The purpose of the Behavior Genetics Association is to promote scientific study of the interrelationship of
genetic mechanisms and behavior, both human and animal; to encourage and aid the education and training of
research workers in the field of behavior genetics; and to aid in the dissemination and interpretation to the general
public of knowledge concerning the interrelationship of genetics and behavior, and its implications for health and
human development and education.
For additional information about the Behavior Genetics Association, please contact Dr. Hermine Maes BGA
Secretary, Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Box
980003, Richmond VA 23298
Vancouver is situated on the coast of British Columbia, nestled between the Rocky and Coast mountain ranges
and the Pacific Ocean, and is a popular resort and tourist destination. Vancouver has consistently been named one
of the best cities to live in North America and is known for the natural beauty of its environs. The Coast Plaza
Hotel is located next to the warm beaches of English Bay harbor (a two minute walk from the hotel) and Stanley
Park consisting of five square miles of old growth forest in the heart of downtown Vancouver. Mountain hiking or
sailing excursions are no more than an hour away from the hotel. Nearby attractions include Vancouver Island and
Victoria a few hours away by ferry. The climate in July is warm and sunny, average temperature being 26C.
Local Host: Dr. Kerry L. Jang
Department of Psychiatry
University of British Columbia
2255 Wesbrook Mall Tel: (604) 822-7895
Vancouver B.C. Fax: (604) 822-7756
Canada, V6T 2A1 Email: kjang@unixg.ubc.ca
1 Faculty of Letters, Keio University, Mita, Tokyo 180-8345 2 School
of Medicine, Keio University, Shinanomachi, Tokyo 160-8582 3 Supported by Grant-For-Aid of
Scientific Research
Address:2-15-45, Mita, Minato-ku, Tokyo 108-8345, Japan Tel 81+3+3453+4511 /Fax
81+3+3798+7480 /Email juko@media.or.jp
Two comprehensive personality inventories, NEO-PI-R(P.T.Costa, & R.R.McCrae,1985) and TCI
(Temperament and Character Inventory; C.R. Cloninger, D.M. Svrakic., & T.R. Przybeck, 1993; N. Kijima, et
al.,1996) were carried out for 257 pairs of the Japanese twins from 15 to 27 years old (118 MZf, 40 MZf, 48 DZf,
19 DZm, and 32 DZo). All of the Big Five factors measured by NEO-PI-R showed substantial additive genetic
influences (33 - 56 % ) with no nonadditive genetic and shared environmental contribution. (Among seven factors
by TCI, all but one (self-transcendence which showed no genetic contribution) showed additive genetic
influences. Contrary to Cloninger's theory, two character dimensions, self-directedness and
cooperativeness, which are claimed to be environmental, had more genetic variances (42 and 43 %) than
four temperament dimensions ( 17 - 36 % ) which are claimed to be genetic. Multivariate genetic analyses indicated
that the Big Five factors are not genetically independent whereas three temperament dimensions, novelty
seeking, harm avoidance, and reward dependence are genetically independent. It also showed
that the genetic component of persistence is completely derived from novelty seeking and harm
avoidance, and that genetic components of self-directedness and cooperativeness are partially
mediated by those of harm avoidance, and reward dependence. [Poster]
Andrei P.Anokhin1, J.W.Rohrbaugh1,
A.A.Todorov1, and A.B.Vedeniapin1.
The P300 event-related brain potential in neuropsychiatric disorders: a moderator of genetic risk?
1 Washington University School of Medicine, St.Louis, MO 63108
Address: A.P.Anokhin, Dept. of Psychiatry, Washington University School of Medicine, 40
N.Kingshighway, St.Louis, MO 63108
Reduced P300 amplitude has been implicated as a phenotypic marker of genetic risk for substance use disorders
and neuropsychiatric diseases. However, the mechanism through which P300 contributes to risk is unclear, given
the non-specificity of P300 as a marker of risk and a lack of clear evidence for its neural substrates. We propose a
model of predisposition that includes (a) disease-specific liability factors, (b) relatively independent non-specific
factor(s) that modulate the expression of this liability. Psychophysiological and clinical studies, as well as
personality correlates of P300 suggest that low P300 amplitude may indicate a deficit in inhibitory self-regulation of
behavior, a non-specific factor that can facilitate the expression of liability for a variety of disorders. Conversely,
higher self-regulation ability indicated by larger P300 amplitude can operate as a protective factor by suppressing
the manifestation of disease symptoms. We illustrate this model using our analysis of smoking as a prototypical
form of substance abuse (based on data from Collaborative Study on the Genetics of Alcoholism (COGA). P300 is
reduced in current, but not in ex-smokers (P<0.001), and larger P300 is associated with higher probability for
ever-smokers to quit (P<0.01), polysubstance abusers have lowest P300. Finally, we have recently shown that P300
may moderate a relationship between smoking and the dopamine D2 receptor gene polymorphism (DRD2), a
controversial candidate gene for addictive behaviors (A.P. Anokhin, A.A. Todorov, P.A.F. Madden, J.D. Grant, and
A.C. Heath, 1999,Genet. Epidemiol., in print). In individuals with lower P300 amplitudes, there is a
significant association between A1 allele and smoking (p < 0.01). On the other hand, this association is not
observed in individuals with higher P300 amplitudes. In conclusion, we suggest that P300 may be an important
moderator variable which is useful to consider in genetic association studies of addiction and neuropsychiatric
disorders.
Arthur P. Arnold1,2
Non-hormonal mechanisms of brain sexual differentiation
1 Departments of Physiological Science and Neurobiology, Laboratory of
Neuroendocrinology of the Brain Research Institute, UCLA, Los Angeles CA 90095-1527 2Supported
by NIH grants DC00217 and MH59268
Address: Department of Physiological Science UCLA PO Box 951527 Los Angeles CA 90095-1527
USA phone 310-825-2169 fax 310-825-8081 email arnold@ucla.edu
Sexual differentiation of the brain is classically attributed to the action of gonadal steroid hormones. In
mammals, males secrete testosterone during fetal and neonatal life, which acts by itself, or after conversion to
estradiol, to induce masculine patterns of neural development. In the relative absence of androgen, feminine
patterns of neural development occur. These sex differences in brain development are manifested in sex differences
in adult behavior. Despite the wealth of evidence supporting the dominant role of gonadal steroids as inducers of
sexually dimorphic neural development, several recent studies suggest that not all sex differences in development
are attributable to the effects of sex steroids. We have studied sexual differentiation of the zebra finch brain. Male
but not female zebra finches sing a courtship song, and males have a much larger neural circuit for song. Although
estrogen treatment of female finches at hatching causes some masculinization of neural development, it has not been
possible to prevent masculine patterns of neural development in males using various manipulations of the levels of
gonadal hormones. Moreover, it has been possible to create genetic females that possess large amounts of testicular
tissue that secretes androgen. These females retain feminine plumage and a feminine neural song system. Thus, it
appears that testicular secretions are not by themselves sufficient to masculinize neural development. An alternative
to hormonally induced sexual differentiation is "direct genetic" induction of sexual differentiation, in which a
non-hormonal gene product, expressed in brain of one sex, triggers sex-specific patterns of neural development.
The best model for direct genetic induction of sexual differentiation is the induction of masculine differentiation of
the mammalian gonadal ridge by the Y chromosome gene Sry. Future studies will address sex chromosomal gene
effects on neural development. A.P. Arnold (1997) J. Neurobiology 33:572-584.
