Wang et al. 2004)and gene-environment interactions (Caspi et al.2003, Risch et al. 2009).

Parallel to the way that evidence for heri-tability from twin studies for a given outcomewas used to justify searching for specific genesinvolved in that outcome, evidence for gene-environment interaction from twin studiescan also be used to develop hypotheses to testgene-environment interactions associated withspecific, identified genes. Change in the overallheritability across environmental contexts doesnot necessarily dictate that any one specificsusceptibility gene will operate in a parallelmanner; however, a change in heritability

suggests that at least a good portion of theinvolved genes (assuming many genes ofapproximately equal and small effect) must beoperating in that manner for a difference inheritability by environment to be detected. Inthis sense, you are loading the dice when youtest for specific candidate gene-environmentinteraction effects with an environment that hasalready been shown to moderate the overall im-portance of genetic influences on that outcome.This is the strategy we have used to furthercharacterize the risk related to genes associatedwith alcohol dependence in the CollaborativeStudy on the Genetics of Alcoholism. Onthe basis of our twin studies suggesting thatgenetic influences on adolescent substance useare moderated by parental monitoring (Dicket al. 2007c) and peer substance use (Dicket al. 2007a), we tested for moderation of theassociation of GABRA2 (Edenberg et al. 2004)and CHRM2 (Wang et al. 2004) as a function ofparental monitoring and peer group antisocialbehavior, respectively. We found evidencefor gene-environment interaction effects inthe direction predicted by the twin studies,namely, genetic effects were enhanced underconditions of lower parental monitoring (Dicket al. 2009) and higher peer group antisocialbehavior (Latendresse et al. 2010).

Twin studies are not the only place fromwhich to draw hypotheses about environmentalinfluences that are likely to moderate geneticeffects. The developmental literature containsa wealth of studies demonstrating differentialeffects of the environment across childrenwith differing temperaments and/or whodiffer on family history. Because temperamentand family history both provide informationabout the child’s genetic predisposition, thesekinds of interactions can also serve as startingpoints for developing hypotheses about gene-environment interaction effects associatedwith specific genes. In addition to the twinevidence suggesting that parental monitoringmoderated the importance of genetic effects,numerous studies in the developmental liter-ature suggest the importance of this constructin moderating associations between early

400 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


temperament/family history and the subse-quent development of child behavior problems.For example, Bates and colleagues found thatacross two independent samples, a difficultchildhood temperament was related to thesubsequent development of externalizing be-havior, but only in the context of lower parentalcontrol (Bates et al. 1998). Further, Molina andcolleagues have found that density of familyhistory of alcoholism is related to the develop-ment of behavior problems in children, but onlyin the context of poor parenting (a measure thatincluded reduced parental monitoring) (Molinaet al. 2010). These studies both find that as-sociations between predisposing factors (bothknown to at least partially reflect genetic influ-ence) and child behavior problems are strongerunder conditions of lower parental monitoring,paralleling the finding from twin studiesthat genetic influences were stronger underconditions of lower parental monitoring. Theyprovide a compelling rationale to study parentalmonitoring as a moderator of the effects asso-ciated with specific candidate genes involvedin substance use and externalizing behavior, aneffect which has now been demonstrated withrespect to GABRA2 (Dick et al. 2009).

Statistics Are Not a Substitutefor Critical Thinking

One of the take-home messages from thisreview is that interpreting results fromgene-environment interaction studies is notstraightforward. There are valid reasons why“real” gene-environment interaction effectsmight not be detected, why results might varyacross studies, or why our statistics might yield“significant evidence” for gene-environmentinteraction when none really exists. An uncrit-ical tally of whether specific gene-environmentinteraction effects replicate or not, withoutpaying attention to issues such as the mea-surement of the outcome, measurement of theenvironment, statistics employed, samplingstrategy, and sample size, across studies willundoubtedly lead to mixed results for any givengene-environment interaction effect in the

literature. Statistics such as meta-analyses havebecome very popular in genetics, largely due torecognition of the incredibly large sample sizesit will take to identify genes of small effect.Meta-analytic techniques are particularlyappropriate in this area, where genotypic datacan be standardized across studies, and out-comes are often measured using standardizedassessments (e.g., DSM diagnoses as assessedin structured clinical interviews). However,in the area of gene-environment interactions,where studies often have very different designsand goals, some of which include the explicitattempt to explore the boundaries of origi-nally reported gene-environment interactions,meta-analyses should not be conducted withoutcareful attention to these issues.