Allyson J. Bennett1, Klaus-Peter Lesch2, Armin
Heils2, Jeff Long1, Joseph P. Lorenz1, Susan E. Shoaf1,
Maribeth Champoux3, Steven J. Suomi3, and J. Dee Higley1
Serotonin transporter gene variation and early rearing environment interact to affect CSF 5-HIAA
concentrations, aggressive behavior, and alcohol consumption in rhesus monkeys
1National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD 20892
2Department of Psychiatry, University of Würzburg, Würzburg, Germany 97080
3National Institute of Child Health and Human Development, Bethesda, MD 20892
Address: Laboratory of Clinical Studies, NIAAA, NIHAC, P.O. Box 529, Building 112, Poolesville,
MD 20837. Tel (301) 496-9550. Fax (301) 496-0630. Email: allysonb@exchange.nih.gov
A polymorphism in the serotonin (5-HT) transporter gene regulatory region has been associated with measures
of human 5-HT transporter (5-HTT) expression (G.L. Hanna, et al., 1998, Neuropsychopharm., 18
102-111) and with 5-HT-mediated behaviors (K.P. Lesch, et al., Science, 274 1527-1531). We
examined the effect of an analogous length variation of the gene's regulatory region (rh5-HTTLPR) and early
rearing history on central 5-HT functioning, as measured by cerebrospinal fluid 5-hydroxyindoleacetic acid (CSF
5-HIAA), competitive aggressive behavior and alcohol consumption. There was a significant genotype by rearing
interaction on CSF 5-HIAA concentrations (n=132). Parentally deprived, peer-reared monkeys exhibited
differences in CSF 5-HIAA concentrations that were dependent upon genotype. Peer-reared monkeys with the short
rh5-HTTLPR allele exhibited lower CSF 5-HIAA than either their homozygous long allele counterparts or
mother-reared monkeys. Peer-reared homozygotes had higher CSF 5-HIAA than any other group. CSF 5-HIAA
concentrations for mother-reared monkeys were not differentiated by genotype. A gene/environment interaction was
also suggested by a marginally significant interaction effect for alcohol consumption (n=115). Monkeys
reared in absence of adult conspecifics consumed significantly more alcohol than monkeys reared by their mothers.
The effect of peer-rearing on alcohol consumption was exacerbated in monkeys with the short allele, whereas
mother-reared monkeys with the short allele consumed less alcohol than their peer-reared counterparts, although the
effect did not quite reach statistical significance. Mother-reared monkeys were more likely than peer-reared
monkeys to engage in competitive aggression and, in both rearing conditions, monkeys with the short allele
exhibited significantly more aggressive behaviors than their counterparts with the homozygous long allele
(n=100). This study provides evidence of an environment-dependent association between variation in the
5›-regulatory region of the 5-HT transporter gene and CSF 5-HIAA concentrations and suggests a similar effect in
voluntary alcohol consumption, as well as genotype and rearing group differences in competitive aggression.
[Poster]
Michelle L. Bohl1, C. Cykowski1, B. Bowers1, Norm
Henderson2
Using diazepam to reduce anxiety in mouse strains derived from lines selected for low activity in an
Open Field3
1 Institute for Behavioral Genetics, University of Colorado, Boulder, CO 80309
2 Department of Psychology, Oberlin College, Oberlin, OH 44074 3 Supported by
NIMH Grant MH-53480
Address: Institute for Behavioral Genetics, University of Colorado, Boulder, CO 80309 Phone: (303)
735-0077 Fax: (303) 492-8063 e-mail: bohlm@Colorado.edu
DeFries, Gervais, and Thomas, 1978, Behavior Genetics, 8, 3-13, used two replicate sets of F3
mice derived from BALB/cJ X C57BL/6J cross to select for High- or Low open field (OF) activity. Following 30
generations of bi-directional selection the lines were inbred using brother-sister mating. Mice from the two replicate
strains selected for high OF activity have exhibited behaviors associated with low anxiety in a variety of test
situations. Conversely, the low active OF strains exhibit behaviors associated with high anxiety. The goal of the
experiment was to attempt to reduce the difference in anxiety level between the high and low strains by
administering diazepam to the two high anxiety strains. Mice from both replicates were administered either 0.5 or
1.0 mg/kg diazepam or cyclodextrin vehicle. Half of each group were tested sequentially in an open field, elevated
square maze, light/dark box and an elevated plus maze and the remaining half in the reverse test order. Diazepam
reduced the anxiety of both replicate strains in all but the open field. On all tests however, the two low active OF
strains showed considerably more anxiety than the two high active OF strains administered only the vehicle. The
magnitude of the genetic differences in the many measures of anxiety were always substantially greater than the
changes brought about by the diazepam. An attempt was also made to further lower anxiety levels in the two low
anxiety strains. This was unsuccessful at all doses attempted. At the highest dose used (4.0 mg/kg), the low anxiety
strains exhibited sedation effects and a general decrease in activity, even in low-threatening environments.
[Poster]
Dorret I. Boomsma1, Mireille van den Berg1, Conor V.
Dolan2, AL Beem1, P. Eline Slagboom3, Judith R.
Koopmans1, and Eco J.C. de Geus1
Genetics of depression in a selected sample of twins and siblings4
1Vrije Universiteit, Dept of Biological Psychology 2Universiteit van
Amsterdam, Dept of Developmental Psychology 3TNO/PG, Gaubius Laboratory, Leiden
4Supported by NWO grant 904-61-090
Address: De Boelelaan 1111, 1081 HV Amsterdam, The Netherlands Phone 31-20-4448787, Fax
31-20-4448832, Email: dorret@psy.vu.nl
In a longitudinal study of Dutch twins, their parents and siblings we collected questionnaire data on depression,
anxiety and correlated personality traits. Data were collected by mailed surveys in 1991, 1993, 1995 and 1997. Over
13,300 subjects from 3381 families were included in the study. A total of 532 families participated on all 4
occasions. The number of families that participated on three, two and one occasion is 855, 833 and 1161,
respectively. Genetic analyses of anxiety, depression and neuroticism showed individual differences in these traits to
be heritable. Genetic factors accounted for roughly 50% of the variance at each measurement occasion and for most
of the stability in these traits. Based on these data, families were selected for the localization of QTLs involved in
anxiety and depression. A family was selected if at least 2 siblings (or DZ twins) scored extreme on a multivariate
anxiety/depression/neuroticism criterion. Both discordant (high-low) and concordant (high-high and low-low) pairs
were included. Once a family was selected, all family members (parents and offspring) who had at least once
returned a questionnaire booklet were asked to provide a DNA sample. All offspring were asked for a telephone
interview, during which they received part of the computerized version of the WHO-Composite International
Diagnostic Interview (CIDI-Auto). CIDI interview data were obtained from 1251 offspring. These data were used to
obtain DSM-IV (single or recurrent) depression status. There is a clear relationship between extreme scores on the
neuroticism, anxiety and depression questionnaires and DSM-IV diagnoses. Correlations between CIDI scores of
family members for DSM-IV depression are higher in MZ twins than in DZ twins and siblings and suggest genetic
influences.
Nathan Brody1
Secular and non-secular change in Intelligence
1 Department of Psychology, Wesleyan University, Middletown, USA
Address: nbrody@mail.wesleyan.edu
This paper reviews research on spontaneous changes in intelligence over the life-span and as a result of planned
interventions. This research is reviewed as a basis for establishing what changes and what remains constant. These
findings are related to research on secular changes in intelligence. Evidence is presented that indicates that secular
and non-secular changes are partially congruent and that the latter may help to understand the former. In particular,
educational interventions are likely to influence changes in fluid intelligence. These findings indicate that
educational changes have contributed to secular increases in intelligence. Changes in intelligence (whether secular
or non-secular) are considered from a behavioral genetic perspective. The paper also discusses evidence for the
construct validity of tests of intelligence and considers whether or not changes in intelligence change or compromise
the predictive validity of test scores.
S. Alexandra Burt1, Lisa N. Legrand1, Matthew K.
McGue1, and William G. Iacono1
An examination of the heritability common to attitude and achievement2
1 Department of Psychology, University of Minnesota, Minneapolis, MN 55455
2 Supported in part by NIH Grants DA05147, AA09367, and AA00175
Address: Department of Psychology, Elliott Hall University of Minnesota 75 East River Road
Minneapolis, MN 55455-0344 (612) 874-8821 burt0105@tc.umn.edu
Recent studies have found that heritability accounts for a moderate to large amount of the variance in scholastic
achievement, (S. A. Petrill & L. A. Thompson, 1993, Personality and Individual Differences 16,
631-640). However, other studies have determined that attitude towards school has an effect on scholastic
achievement, (V B. Hinsz & R. E. Ployhart, 1998, Journal of Applied Social Psychology 28,
1051-1066). This study seeks to better understand this relationship between attitude towards school and scholastic
achievement. Specifically, to what extent is it a consequence of common genetic or shared environmental factors?
Using 11 and 17 year-old male and female cohorts from the Minnesota Twin Family Study (DZ=926 and
MZ=1706), we assessed IQ, attitude towards school, and school achievement. Preliminary findings partially
replicated earlier studies, (L. Eaves, et al., 1997, Behavior Genetics 27, 121-124.), of genetic
influence on attitude, (rmz=.42, rdz=.35 for 17 year-olds; rmz=.39,
rdz=.16 for 11 year-olds). Correlations for scholastic achievement suggest genetic influence, as well
(rmz=.63, rdz=.30 for 17 year-olds; rmz=.51, rdz=.35 for 11
year-olds). Moreover, for both attitudes and scholastic achievement, twin correlations also suggest shared
environmental influence, albeit moderated by age. In order to explore the relationship between attitude, ability, and
achievement, results from a multivariate genetic analysis will be presented. [Poster]
Andreas Busjahn1, H.Knoblauch2,
H.-D.Faulhaber1, M.Rosenthal1, R.Uhlmann1, H.Schuster1,
F.C.Luft1, and B.Müller-Myhsok3
Peroxisome Proliferator-Activated Receptor PPAR
as a DZ Twinning Gene in Man.
1Franz Volhard Clinic and Max Delbrück Center 2Department
of Medical Genetics, Medical Faculty of the Charite, Humboldt University of Berlin 3Bernhard Nocht
Institute for Tropical Medicine, University of Hamburg, Germany
Address: busjahn@fvk-berlin.de
Weinberg suggested that dizygotic (DZ) twinning is transmitted through the female line. Mormon records
support a recessive mode of inheritance. Data from Finland support a role for natural selection. Increased DZ
twinning in fragile X has implicated CCG repeats. The Booroola fecundity gene is responsible for multiple
ovulation in sheep. However, the genetic mechanisms for DZ twinning in man remain a mystery. About 40% of
spontaneous DZ twin pregnancies result in singleton births, a phenomenon termed "vanishing twins". We tested
the hypothesis that PPAR is a "human DZ twinning" gene and present three lines of evidence. We studied
100 pairs of MZ twins, 64 sets of DZ twins, and 40 parental couples. First, a linkage analysis gave a maximum
multipoint MLB-LOD of 3.09, right on D3S3608. The mean sharing among DZ sibs where DNA from both parents
was available for study on D3S3608 was 0.72. Second, we used a biallelic polymorphism (C->T) in exon 6 of
PPAR and found that the T allele was remarkably lower (P<0.0006) in frequency in MZ twins (0.09) than
in DZ twins (0.23). Third, the biallelic polymorphism was not in Hardy-Weinberg equilibrium in the population of
DZ twins (P<0.0005) with a sizeable over-representation of homozygotes. This over-representation was not present
in the MZ twins or in the parents of DZ twins, where Hardy-Weinberg proportions were maintained.
PPAR is a transcription factor involved in adipocyte differentiation, insulin-related effects, lipid
metabolism, body mass index, and appears pivotal to the growth process. We suggest that intra-uterine selection
based on genetic variation in PPAR may be responsible for the deviation from Hardy-Weinberg
equilibrium.
Andreas Busjahn1, H.Knoblauch2,
H.-D.Faulhaber1, M.Rosenthal1, R.Uhlmann1, H.Schuster1,
F.C.Luft1, and B.Müller-Myhsok3
Peroxisome proliferator activated receptor PPAR gene locus is related to body mass index
and lipid values in normal subjects
1Franz Volhard Clinic and Max Delbrück Center 2Department
of Medical Genetics, Medical Faculty of the Charite, Humboldt University of Berlin 3Bernhard Nocht
Institute for Tropical Medicine, University of Hamburg, Germany
Address: busjahn@fvk-berlin.de
The peroxisome proliferator activated receptor (PPAR) is a transcription factor involved in adipocyte
differentiation, insulin-related effects, lipid metabolism, body mass index, and appears pivotal to the growth
process. The PPAR gene has been implicated in morbid obesity and is important to lipid and carbohydrate
metabolism. However, the relevance of gene variations in normal subjects is not defined. We recruited
monozygotic (MZ) and dizygotic (DZ) normal twin subjects to test the hypothesis that the PPAR gene is
important to body mass index and lipid concentrations in normal subjects. Both linkage and association strategies
were employed in the same DZ twin subjects. The PPAR gene locus was linked (p<0.01) to high density
lipoprotein cholesterol (HDL), low density lipoprotein cholesterol (LDL) and body mass index (BMI) as
quantitative traits. A biallelic variant in the PPAR gene was associated with the HDL and BMI (p<0.05).
We also looked for linkage between the same variables and the retinoic X receptor gene locus. This locus was
linked to total and LDL cholesterol as well as triglycerides. We conclude that the PPAR gene is highly
relevant to lipid metabolism and body mass index, not only in the morbidly obese, but in normal subjects as well.
The same appears to be true for its binding partner. Sequencing these genes in the twin subjects would serve to
identify gene variations contributing to BMI and lipid concentrations in normal subjects.
Lon R. Cardon1, G.R. Abecasis1, and W.O.C.
Cookson1
Testing linkage and linkage disequilibrium with quantitative trait loci in nuclear families: A DF
regression model and variance components extensions
Address: Wellcome Trust Centre for Human Genetics, University of Oxford, Windmill Road, Oxford
OX3 7BN United Kingdom Tel: +44 01865 740 002 Fax: +44 01865 742 193 Email: lon@well.ox.ac.uk
Fulker et al. (D. W. Fulker, S. S. Cherny, P. C. Sham and J. K. Hewitt, 1999, Am J Hum Genet.
64, 259-267) recently described a combined test of association and linkage in sib-pair families using the
variance components framework. This model has considerable advantages over existing methods for fine-mapping
quantitative trait loci (QTLs), including a direct test of population substructure vs. linkage disequilibrium and the
absence of model dependence on parental genotypes. We describe extensions of this approach to the DF regression
model and to sibships of any size using variance components. Analytical expectations for the regression coefficients
are described, allowing direct interpretation of the parameter estimates. Parental data are not required in the
extensions, although such data do give additional power to test QTL-marker allele association and to determine
whether any such association is attributable to linkage disequilibrium or population admixture. The relationship
between power and family structure, explored using simulation studies, indicates that when parental genotypes are
available, power is largely independent of the number of offspring in each family. Power is reduced in the absence
of parental genotypes, but the loss in power is negligible when four or more offspring per family are genotyped.