Play Nice in the Sandbox

Research on gene-environment interactionsis inherently interdisciplinary; it sits at theintersection between genetics and psychol-ogy. However, this perspective has not beenembraced to the extent that it could be, and(in my opinion) must be, in order to doreally good research in this area. Most ofthe gene-environment interaction researchin psychology has been limited to the “usualsuspects”—purportedly functional polymor-phisms in MAOA, 5-HTT, DRD4, and afew others (Belsky et al. 2009). However,the evidence for those polymorphisms trulybeing functional is often ambiguous (Cirulli &Goldstein 2007). In addition, in the field ofgenetics, we would never test a single markerin a gene in order to make conclusions aboutthe relevance of that gene in a given genotype.Rather, with data from the human genomeproject and the HapMap project, we now knowsomething about the structure of most genesin the human genome (Manolio et al. 2008).Further, there are many polymorphic markersavailable across most genes of interest. It ispossible that multiple locations in a gene couldhave various forms that lead to differentialfunction of that gene contributing to differ-ential susceptibility to an outcome (McClellan


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


& King 2010). Today it would be nearlyimpossible to publish a paper in a geneticsjournal without saying something about thecoverage of the gene provided by the set ofgenotyped markers (Pettersson et al. 2009).

As one example, the Taq1A allele, originallythought to be in the DRD2 gene, has an exten-sive literature surrounding it, with reported as-sociations with a number of phenotypes relatedto substance use, smoking, and a variety of otherphenotypes related to impulsivity (Dick et al.2007d, Noble 2000). Figure 2 shows a screen-shot of output from the program Haploview(Barrett et al. 2005) illustrating the linkage dise-quilibrium (LD) structure of the chromosomalregion surrounding the DRD2 gene. Along thetop of the figure is the base pair position of thechromosomal region, as listed in kilobases togive an idea of scale. Directly underneath, thetriangles indicate SNPs that were genotyped inthe samples on which the Haploview data arebased. The SNP highlighted in green is themarker commonly referred to as DRD2 Taq1A.Beneath the SNPs, the genes in the region arelisted. The length of the line reflects the lengthof the gene. One will note that the Taq1A alleleis actually located in a small gene next to DRD2called ANKK1. It is not located in DRD2 de-spite the large literature making claims aboutwhether DRD2 was involved in many differentphenotypes of interest based on genotyping atthis marker. Below the genes is the LD plot,where shading indicates the degree of correla-tion between markers (shown here as the smallhash marks at the top of the figure) as measuredby D′ (Hedrick & Kumar 2001), with darkerred shading indicating higher correlations; blueor white shading indicates the markers areunlinked or uncorrelated. What stands out isthe block-like correlational structure, yield-ing inverted red triangles that indicate groupsof SNPs where there is high LD across thatgroup of SNPs and low LD with surroundingSNPs located outside the block. This block-likestructure is observed throughout the genome(Gabriel et al. 2002). Knowing the correlationpattern is critical for the selection of markers

and the interpretation of genetic association re-sults. For example, DRD2 spans at least twoblocks on the figure. If different studies geno-typed SNPs from different blocks, they couldreach different conclusions about whetherDRD2 was “associated” with outcome, depend-ing on the location of the marker they chose andwhere the actual associated SNP was. Further,also note that the LD block that contains Taq1Aspans the genes DRD2, ANKK1, and TTC12.This makes it very difficult to know which geneis actually important for an observed associa-tion. More extensive genotyping across thesegenes and the other gene located very nearbyin the region, NCAM1, has suggested that theassociation with substance-use phenotypes ex-tends to multiple genes in this region (Dick et al.2007d, Gelernter et al. 2006, Yang et al. 2008).This underscores the necessity of understand-ing genomic structure in order to evaluate therole of hypothesized genes of interest.

In the same way that psychologists pay care-ful attention to the measurement of their out-comes of interest and potential environmen-tal factors of relevance, care must be taken incharacterizing genes of interest. The genet-ics research being carried out by psychologistsshould be of the same caliber as that being con-ducted in other areas of genetics, and it mustkeep up with the rapid advances going on in thatfield. Otherwise it will not be taken seriously.This does not mean that all psychologists needto be “gene-finders” or to carry out GWAS. Butit does mean that anyone involved in this kindof research should understand the complexitiesof studying genetics and be connected to thelatest developments in genetics. Because of therapid pace at which the field of genetics moves,this necessitates having collaborators who aretied more centrally to the world of geneticsand/or (for the younger generation of psycholo-gists with interest in this area) to obtain focusedtraining in genetics, ideally through a post-doctoral training experience. With ∼25,000genes in the human genome, thousands of ge-netic association papers published, and GWASpapers being turned out every day, I find it

402 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


impossible to believe that the handful of usualsuspects are the only genes of interest for clini-cal outcomes. As large-scale gene-finding stud-ies continue to report associations with new andnovel genes of interest, these genes too deservefurther study by psychologists to delineate howassociated risk may be modified by environmen-tal factors.

The bridge between psychology and genet-ics is not a one-way street. Although psycholo-gists have much to learn from geneticists, theyalso have much to offer. As geneticists havegrown interested in incorporating environmen-tal information into genetic studies, they havebeen guilty of using environmental measuresthat would be considered naive by psychologistswho are devoted to careful characterization ofthese constructs. Much of the large-scale geneidentification work to date has been dominated

by studies of binary diagnostic outcomes. Al-though the use of binary diagnoses brings theadvantage of standardized, reliable assessmentsacross studies and sites, there is reason to be-lieve that these phenotypes are not ideal forgene finding. Psychologists have a long historyof careful measurement of phenotype and ofstudying intermediate phenotypes and mecha-nistic processes. The application of these skillsto the field of genetics holds great promise bothto aid in gene identification and to help with thecharacterization of risk associated with identi-fied genes. However, this promise will only berecognized when psychologists and geneticistswork closely together, have patience with oneanother concerning differences in training andideology, and respect the relative contributionsthat each field can bring to the study of gene-environment interaction.

SUMMARY POINTS

1. Gene-environment interaction refers to the phenomenon whereby the effect of genesdepends on the environment, or the effect of the environment depends on genotype.There is now widespread documentation of gene-environment interaction effects acrossmany clinical disorders, leading to more integrated etiological models of the developmentof clinical outcomes.

2. Twin, family, and adoption studies provide methods to study gene-environment interac-tion with genetic effects modeled latently, meaning that genes are not directly measured,but rather genetic influence is inferred based on correlations across relatives. Advancesin genotyping technology have contributed to a proliferation of studies testing for gene-environment interaction with specific measured genes. Each of these designs has its ownstrengths and limitations.

3. Two types of gene-environment interaction have been discussed in greatest detail in theliterature: fan-shaped interactions, in which the influence of genotype is greater in oneenvironmental context than in another; and crossover interactions, in which the sameindividuals who are most adversely affected by negative environments may also be thosewho are most likely to benefit from positive environments. Distinguishing between thesetypes of interactions poses a number of challenges.

4. The range of environments studied and the lack of a true metric for many environmentalmeasures of interest create difficulties for studying gene-environment interactions. Is-sues surrounding power, and the use of logistic regression and selected samples, furthercompound the difficulty of studying gene-environment interactions. These issues havenot received adequate attention by many researchers in this field.


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


5. Epigenetic processes may tell us something about the biological mechanisms by whichthe environment can affect gene expression and impact behavior. The growing literaturedemonstrating differences in gene expression in humans as a function of environmentalexperience demonstrates that our social worlds can exert biologically significant effectson gene expression in humans.

6. Much of the current work on gene-environment interactions does not take advantageof the state of the science in genetics or psychology; advancing this area of study willrequire close collaborations between psychologists and geneticists.

DISCLOSURE STATEMENT

The author is not aware of any affiliations, memberships, funding, or financial holdings that mightbe perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

This manuscript was prepared with support by AA15416 and K02AA018755 from the NationalInstitute of Alcohol Abuse and Alcoholism. The author also acknowledges helpful comments fromDr. Kenneth Kendler on an earlier version of this manuscript.

LITERATURE CITED

Bakermans-Kranenburg MJ, van Ijzendoorn MH. 2006. Gene-environment interaction of the dopamine D4receptor (DRD4) and observed maternal insensitivity predicting externalizing behavior in preschoolers.Dev. Psychobiol. 48(5):406–9

Barrett JC, Fry B, Maller J, Daly MJ. 2005. Haploview: analysis and visualization of LD and haplotype maps.Bioinformatics 21:263–65

Bates JE, Pettit GS, Dodge KA, Ridge B. 1998. Interaction of temperamental resistance to control andrestrictive parenting in the development of externalizing behavior. Dev. Psychol. 34:982–95

Belsky J, Jonassaint C, Pluess M, Stanton M, Brummett B, Williams R. 2009. Vulnerability genes or plasticitygenes? Mol. Psychiatry 14(8):746–54

Bergeman CS, Plomin R, McClearn GE, Pedersen NL, Friberg LT. 1988. Genotype-environment interactionin personality development: identical twins reared apart. Psychol. Aging 3:399–406

Boardman JD, Blalock CL, Pampel FC. 2010. Trends in genetic influences on smoking. J. Health Soc. Behav.51:108–23

Boyce WT, Ellis BJ. 2005. Biological sensitivity to context: I. An evolutionary-developmental theory of theorigins and functions of stress reactivity. Dev. Psychopathol. 17(2):271–301

Button TM, Corley RP, Rhee SH, Hewitt JK, Young SE, Stallings MC. 2007. Delinquent peer affiliation andconduct problems: a twin study. J. Abnorm. Psychol. 116(3):554–64

Button TM, Lau JY, Maughan B, Eley TC. 2008. Parental punitive discipline, negative life events and gene-environment interplay in the development of externalizing behavior. Psychol. Med. 38(1):29–39

Cadoret RJ, Troughton E, Merchant LM, Whitters A. 1990. Early life psychosocial events and adult affectivesymptoms. In Robbins & Rutter 1990, pp. 300–13