When multiple siblings are available, the total number of genotypes required to achieve comparable power is
smaller when parents are not genotyped. These results are discussed in the context of sampling designs and
genotyping strategies for fine-mapping quantitative trait loci.
Stacey S. Cherny1,2, Pak C. Sham2, Shaun Purcell2,
and John K. Hewitt1
Selecting maximally informative sibships for QTL linkage analysis3
1Institute for Behavior Genetics, University of Colorado, Boulder, CO
80309-0447 2Social, Genetic and Development Psychiatry Research Centre, Institute of Psychiatry,
DeCrespigny Park, Camberwell, London SE5 8AF, United Kingdom 3Supported in part by EY-12562
and a Programme Project grant from the Medical Research Council of Great Britain
Address: Institute for Behavioral Genetics Campus Box 447 University of Colorado Boulder, CO
80309-0447 Phone: +1 303 492 0835 FAX: +1 303 492 8063 Email: Stacey.Cherny@Colorado.EDU WWW:
http://ibgwww.colorado.edu/~cherny/
Much work has been done in the area of increasing power of linkage studies by use of selective sampling. It is
generally known that maximally discordant sib pairs are most informative, although affected pairs can be more
informative still in the presence of a rare recessive. However little work has been done in the area of construction of
an optimal sample, once the sample size for genotyping has been determined. For example, clearly even under a
simple additive model, the most extremely affected pairs will be more informative than all most but the most
extremely discordant pairs. When conducting a large scale study where all sibships have been phenotyped and the
next task is to select the potentially most informative sibships or members of a sibship for genotyping, a method for
rank ordering all the sibships in a study by their potential informativeness for linkage would be most desirable. This
will describe such a method and its implementation in a freely-distributed Fortran program.
Frederick L. Coolidge1, Linda L. Thede1, and Kerry L.
Jang1
Genetic Contributions to Personality Disorders in Childhood
1 Psychology Department, University of Colorado at Colorado Springs, CO
80933-7150
Address: Psychology Department, P. O. Box 7150, University of Colorado, Colorado Springs, CO
80933-7150. Telephone: 719/262-4146; fax: 719/262-4166; e-mail: fcoolidg@mail.uccs.edu
Evidence for a genetic basis of normal and abnormal personality traits has gathered increasing attention in
recent years. Much of this research has been confined to adults, particularly with respect to abnormal personality
traits. The present study examined heritability of personality disorders in childhood and early adolescence.
Personality disorders are chronic and pervasive maladaptive behaviors that cause significant disruption in social and
occupational functioning. There is evidence that many personality disorders begin in adolescence or even earlier,
but the lack of evidence of a genetic basis of personality disorders in childhood, in part, may be due to the lack of
assessment measures. The present study assessed the heritability of 12 personality disorders in 82 monozygotic
(MZ) twins and 68 dizygotic (DZ) twins between the ages of 5 and 14 years old. The Coolidge Personality and
Neuropsychological Inventory (CPNI; F. L. Coolidge, 1998, CPNI ManualAuthor, Colorado Springs), a
DSM-IV criteria based, standardized, 200-item, parent-as-respondent questionnaire was given to the parents of the
150 twins. The CPNI has been normed on 329 children ages 5-17, the median scale reliability of the 12 personality
disorder scales is .67, and validity studies support its use in a variety of clinical settings. Heritability estimates were
determined by using Holzinger's H statistic (L. J. Eaves, H. J. Eysenck, & N. G. Martin, 1989, Genes, Culture
and PersonalityAcademic Press, San Diego). It was found that the median heritability for the 12 personality
disorder scales was 51%, ranging from 27% for the passive aggressive personality disorder scale to 68% for the
conduct disorder scale. Six of the 12 personality disorder scales had significantly greater MZ correlations than DZ
correlations (borderline, conduct, dependent, histrionic, paranoid, and schizotypal personality disorders). It was
concluded that personality disorders appear to have a strong genetic component measurable in childhood, and the
sizes of the heritability coefficients appear to be similar to those found in most adult studies.
Robin P. Corley1, and S. A. Rhea1
Substance experimentation in adopted adolescents in the Colorado Adoption Project: Uncommon
environmental effects?2
1 Institute for Behavioral Genetics, University of Colorado, Boulder, CO
80309-0447. 2 Supported by NICHD grant HD-10333, NIMH grant MH-43899, and NIDA grants
DA-05131 and DA-11015
Address: IBG, Campus Box 447,University of Colorado, Boulder, Boulder, CO 80309-0447. Phone:
303-492-5189, FAX #: 303-492-8063, E-mail: corley@colorado.edu
As part of an ongoing study of adolescent substance experimentation in subjects participating in the Colorado
Adoption Project, we examined self-reported use of nicotine, alcohol, marijuana, and other drugs for 199 adopted
adolescents, and 212 non-adopted adolescents, obtained when the subjects were age 16 years. Adopted subjects
were more likely, though not always significantly more likely, to have experimented with each individual substance,
as well as with multiple substances, with odds ratios ranging from 1.20 for ever tried marijuana to 1.87 for ever tried
other drugs. Both genetic and sociocultural factors may help explain these differences. We examine the predictive
power for experimentation within the adoptee sample of several different measures including the personality
domains of harm-avoidance and novelty-seeking, major life-events such as divorce, a known positive history for
substance use in biological parents, and measures specific to adoptive families adapted from the Search Institute's
national study of adoptive families (A.R. Sharma, M.K. McGue, and P.L. Benson, 1996, Children and Youth
Services Review, 18, 83-100). [Poster]
Nikole J. Cronk1, Wendy S. Slutske1, Pamela A. F.
Madden2, Kathleen K. Bucholz2, and Andrew C. Heath2
The equal environments assumption and similarity of mother reports of emotional and behavioral
problems among female adolescent twins3
1Department of Psychology, University of Missouri-Columbia, Columbia, MO
65211 2Department of Psychiatry, Washington University School of Medicine, St. Louis, MO 63108
3 Supported by NIH grants AA09022, AA00264, AA07728, and DA00272
Address: Department of Psychology, University of Missouri, 210 McAlester Hall, Columbia, MO
65211 phone: (573) 882-4043 fax: (573) 882-7710 email: nikole@taxa.psyc.missouri.edu
Critics of the twin study method often assert that the greater similarity of monozygotic (MZ) compared to
dizygotic (DZ) twins is caused by the relatively greater degree of experiences shared by MZ versus DZ twins, rather
than by the greater degree of genetic resemblance of MZ versus DZ twins. We assessed the influence of
environmental similarity on MZ and DZ twin resemblance for mother-reported emotional and behavioral problems
in a sample of 1,948 female adolescent twin pairs. Twins were identified from state of Missouri birth records and
the measures of environmental similarity (perceptions of twin zygosity, having the same friends, being in the same
classes, and dressing alike) and emotional and behavioral problems (symptoms of separation anxiety disorder,
attention-deficit hyperactivity disorder, oppositional-defiant disorder, and conduct disorder) were obtained via
structured telephone interviews with the twins' biological mothers. There were significant differences between MZ
and DZ twin correlations for all of the symptom scales prior to accounting for environmental similarity. After
controlling for environmental similarity, there were still significant differences between MZ and DZ twin
correlations for nearly every combination of symptom scale and environmental similarity measure. In most cases,
within-zygosity comparisons of twins differing in greater and lesser degrees of environmental similarity yielded
non-significant differences, although there were a few consistent effects of environmental similarity on twin
resemblance. Overall, this study supports the conclusion that genetic similarity is the predominant factor leading to
greater MZ than DZ similarity in twin studies of mother-reported emotional and behavioral problems.
Chayna J. Davis1, Valerie S. Knopik1, Sally J.
Wadsworth1, and John C. DeFries1
Etiology of the relationship between reading performance and rapid automatic naming: A twin
study2
1Institute for Behavioral Genetics, University of Colorado, Boulder, CO 80309
2Supported by grant HD-27802 from NICHD, and grant MH-16880 from NIMH
Address: Chayna J. Davis Institute for Behavioral Genetics Campus Box 447 Boulder, Colorado 80309
Phone: 303-492-2817 Fax: 303-492-8063 E-mail: davisc@colorado.edu
Scarborough (1998, Annals of Dyslexia 48, 115-136) recently found that early literacy scores
were the best predictor of later reading performance in a normally achieving sample, but prediction was
significantly improved by including rapid naming tasks in a reading-disabled sample. In the present study, the
relationship between reading performance and rapid automatic naming (RAN) was examined by fitting structural
equation models to data from 679 twin pairs (291 monozygotic and 388 dizygotic) in which at least one member of
the pair was reading-disabled (RD) and from 439 control twin pairs (218 monozygotic and 221 dizygotic) tested in
the Colorado Learning Disabilities Research Center. The measures included the reading recognition, reading
comprehension and spelling subtests (READ) of the Peabody Individual Achievement Test, as well as 4 subtests
(numbers, colors, pictures, and letters) of the RAN paradigm. Results from a bivariate phenotypic model with two
hypothesized latent factors, READ and RAN, indicated that the correlation between reading and RAN performance
for the reading-disabled sample (.58) was significantly different from that of the control sample (.29). When this
model was partitioned to include estimates of genetic, shared environmental and non-shared environmental
influences, resulting heritability estimates did not differ significantly for the RD and control samples for either
READ (h2 = .86 and .71, respectively) or RAN (h2 = .72 and .64, respectively).
However, the genetic correlation between the READ and RAN latent factors could not be equated for the two
groups (rg = .66 for probands and .32 for controls). Thus, the etiology of the relationship between
reading performance and RAN may differ for reading-disabled and normally-achieving readers. Moreover, these
results support previous findings that the best predictors of reading skills may differ for samples of children with
normal reading levels and those with reading difficulties (Scarborough, 1998).[Poster]
Filip De Fruyt1, and Ivan Mervielde1
A behavioral genetic analysis of children's traits
1 Department of Psychology, University of Ghent, Belgium
Address: H. Dunantlaan 2, B-9000, Gent, Belgium
A behavioral genetic analysis of children's traits. The construction and cross-validation of the Hierarchical
Personality Inventory for Children (HiPIC) will be described and data will be reported from a Flemish twin study in
which the HiPIC was used to describe individual differences among children aged 6 to 12. The HiPIC is a 144-item
inventory assessing 5 broad personality domains and 18 more specific personality facets. The facets are empirically
constructed and closely cover individual differences denoted in parental free descriptions of childhood personality.
284 parents of mono- (MZ) and dizygotic (DZ) twins aged 6 to 12 provided independent ratings of their twins using
the HiPIC. The fit of different models to represent the trait variance, including models to examine contrast effects,
were investigated, showing substantial additive genetic effects and a relatively small contribution of the shared
environment. The results are in line with previous behavioural genetic studies with children using temperament
inventories and extend the findings obtained for adults using hierarchical conceptualisations of the Five-Factor
Model to children. [Poster]
V.H Denenberg, and A.J. Stavnezer
Spatial Ability of XY Sex-reversed Female Mice
Perinatal gonadal hormones significantly affect sex differences in rodents for both reproductive and
non-reproductive behaviors. However, the influence of the sex chromosomes has been largely ignored. To assess
the influence of the non-pairing region of the Y chromosome, C57BL/6J male and female mice and mice from
theC57BL/6JEi-YPOS consomic strain were given behavioral tests known to distinguish males from females. The
C57BL/JEi-YPOS strain contains sex-reversed XY -females which, when compared to their XX-female siblings,
allow assessment of the influence of the Y chromosome in a female phenotype. XX females and XY-females did
not differ on open-field activity, the Lasley maze, or active avoidance learning, but the XY -females were
significantly better than the XX-females on the Morris hidden platform spatial maze. These findings suggest that
normal males may have two functional mechanisms (hormonal and genetic) to ensure visuospatial superiority.
Danielle M. Dick1, Richard J. Rose1, Richard J. Viken1,
and Jaakko Kaprio2
Longitudinal analyses of genetic and environmental influences on drinking across late
adolescence3
1Indiana University, Department of Psychology, Bloomington, IN 47405
2University of Helsinki, Department of Public Health, Helsinki, Finland, 00014
3Supported by NIAAA grants AA-09203 and AA-07611, and the Academy of Finland
Address: ddick@indiana.edu, 812-855-4101 (ph), 812-855-4691 (fax)
Late adolescence is a critical time period for the development of patterns of alcohol use and abuse, underscoring
the importance of research among adolescent populations. Using three waves of data collected at ages 16, 17, and
18.5 from FinnTwin16, we analyzed longitudinal change in the influence of genetic and environmental
factors on self-reported drinking frequency. Complete drinking data were provided at all three time points by 1,570
pairs of same-sex Finnish twins, ascertained through Finland's Central Population Register: 296 male MZ pairs, 389
male DZ pairs, 480 female MZ pairs, and 405 female DZ pairs. Trivariate analyses were conducted using the
statistical package Mx (Neale, M., 1997, Mx: Statistical modeling, Box 126 MCV, Richmond, VA 23298:
Department of Psychiatry, 4th Edition). At the baseline assessment at age 16, genetic factors accounted for
approximately 20% of the variance in drinking frequency, and common environmental factors accounted for 60%,
with the remaining variance attributed to unique environmental factors and error. The influence of genetic factors
increased substantially from age 16 to age 17 and remained relatively stable thereafter. No new genetic effects were
evident at age 17; however, new genetic effects were apparent at age 18. The influence of common environment
decreased progressively over the study period, while the influence of unique environmental effects increased. The
model in which parameter estimates for males and females were constrained to be equal provided a parsimonious
and adequate fit to the data. Differences in longitudinal influences on drinking among twins differing in
socioregional background and family history will be explored in further analyses.
Michael G. DuPree1,2, Leo A. Sirota1, Dean H.
Hamer1
A Genome Screen for Sexual Orientation: Progress Report3
1Laboratory of Biochemistry, National Cancer Institute, The National Institutes of
Health, Bethesda, MD 20892 2Department of Anthropology, The Pennsylvania State University,
University Park, PA 16802
Address: Laboratory of Biochemistry, Section on Gene Structure and Regulation, National Cancer
Institute, National Institutes of Health, Bldg 37: 4A-13, Bethesda, MD 20892, (301)496-1747,
dupreem@pop.nci.nih.gov
Despite twin evidence that male and female homosexuality have a heritable component(J.M. Bailey, D.S.