Cadoret RJ, Winokur G, Langbehn D, Troughton E, Yates WR, Stewart MA. 1996. Depression spectrumdisease, I: the role of gene-environment interaction. Am. J. Psychiatry 153:892–99

Cannon TD, Mednick SA, Parnas J. 1990. Two pathways to schizophrenia in children at risk. In Robbins &Rutter 1990, pp. 314–27

404 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Cannon TD, Mednick SA, Parnas J, Schulsinger F, Praestholm J, Vestergaard A. 1993. Developmental brainabnormalities in the offspring of schizophrenic mothers. I. Contributions of genetic and perinatal factors.Arch. Gen. Psychiatry 50:551–64

Cardon LR. 2006. Genetics. Delivering new disease genes. Science 314(5804):1403–5Caspi A, Hariri AR, Holmes A, Uher R, Moffitt TE. 2010. Genetic sensitivity to the environment: the case

of the serotonin transporter gene and its implications for studying complex diseases and traits. Am. J.Psychiatry 167(5):509–27

Caspi A, McClay J, Moffitt TE, Mill J, Martin J, et al. 2002. Role of genotype in the cycle of violence inmaltreated children. Science 297:851–54

Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, et al. 2003. Influence of life stress on depression:moderation by a polymorphism in the 5-HTT gene. Science 301:386–89

Champagne DL, Bagot RC, van HF, Ramakers G, Meaney MJ, et al. 2008. Maternal care and hippocam-pal plasticity: evidence for experience-dependent structural plasticity, altered synaptic functioning, anddifferential responsiveness to glucocorticoids and stress. J. Neurosci. 28(23):6037–45

Cirulli ET, Goldstein DB. 2007. In vitro assays fail to predict in vivo effects of regulatory polymorphisms.Hum. Mol. Genet. 16(16):1931–39

Cloninger CR, Bohman M, Sigvardsson S. 1981. Inheritance of alcohol abuse: cross-fostering analysis ofadopted men. Arch. Gen. Psychiatry 38:861–68

Cole SW. 2009. Social regulation of human gene expression. Curr. Dir. Psychol. Sci. 18:132–37Cole SW, Hawkley LC, Arevalo JM, Sung CY, Rose RM, Cacioppo JT. 2007. Social regulation of gene

expression in human leukocytes. Genome Biol. 8(9):R189Cooper RM, Zubek JP. 1958. Effects of enriched and restricted early environments on the learning ability of

bright and dull rats. Can. J. Psychol. 12:159–64Dick DM, Latendresse SJ, Lansford JE, Budde J, Goate A, et al. 2009. The role of GABRA2 in trajectories

of externalizing behavior across development and evidence of moderation by parental monitoring. Arch.Gen. Psychiatry 66:649–57

Dick DM, Meyers JL, Latendresse SJ, Creemers HE, Lansford JE, et al. 2011. CHRM2, parental monitoring,and adolescent externalizing behavior: evidence for gene-environment interaction. Psychol. Sci. In press

Dick DM, Pagan JL, Holliday C, Viken R, Pulkkinen L, et al. 2007a. Gender differences in friends’ influenceson adolescent drinking: a genetic epidemiological study. Alcohol. Clin. Exp. Res. 31:2012–19

Dick DM, Pagan JL, Viken R, Purcell S, Kaprio J, et al. 2007b. Changing environmental influences onsubstance use across development. Twin Res. Hum. Genet. 10(2):315–26

Dick DM, Rose RJ, Viken RJ, Kaprio J, Koskenvuo M. 2001. Exploring gene-environment interactions:socioregional moderation of alcohol use. J. Abnorm. Psychol. 110:625–32

Dick DM, Viken R, Purcell S, Kaprio J, Pulkkinen L, Rose RJ. 2007c. Parental monitoring moderates the im-portance of genetic and environmental influences on adolescent smoking. J. Abnorm. Psychol. 116(1):213–18

Dick DM, Wang J, Plunkett J, Aliev F, Hinrichs AL, et al. 2007d. Family-based association analyses of alcoholdependence phenotypes across DRD2 and neighboring gene ANKK1. Alcohol. Clin. Exp. Res. 31:1645–53

Eaves LJ. 2006. Genotype x environment interaction in psychopathology: fact or artifact? Twin Res. Hum.Genet. 9(1):1–8

Edenberg HJ, Dick DM, Xuei X, Tian H, Almasy L, et al. 2004. Variations in GABRA2, encoding the α2subunit of the GABA-A receptor are associated with alcohol dependence and with brain oscillations. Am.J. Hum. Genet. 74:705–14

Edwards AC, Dodge KA, Latendresse SJ, Lansford JE, Bates J, et al. 2010. MAOA uVNTR and early physicaldiscipline interact to influence delinquent behavior. J. Child Psychol. Psychiatry 51(6):679–87