Benishay, 1993, Am.J.Psychiatry 150(2), 272-277.; J.M. Bailey, A.P. Bell, 1993, Behavior
Genetics 23, 313-322.; F.L. Whitman, M. Diamond, J. Martin, 1993, Archives of Sexual
Behavior 22,187-206.), nothing is known about the underlying molecular genetic factors in the
development of sexual orientation. The finding of X chromosome linkage for male homosexuality (D.H. Hamer, S.
Hu, V.L Magnuson, N. Hu, A.M. Pattatucci, 1993, Science 261(5119), 321-327.; S. Hu, A.M.
Pattatucci, C. Patterson, L. Li, D.W. Fulker, S.S. Cherny, L. Kruglyak, D.H. Hamer, 1995, Nature Genetics
11(3), 248-256.) does not fully explain the occurrence of the behavior and there have been few
investigations into the possible contributions of loci on other chromosomes (J.P. Macke, N. Hu, S. Hu, M. Bailey,
V.L. King, T. Brown, D. Hamer, J. Nathans, 1993, Am. J.Hum.Genet 53, 844-852.). Using a
battery of 400 markers, we are conducting a genome screen of a sample of nearly 500 individuals consisting of 135
sib-pairs and 10 sib-trios concordant for homosexuality and any heterosexual family members available for
genotyping. We report on our progress midway through the research, discussing findings to date as well as our
statistical methods. [Poster]
Lindon J. Eaves1, and Patrick F. Sullivan1
Genotype x Environment Interaction in Transmission Disequilibrium Tests2
1 Virginia Institute for Psychiatric & Behavioral Genetics, Department of Human
Genetics and Psychiatry, Virginia Commonwealth University2 Supported by NIH grants MH45268
Address: VIPBG, MCV/VCU, P.O. Box 980003, Richmond VA 23298-0003. 804/828-8155 (phone);
804/828-8801(fax); eaves@hsc.vcu.edu; http://www.vipbg.vcu.edu/vipbg/
Transmission disequilibrium tests (TDTs) provide an approach to the detection of associations between alleles
at marker loci and risk to complex disorders. The logistic regression approach to TDTs proposed by Sham and
Curtis (1995) is generalized to provide separate tests of the main effects of marker loci on genetic risk and genotype
x environment interaction (GxE) arising because alleles differ in their sensitivity to specified environmental
covariates. The same model may be used to detect the effects of genomic imprinting on the expression of
susceptibility loci. In the presence of GxE, highly significant genetic effects may be present that will not produce
marked twin or sibling resemblance and not yield significant associations in conventional TDTs. However,
simulation studies show how the logistic regression model can be used to detect the main effects of marker alleles
and their interaction with covariates on continuous and dichotomous outcomes in offspring-parent trios, pairs of
siblings and their parents, and monozygotic twins pairs and their parents. TDT tests with MZ twin pairs permit the
detection of alleles whose primary effects on the phenotype are mediated through the control of sensitivity to latent
features of the within-family environment. It is shown the genotype-environment correlations, caused by the
environmental effects of parental alleles on offspring phenotypes can produce spurious marker-phenotype
association in population studies that do not bias the outcome of TDTs
Thalia Eley1, and the Twins' Early Development Study (TEDS)
The Aetiology of Anxiety Symptoms in Pre-school Children: Temperament or Psychopathology?
1 Social, Genetic, & Developmental Psychiatry Research Centre, Institute of
Psychiatry, London
Address: Social, Genetic, & Developmental Psychiatry Research Centre, Institute of Psychiatry, 111
Denmark Hill, London, UK SE5 8AF Phone: +44 171 919 3890 Fax: +44 171 919 3866 Email:
T.Eley@iop.bpmf.ac.uk
This study examines the aetiology of anxious temperament and psychopathology in both the normal and
abnormal range in order to assess whether these two aspects of anxiety are better considered together or apart. The
sample included over 3000 pairs of twins, recruited from the entire birth cohort of twin pairs born in England and
Wales in 1994. Anxiety symptoms were assessed using the pre-school version of the Behar scale, which includes
three items assessing emotional temperament, and 3 items assessing anxiety and depression psychopathology.
Highly different patterns of response were revealed for these two types of items, with roughly normally distributed
answers on the temperament items as compared to highly skewed responses for the psychopathology items. Two
sub-scales were created: emotional temperament and anxiety psychopathology. The emotional temperament scale
showed a pattern of very low DZ correlations relative to MZ, indicating substantial non-additive genetic variance.
In fact, the model of best fit included significant non-additive genetic (d2 = .52) and non-shared environmental (e2
= .48) influences. Shared environment was non-significant. For extreme group membership (highest 8%) the same
pattern resulted, d2 = .35 and e2 = .65. In contrast, DZ twin correlations on the anxiety psychopathology scale were
at least half as great as MZ correlations indicating both additive genetic and shared environment contributions. This
was also confirmed using model-fitting in which additive genetic (a2 = .52), shared environment (c2 = .10), and
non-shared environment (e2 = .39) parameters were all significant contributors to the variance. For the extreme
group (top 8%) the same pattern was found, a2 = .40, c2 = .13, and e2 = .47. In summary, emotional temperament
and anxious psychopathology have very different aetiologies, yet both scales appear influenced by similar factors
both in the normal and abnormal range. Bivariate genetic analyses will also be presented.
Susan Felsenfeld1, G. Zhu2, D. Statham2, and N.
Martin2
Examining the heritability of stuttering in Australian twins: Descriptive and epidemiological
analyses
1 Department of Speech-Language Pathology, Duquesne University, Pittsburgh,
PA 15282 2 Queensland Institute of Medical Research, Brisbane, QLD 4029, Australia 3
Supported by NIH Grant RO1 DC03776-01
Address: Susan Felsenfeld, Ph.D. Department of Speech-Language Pathology 600 Forbes Avenue
Duquesne University Pittsburgh, PA 15282-2231 Phone: 412-396-4205 Fax: 412-396-4196 Email:
felsenfeld@duq.edu
The occurrence of stuttering has been found to be at least moderately heritable in the small number of twin and
family studies of this disorder completed to date (N. Ambrose, E. Yairi, & N. Cox, 1993, Journal of Speech and
Hearing Research 36, 701-706; G. Andrews, A. Morris-Yates, P. Howie, & N. Martin, 1991,
Archives of General Psychiatry 48, 1034-1035; P. Howie, 1981, Journal of Speech and
Hearing Disorders 24, 317-321; K. Kidd, 1984, in R.F. Curlee and W.H. Perkins, eds., Nature and
Treatment of Stuttering: New Directions, Boston: Allyn and Bacon). The present study adds significantly to
these findings by assessing directly a large sample of stuttering twins drawn from the Australian Twin Registry.
From a database of 6,717 subjects in two age cohorts who responded to a questionnaire item about stuttering, 407
subjects who indicated a positive stuttering history were identified. These positive cases, their cotwin, and a small
age-matched control sample were subsequently interviewed by telephone to confirm the diagnosis and to provide
descriptive information (N=589 completed interviews). In the present study, selected descriptive findings (i.e.,
symptom frequencies, age of onset, recovery status, relapse history, frequency of comorbid speech problems) for the
confirmed affected cases will be reported. In addition, probandwise concordance values for confirmed cases will be
presented, and discordant twin pairs will be examined to identify for future study possible nongenetic risk factors
for this disorder.