Ellis BJ, Boyce WT. 2008. Biological sensitivity to context. Curr. Dir. Psychol. Sci. 17(3):183–87Ellis BJ, Essex MJ, Boyce WT. 2005. Biological sensitivity to context: II. Empirical explorations of an

evolutionary-developmental theory. Dev. Psychopathol. 17(2):303–28Engelman CD, Baurley JW, Chiu YF, Joubert BR, Lewinger JP, et al. 2009. Detecting gene-environment

interactions in genome-wide association data. Genet. Epidemiol. 33(Suppl. 1):S68–73Enoch MA, Steer CD, Newman TK, Gibson N, Goldman D. 2010. Early life stress, MAOA, and gene-

environment interactions predict behavioral disinhibition in children. Genes Brain Behav. 9(1):65–74


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Feinberg ME, Button TM, Neiderhiser JM, Reiss D, Hetherington EM. 2007. Parenting and adolescentantisocial behavior and depression: evidence of genotype x parenting environment interaction. Arch. Gen.Psychiatry 64(4):457–65

Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, et al. 2005. Epigenetic differences arise during the lifetimeof monozygotic twins. Proc. Natl. Acad. Sci. USA 102(30):10604–9

Frazzetto G, Di LG, Carola V, Proietti L, Sokolowska E, et al. 2007. Early trauma and increased risk forphysical aggression during adulthood: the moderating role of MAOA genotype. PLoS One 2(5):e486

Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, et al. 2002. The structure of haplotype blocks in thehuman genome. Science 296(5576):2225–29

Gauderman WJ. 2002. Sample size requirements for matched case-control studies of gene-environment in-teraction. Stat. Med. 21:35–50

Gauderman WJ, Thomas DC, Murcray CE, Conti D, Li D, Lewinger JP. 2010. Efficient genome-wideassociation testing of gene-environment interaction in case-parent trios. Am. J. Epidemiol. 172(1):116–22

Gelernter J, Yu Y, Weiss R, Brady K, Panhuysen C, et al. 2006. Haplotype spanning TTC12 and ANKK1,flanked by the DRD2 and NCAM1 loci, is strongly associated to nicotine dependence in two distinctAmerican populations. Hum. Mol. Genet. 15:3498–507

Gottesman II. 1991. Schizophrenia Genesis: The Origins of Madness. New York: FreemanHeath AC, Eaves LJ, Martin NG. 1998. Interaction of marital status and genetic risk for symptoms of depres-

sion. Twin Res. 1(3):119–22Heath AC, Jardine R, Martin NG. 1989. Interactive effects of genotype and social environment on alcohol

consumption in female twins. J. Stud. Alcohol 50:38–48Hedrick P, Kumar S. 2001. Mutation and linkage disequilibrium in human mtDNA. Eur. J. Hum. Genet.

9:969–72Henderson ND. 1970. Genetic influences on the behavior of mice can be obscured by laboratory rearing.

J. Comp. Physiol. Psychol. 72:505–11Henderson ND. 1972. Relative effects of early rearing environment and genotype on discrimination learning

in house mice. J. Comp. Physiol. Psychol. 79:243–53Heston LL, Denney D. 1967. Interactions between early life experience and biological factors in schizophrenia.

In The Transmission of Schizophrenia, ed. D Rosenthal, SS Kety, pp. 363–76. Oxford: PergamonHewitt JK, Turner JR. 1995. Behavior genetic approaches in behavioral medicine: an introduction. In Behavior

Genetic Approaches in Behavioral Medicine, ed. JR Turner, LR Cardon, JK Hewitt, pp. 3–16. New York:Plenum

Kendler KS. 1993. Twin studies of psychiatric illness: current status and future directions. Arch. Gen. Psychiatry50:905–15

Kendler KS. 2005. Psychiatric genetics: a methodologic critique. Am. J. Psychiatry 162:3–11Kendler KS. 2011. A conceptual overview of gene-environment interaction and correlation in a developmental

context. In The Dynamic Genome and Mental Health, ed. KS Kendler, S Jaffee, D Romer. New York: OxfordUniv. Press. In press

Kendler KS, Baker JH. 2007. Genetic influences on measures of the environment: a systematic review. Psychol.Med. 37:615–26

Kendler KS, Eaves LJ. 1986. Models for the joint effect of genotype and environment on liability to psychiatricillness. Am. J. Psychiatry 143:279–89

Kendler KS, Neale MC, Heath AC, Kessler RC, Eaves LJ. 1991. Life events and depressive symptoms: a twinstudy perspective. In The New Genetics of Mental Illness, ed. P McGuffin, R Murray, pp. 146–64. Oxford:Butterworth-Heinemann

Kendler KS, Neale MC, Kessler R, Heath A, Eaves L. 1993. A twin study of recent life events and difficulties.Arch. Gen. Psychiatry 50:789–96

Kim-Cohen J, Caspi A, Taylor A, Williams B, Newcombe R, et al. 2006. MAOA, maltreatment, and gene-environment interaction predicting children’s mental health: new evidence and a meta-analysis. Mol.Psychiatry 11(10):903–13

Koopmans JR, Slutske WS, van Baal GC, Boomsma DI. 1999. The influence of religion on alcohol useinitiation: evidence for genotype X environment interaction. Behav. Genet. 29(6):445–53

406 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Landi MT, Chatterjee N, Yu K, Goldin LR, Goldstein AM, et al. 2009. A genome-wide association studyof lung cancer identifies a region of chromosome 5p15 associated with risk for adenocarcinoma. Am. J.Hum. Genet. 85(5):679–91

Latendresse SJ, Bates J, Goodnight JA, Dodge KA, Lansford JE, et al. 2010. Differential susceptibility toadolescent externalizing trajectories: examining the interplay between CHRM2 and peer group antisocialbehavior. Child Dev. DOI: 10.1111/j.1467-8624.2010.01517.x

Legrand LN, Keyes M, McGue M, Iacono WG, Krueger RF. 2008. Rural environments reduce the geneticinfluence on adolescent substance use and rule-breaking behavior. Psychol. Med. 38(9):1341–50

Leve LD, Neiderhiser JM, Scaramella LV, Reiss D. 2010. The early growth and development study: usingthe prospective adoption design to examine genotype-environment interplay. Behav. Genet. 40:306–14

Lind PA, Luciano M, Horan MA, Marioni RE, Wright MJ, et al. 2009. No association between cholinergicmuscarinic receptor 2 (CHRM2) genetic variation and cognitive abilities in three independent samples.Behav. Genet. 39(5):513–23

Lindgren CM, Heid IM, Randall JC, Lamina C, Steinthorsdottir V, et al. 2009. Genome-wide association scanmeta-analysis identifies three loci influencing adiposity and fat distribution. PLoS Genet. 5(6):e1000508

Loman MM, Gunnar MR. 2010. Early experience and the development of stress reactivity and regulation inchildren. Neurosci. Biobehav. Rev. 34(6):867–76

Malaspina D, Sohler NL, Susser ES. 1999. Interaction of genes and prenatal exposures in schizophrenia. InPrenatal Exposures in Schizophrenia, ed. ES Susser, AS Brown, JM Gorman, pp. 35–59. Washington, DC:Am. Psychiatr. Press

Manolio TA, Brooks LD, Collins FS. 2008. A HapMap harvest of insights into the genetics of common disease.J. Clin. Invest. 118(5):1590–605

Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, et al. 2009. Finding the missing heritability ofcomplex diseases. Nature 461(7265):747–53

Martin N, Boomsma D, Machin G. 1997. A twin-pronged attack on complex traits. Nat. Genet. 17:387–92Mather K, Jinks J. 1982. Biometrical Genetics. London: Chapman & HallMcClellan J, King MC. 2010. Genetic heterogeneity in human disease. Cell 141(2):210–17McDermott R, Tingley D, Cowden J, Frazzetto G, Johnson DD. 2009. Monoamine oxidase A gene (MAOA)

predicts behavioral aggression following provocation. Proc. Natl. Acad. Sci. USA 106(7):2118–23McGowan PO, Sasaki A, D’Alessio AC, Dymov S, Labonte B, et al. 2009. Epigenetic regulation of the

glucocorticoid receptor in human brain associates with childhood abuse. Nat. Neurosci. 12(3):342–48McGue M, Bouchard TJ Jr. 1998. Genetic and environmental influences on human behavioral differences.

Annu. Rev. Neurosci. 21:1–24McGue M, Sharma A, Benson P. 1995. Parent and sibling influences on adolescent alcohol use and misuse:

evidence from a US adoption cohort. J. Stud. Alcohol 57:8–18McGue M, Sharma A, Benson P. 1996. The effect of common rearing on adolescent adjustment: evidence

from a U.S. adoption cohort. Dev. Psychol. 32(6):604–13McKenna AH, Hanna M, Banks E, Sivachenko A, Cibulskis K, et al. 2010. The Genome Analysis Toolkit: a

MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20(9):1297–303Meaney MJ. 2010. Epigenetics and the biological definition of gene x environment interactions. Child Dev.