Deborah Finkel1, and Nancy L. Pedersen2
Contribution of age, genes, and environment to the relationship between perceptual speed and
cognitive ability3
1 Division of Social Sciences, Indiana University Southeast, New Albany, IN
47150. 2 Genetic Epidemiology, Karolinska Institute, Stockholm S-171 77 and Department of
Psychology, University of Southern California, Los Angeles, CA. 3 Supported by
American-Scandinavian Foundation fellowship and NIA grant AG15211 awarded to Dr. Finkel. SATSA is
supported by NIA (AG04563, AG10175), The MacArthur Foundation Research Network on Successful Aging, and
the Swedish Council for Social Research (97:0147:1B)
Address: Division of Social Sciences, Indiana University Southeast, 4201 Grant Line Road, New
Albany, IN 47150, phone: 812-941-2668, fax: 812-941-2591, e-mail:dfinkel@ius.edu
The aim of the present analysis was to examine genetic influences on cognitive ability in adulthood in the
context of the relationship between perceptual speed and cognitive aging. Quantitative genetic analysis of data from
the Swedish Adoption/Twin Study of Aging allowed for estimation of the contribution of age, genetic, and
environmental effects to the variance in a latent cognitive factor and to the covariance between the cognitive factor
and perceptual speed. The sample included 302 pairs of monozygotic and dizygotic twins, both reared together and
reared apart, ranging in age from 40 to 84 years. Analysis of components of total variance in the cognitive factor
indicated that 90% of the age-related variance in the cognitive factor was shared with perceptual speed, and 70% of
the genetic variance in the cognitive factor was shared with perceptual speed. The correlation between the speed and
cognitive factors was primarily genetically mediated. [Poster]
Deborah Finkel1, Nancy L. Pedersen2, and Maria
Larsson3
Odor perception and its relationship with cognitive functioning: A twin study4
1 Division of Social Sciences, Indiana University Southeast, New Albany, IN
47150. 2 Division of Genetic Epidemiology, Karolinska Institute, Stockholm S-171 77 and Department
of Psychology, University of Southern California, Los Angeles, CA. 3 Department of Clinical
Neuroscience and Family Medicine, Division of Geriatric Medicine, Karolinska Institute, and Department of
Psychology, Uppsala University, Sweden 4 Supported in part by NIA grant AG15211 awarded to Dr.
Finkel. SATSA is supported by NIA (AG04563, AG10175), The MacArthur Foundation Research Network on
Successful Aging, and the Swedish Council for Social Research (97:0147:1B)
Address: Division of Social Sciences, Indiana University Southeast, 4201 Grant Line Road, New
Albany, IN 47150, phone: 812-941-2668, fax: 812-941-2591, e-mail:dfinkel@ius.edu
Results from the National Geographic study of odor perception demonstrated marked individual differences and
age differences in the sense of smell (R.J. Russell, B.J. Cummings, B.F. Proffitt, C.J. Wysocki, A.N. Gilbert, and
C.W. Cottman, 1993, J. Gero., 48, P49-P53.). Twin studies have reported significant heritabilities
for odor thresholds for some, but not all, odors (N.L. Segal, T.D. Topolski, S.M. Wilson, K.W. Brown, and L.
Araki, 1995, Phys. and Beh., 57, 605-609.). The purpose of the present analysis was to incorporate
data from both reared apart and reared together twins in a behavior genetic investigation of odor perception. A
sample of 37 MZA pairs, 62 MZT pairs, 87 DZA pairs, and 77 DZT pairs from SATSA, age 40 to 84 years,
completed the National Geographic smell survey. Measures included odor detection, odor identification, and
qualitative measures of odor strength for 6 different odorants. Responses were corrected for the effects of sex, age,
and smoking (packyears). Univariate analysis indicated heritabilities ranging from .40 for the detection of
mercaptans (natural gas) to .60 for the detection of galaxolide (musk). Heritability of perceived odor strength was
.36, with the remaining variance attributable to nonshared environmental influences. Signicant correlations between
percent correct odor identification and cognitive measures such as Analogies and Synonyms were predominantly
genetically mediated. The results suggest a possible identification mechanism that may be common both to odor
identification and more verbal measures of identification.
James R. Flynn1
IQ gains and fluid g: a research design to discover causes
1 Political Studies, University of Otago, Dunedin, New Zealand
Address: email jim.flynn@stonebow.otago.ac.nz
It has long been known that IQ gains are greatest on tests of fluid g. An analysis of differential gains on WISC
subtests adds confirmation. The magnitude of gains on various subtests is correlated with the the fluid-g loading of
the subtests. Ravens is used as an index of a subtest's fluid-g loading. In so far as fluid g plays a role in the theory
of intelligence, the fact that IQ gains are unaccompanied by the real-world effects normally associated with g poses
theoretical problems. These are exacerbated, not solved, by the relatively low gains in crystallized g. The pattern of
gains suggests a research model for determining causes. Gains are highest on Ravens, Block Design, and Verbal
Similarities; they are lowest on Arithmetic, General Information, and Vocabulary. Therefore, if contemporary
samples were divided into two groups who are matched on the low-gain tests, but diverge by about one SD on the
high-gain tests, we might simulate two generations separated by 'time'. The problems and promise of this design are
discussed. The author speculates that the proximate cause of gains is a new motivational disposition to take certain
kinds of problems seriously. This hypothesis could be tested by an attitude survey. The ultimate causes suggested
in the literature, ranging from breast feeding and number of siblings to more progressive schooling and parenting,
could be tested by perceived correlations, or lack of such, with the score patterns that separate the two groups.
Nadine Forget-Dubois1, Marie-Claude Martel2, Daniel
Pérusse1, George Tarabulsy2, Michel Boivin3, and Richard E.
Tremblay4
A study of maternal sensitivity in 5-month-old twins5
1Department of Anthropology, University of Montreal/Centre de recherche
Fernand-Seguin, L.-H. Lafontaine Hospital, Montreal, Quebec, Canada 2Department of Psychology,
University of Quebec at Trois-Rivieres, Trois-Rivieres, Quebec, Canada 3School of Psychology, Laval
University, Quebec, Quebec, Canada 4 Department of Psychology, University of Montreal, Montreal,
Quebec, Canada 5Supported by grants to D. Pérusse from the Medical Research Council of
Canada, the National Health Research and Development Program/Health Canada, The Social Sciences and
Humanities Research Council of Canada, the Fond de la recherche en sante du Quebec and the Quebec Ministry of
Health and Social Services. Nadine Forget-Dubois is supported by a fellowship from the Fonds pour la formation de
chercheurs et l'aide a la recherche/Fond de la recherche en santé du Quebec
Address: Centre de recherche Fernand-Seguin, 7331 Hochelaga, Montreal, Canada, H1N 3V2. E-mail:
forgetdubois@sympatico.ca. Phone: (514) 251-4015. Fax: (514) 251-2617
Maternal sensitivity is associated with various aspects of socioaffective and cognitive development (e.g.
Ainsworth et al., 1978, Patterns of attachment: A psychological study of the strange situation.