81(1):41–79Moldin SO, Gottesman II. 1997. At issue: genes, experience, and chance in schizophrenia—positioning for

the 21st century. Schizophr. Bull. 23:547–61Molina BSG, Donovan JE, Belendiuk KA. 2010. Familial loading for alcoholism and offspring behavior:

mediating and moderating influences. Alcohol. Clin. Exp. Res. 34(11):1972–84Neale BM, Ferreira MAR, Medland SE, Posthuma D. 2008. Statistical Genetics: Gene Mapping Through Linkage

and Association. New York: Taylor & FrancisNeale MC. 2000. The use of Mx for association and linkage analyses. GeneScreen 1:107–11Noble EP. 2000. The DRD2 gene in psychiatric and neurological disorders and its phenotypes. Pharmacoge-

netics 1:309–33Pettersson FH, Anderson CA, Clarke GM, Barrett JC, Cardon LR, et al. 2009. Marker selection for genetic

case-control association studies. Nat. Protoc. 4(5):743–52


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Plomin R, DeFries JC, Loehlin JC. 1977. Genotype-environment interaction and correlation in the analysisof human behavior. Psychol. Bull. 84:309–22

Plomin R, Hershberger S. 1991. Genotype-environment interaction. In Conceptualization and Measurement ofOrganism-Environment Interaction, ed. TD Wachs, R Plomin, pp. 29–43. Washington, DC: Am. Psychol.Assoc.

Prom-Wormley EC, Eaves LJ, Foley DL, Gardner CO, Archer KJ, et al. 2009. Monoamine oxidase A andchildhood adversity as risk factors for conduct disorder in females. Psychol. Med. 39(4):579–90

Psychiatr. GWAS Consort. Steer. Comm. 2008. A framework for interpreting genome-wide association studiesof psychiatric disorders. Mol. Psychiatry 14:10–17

Purcell S. 2002. Variance components models for gene-environment interaction in twin analysis. Twin Res.5:554–71

Purcell S, Cherny SS, Sham PC. 2003. Genetic power calculator: design of linkage and association geneticmapping studies of complex traits. Bioinformatics 19:149–50

Risch N, Herrell R, Lehner T, Liang KY, Eaves L, et al. 2009. Interaction between the serotonin transportergene (5-HTTLPR), stressful life events, and risk of depression: a meta-analysis. JAMA 301(23):2462–71

Robins LN, Rutter M, eds. 1990. Straight and Devious Pathways from Childhood to Adulthood. London: CambridgeUniv. Press

Rose RJ, Dick DM, Viken RJ, Kaprio J. 2001. Gene-environment interaction in patterns of adolescent drinking:regional residency moderates longitudinal influences on alcohol use. Alcohol. Clin. Exp. Res. 25:637–43

Scarr S, McCartney K. 1983. How people make their own environments: a theory of genotype-environmenteffects. Child Dev. 54:424–35

Shanahan MJ, Hofer SM. 2005. Social context in gene-environment interactions: retrospect and prospect.J. Gerontol. B Psychol. Sci. Soc. Sci. 60(Spec. No. 1):65–76

Sherman SL, DeFries JC, Gottesman II, Loehlin JC, Meyer JM, et al. 1997. Recent developments in humanbehavioral genetics: past accomplishments and future directions. Am. J. Hum. Genet. 60:1265–75

Stevens HE, Leckman JF, Coplan JD, Suomi SJ. 2009. Risk and resilience: early manipulation of macaquesocial experience and persistent behavioral and neurophysiological outcomes. J. Am. Acad. Child Adolesc.Psychiatry 48(2):114–27

Taylor SE, Way BM, Welch WT, Hilmert CJ, Lehman BJ, Eisenberger NI. 2006. Early family environment,current adversity, the serotonin transporter promoter polymorphism, and depressive symptomatology.Biol. Psychiatry 60(7):671–76

Thapar A, Harold G, Rice F, Langley K, O’Donovan M. 2007. The contribution of gene-environment inter-action to psychopathology. Dev. Psychopathol. 19(4):989–1004

Tienari P, Lahti I, Sorri A, Naarala M, Moring J, et al. 1990. Adopted-away offspring of schizophrenics andcontrols: the Finnish Adoptive Family Study of Schizophrenia. In Robins & Rutter 1990, pp. 365–79

Van Den Oord EJ. 2008. Controlling false discoveries in genetic studies. Am. J. Med. Genet. B Neuropsychiatr.Genet. 147B(5):637–44

Vanyukov MM, Maher BS, Devlin B, Kirillova GP, Kirisci L, et al. 2007. The MAOA promoter polymorphism,disruptive behavior disorders, and early onset substance use disorder: gene-environment interaction.Psychiatr. Genet. 17(6):323–32

Wahlberg K, Wynne LC, Oja H, Keskitalo P, Pykalainen L, et al. 1997. Gene-environment interaction invulnerability to schizophrenia: findings from the Finnish Adoptive Family Study of Schizophrenia. Am.J. Psychiatry 154:355–62

Wahlsten D, Metten P, Phillips TJ, Boehm SL, Burkhart-Kasch S, et al. 2003. Different data from differentlabs: lessons from studies of gene-environment interaction. J. Neurobiol. 54(1):283–311

Wang JC, Hinrichs AL, Stock H, Budde J, Allen R, et al. 2004. Evidence of common and specific geneticeffects: association of the muscarinic acetylcholine receptor M2 (CHRM2) gene with alcohol dependenceand major depressive syndrome. Hum. Mol. Genet. 13:1903–11

Weder N, Yang BZ, Douglas-Palumberi H, Massey J, Krystal JH, et al. 2009. MAOA genotype, maltreatment,and aggressive behavior: the changing impact of genotype at varying levels of trauma. Biol. Psychiatry65(5):417–24

408 Dick

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Yang BZ, Kranzler HR, Zhao H, Gruen JR, Luo X, Gelernter J. 2008. Haplotypic variants in DRD2, ANKK1,TTC12, and NCAM1 are associated with comorbid alcohol and drug dependence. Alcohol. Clin. Exp. Res.32(12):2117–27

Zhang TY, Meaney MJ. 2010. Epigenetics and the environmental regulation of the genome and its function.Annu. Rev. Psychol. 61:439–66


Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.


Prob

abili

ty o

f dis

orde

r

Prob

abili

ty o

f dis

orde

r

Prob

abili

ty o

f dis

orde

r2 copies of risk allele

0 copies of risk allele1 copy of risk allele

–2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2

a b

c

Environment risk

Environment risk

Environment risk

Figure 1Different types of gene-environment interactions.

Figure 2Screenshot from Haploview (Barrett et al. 2005) showing the linkage disequilibrium structure surroundingthe DRD2 gene.

www.annualreviews.org • Gene-Environment Interaction C-1

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.

CP07-FrontMatter ARI 8 March 2011 4:13

Annual Review ofClinical Psychology

Volume 7, 2011 Contents

The Origins and Current Status of Behavioral Activation Treatmentsfor DepressionSona Dimidjian, Manuel Barrera Jr., Christopher Martell, Ricardo F. Munoz,

and Peter M. Lewinsohn � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Animal Models of Neuropsychiatric DisordersA.B.P. Fernando and T.W. Robbins � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �39

Diffusion Imaging, White Matter, and PsychopathologyMoriah E. Thomason and Paul M. Thompson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �63

Outcome Measures for PracticeJason L. Whipple and Michael J. Lambert � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �87

Brain Graphs: Graphical Models of the Human Brain ConnectomeEdward T. Bullmore and Danielle S. Bassett � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 113

Open, Aware, and Active: Contextual Approaches as an EmergingTrend in the Behavioral and Cognitive TherapiesSteven C. Hayes, Matthieu Villatte, Michael Levin, and Mikaela Hildebrandt � � � � � � � � 141

The Economic Analysis of Prevention in Mental Health ProgramsCathrine Mihalopoulos, Theo Vos, Jane Pirkis, and Rob Carter � � � � � � � � � � � � � � � � � � � � � � � � � 169

The Nature and Significance of Memory Disturbance in PosttraumaticStress DisorderChris R. Brewin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 203

Treatment of Obsessive Compulsive DisorderMartin E. Franklin and Edna B. Foa � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 229

Acute Stress Disorder RevisitedEtzel Cardena and Eve Carlson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 245

Personality and Depression: Explanatory Models and Reviewof the EvidenceDaniel N. Klein, Roman Kotov, and Sara J. Bufferd � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 269

vi

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.

CP07-FrontMatter ARI 8 March 2011 4:13

Sleep and Circadian Functioning: Critical Mechanismsin the Mood Disorders?Allison G. Harvey � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 297

Personality Disorders in Later Life: Questions About theMeasurement, Course, and Impact of DisordersThomas F. Oltmanns and Steve Balsis � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 321

Efficacy Studies to Large-Scale Transport: The Development andValidation of Multisystemic Therapy ProgramsScott W. Henggeler � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 351

Gene-Environment Interaction in Psychological Traits and DisordersDanielle M. Dick � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 383

Psychological Treatment of Chronic PainRobert D. Kerns, John Sellinger, and Burel R. Goodin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 411

Understanding and Treating InsomniaRichard R. Bootzin and Dana R. Epstein � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 435

Psychologists and Detainee Interrogations: Key Decisions,Opportunities Lost, and Lessons LearnedKenneth S. Pope � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 459

Disordered Gambling: Etiology, Trajectory,and Clinical ConsiderationsHoward J. Shaffer and Ryan Martin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 483

Resilience to Loss and Potential TraumaGeorge A. Bonanno, Maren Westphal, and Anthony D. Mancini � � � � � � � � � � � � � � � � � � � � � � � 511

Indexes

Cumulative Index of Contributing Authors, Volumes 1–7 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 537

Cumulative Index of Chapter Titles, Volumes 1–7 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 540

Errata

An online log of corrections to Annual Review of Clinical Psychology articles may befound at http://clinpsy.annualreviews.org

Contents vii

Ann

u. R

ev. C

lin. P

sych

ol. 2

011.

7:38

3-40

9. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by S

tate

Uni

vers

ity o

f N

ew Y

ork

- B

rook

lyn

on 0

1/23

/12.

For

per

sona

l use

onl

y.

Documents

Gene-Environment Interaction in Psychological Traits and Disorders