Hillsdale, NJ: Erlbaum, Lewis, M.D., 1993, Developmental Psychology 29, 1036-1045, etc.). Bell
and Harper (1977, Child effects on adult, Hillsdale, NJ: Erlbaum) proposed that individual characteristics of
children could influence parental behavior. The present research examines the variation in maternal sensitivity
according to zygosity in a sample of 100 same-sex, five month-old twins and their mothers. Maternal sensitivity was
assessed independently for each dyad (mother-twin 1, mother-twin 2) using the Maternal Behavior Q-Sort
(Pederson et al., 1990, Child Development 61, 1974-1983). Assessment was performed
from videotapes of free interactions occurring during a laboratory visit, between experimental situations. Each visit
lasted between 4 and 5 hours; observation time for each twin ranges between 0,60 and 1,92 hours. We found higher
intra-class correlations of maternal sensitivity scores in MZ than in DZ twins for both sexes. Heritability estimates
for maternal sensitivity reached 50%. These results show that (1) maternal sensitivity varies according to infant's
characteristics and (2) that variation in these latent characteristics are under genetic influence. Considering that
maternal sensitivity may act as a variable of the infant's environment, these results suggest the potential for
genotype-environment correlations in child development. [Poster]
Qiang Fu1, Andrew C. Heath 1, Kathleen K. Bucholz
1, Seth A. Eisen1, Jack Goldberg2, Michael J. Lyons3,
and William R. True 4
Genetic and Environmental Effects on Suicidality: Findings from USA Vietnam Era Twin (VET)
Registry5
1Washington University School of Medicine, St. Louis, MO 63108
2University of Illinois, Chicago School of Public Health and the Cooperative Studies Program
Coordinating Center, VAMC, Hines, IL 3Harvard Medical School, Department of Psychiatry at the
Brockton/West Roxbury VAMC and Department of Psychology, Boston University 4St. Louis
University School of Public Health, St. Louis, MO 6308 5Supported by AA11822 AA10339,
DA04604, LIP No. 41-065, MH37685, MH31302
Address: Washington University, 40 N. Kingshighway Blvd., Suite 2, St. Louis, MO 63108
Suicidal behavior has been found to occur more often among people experiencing psychiatric disorders and
traumatic life events and also aggregates in families. However, general population studies examining genetic and
environmental contributions to suicid al behavior are rare. Recent data from 5,995 twins from the community
yielded heritability extimates for suicidal behavior of 45% (33-51%) (D.J. Statham, A.C. Heath, P.A.F. Madden,
K.K. Bucholz, L. Bierut, S.H. Dinwiddie, W.S. Slutske, M.P. Dunne & N.G. Martin, 19 98, Psychol Med
28, 839-855). The present study explores genetic and environmental contributions to suicidality based on a
national representative sample of same sex U.S. male twins from the VET registry. We studied 3,372 complete twin
pairs who were interviewed by telephone in 1992 with a computer version of the DIS-III-R. Lifetime prevalence of
suicidal thoughts and attempts was 16.1% and 2.4%, respectively. Adjusting for demographic variables and combat
exposure, respondents with suicidal thoughts were more likely to qualify for lifetime DSM-III-R diagnosis on
alcohol (OR=1.36, 95%CI=1.13, 1.64), drugs (OR=1.79, 95%CI=1.15, 2.76), and both substances (OR=2.07,
95%CI=1.55, 2.75), Antisocial Personality Disorder (OR=1.74, 95%CI=1.18, 2.55), Major Depressive Episode
(OR=6.61, 95%CI=5.32, 8.20), Panic Disorder (OR=1.96, 95%CI=1.12, 3.34), and to report a family history of
depression (OR=1.75, 95%CI=1.46, 2.10). Those with suicide attempts were more likely to meet criteria for Major
Depressive Episode (OR=4.54, 95%CI=2.95, 7.00), nicotine dependence (OR=1.99, 95%CI=1.22, 3.26), to have a
cotwin with depression (OR=1.83, 95%CI=1.18, 2.84), and to report family history of alcohol abuse (OR=1.84,
95%CI=1.20, 2.83). Multiple logistic regression ana lyses revealed genetic influence on suicide thoughts and a
familial effect on suicide attempts. A series of models were assessed using model-fitting procedure with MX.
Although none of these models fit the data well, the most improved model yielded heritability estimates for suicidal
thoughts and attempts of 47% and 40%, respectively (X2=6.19, d.f.=1, p<.02). Our results suggest
that both genetic and environmental effects influence on suicidality.
Javier Gayán1 and R. K. Olson1
Behavioral Genetic Analysis of Individual Differences in Printed Word Recognition, Phonological
and Orthographic Coding, Phoneme Awareness and IQ2
1Institute for Behavioral Genetics and Department of Psychology, University of
Colorado, Boulder, CO 80309-0447 2Supported by NICHD Grants HD-11681 and HD-27802, and
RO1 HD-22223
Address: Javier Gayán Institute for Behavioral Genetics University of Colorado Boulder, CO
80309-0447 Phone: 303 492 2817 Fax: 303 492 8063 Email: javier.gayan@colorado.edu
Data from two samples of identical and fraternal twins were used to assess genetic and environmental influences
on individual differences in error-free latent traits for printed word recognition, phonological decoding,
orthographic coding, phoneme awareness, and IQ. One group of twins (239 pairs) had been selected with at least
one member of each pair having some school history for reading problems. A second group (158 pairs) included
twins with no school history for reading problems. In both groups, a series of behavioral-genetic analyses revealed
genetic influences common to all latent traits and IQ, separate shared genetic influences for phonological awareness
and the reading constructs, and independent genetic influences on orthographic coding. Common environment
influences on IQ and each of the latent traits were small, not statistically significant, and could be dropped from the
models assuming additive genetic influence. However, models assuming non-additive genetic influences provided
an equally good fit to the data. Thus, the possible operation of non-additive genetic effects may have tended to
obscure some common environment influences on individual differences under the additive genetic models. There
were significant but generally small unique environment influences, some common and some specific to different
constructs or groups of constructs. The overall pattern for the level of genetic and environmental influences was
similar in both twin samples, although there was some suggestion that the balance of additive and non-additive
genetic influences may be different. [Poster]
Nathan Gillespie1, Katherine M. Kirk1, Andrew C. Heath
2, Ian Hickie3, and Nicholas G. Martin1
The genetic aetiology of somatic distress
1 Queensland Institute of Medical Research, Brisbane, Australia 2
Department of Psychiatry, Washington University School of Medicine, St Louis, Missouri. 3 School
of Psychiatry, University of New South Wales; the Academic Department of Psychiatry, St George Hospital and
Community Health Service; Sydney, Australia. This work was supported by NIH grants AA04535, AA07728, and
AA10249 and NHMRC (Australia) grants 941177, and 971232. We thank the twins, who were drawn from the
Australian NH&MRC Twin Registry for their cooperation.
Address: Queensland Institute of Medical Research Post Office, Royal Brisbane Hospital Brisbane
QLD 4029 Telephone: (61) 7 3362 0228 Fax: (61) 7 3362 0101 E-mail: nathanG@qimr.edu.au
Measures of anxiety, depression, phobic anxiety, somatic distress and sleep difficulty were administered in a
self-report questionnaire format to a community-based sample of 3468 Australian twins aged between 18 to 28
years. Factor analysis using an oblique rotation produced four interpretable factors: depression; phobic anxiety with
panic features; somatic distress; and sleep disturbance. These factors were subject to genetic analysis. Univariate
analysis revealed that best fitting model for the self-report symptoms of depression, phobic anxiety, and somatic
distress, for male and female twins alike, was best explained by additive genetic and specific environmental effects
(AE). Multivariate analysis showed that 38% of the genetic effects in somatic distress were due to specific gene
action unrelated to either depression or phobic anxiety. In addition, 56% of the environmental factors that influence
somatic distress were due to specific/non-shared environmental effects unrelated to either depression or phobic
anxiety. The current results lend additional support to findings that somatic symptoms demonstrate aetiologically
distinct pathways (genetic and environmental) from those of anxiety and depression.
E. R. Goldberg1, A. M. Johnson1, and P. A.
Vernon1
Effects of birth weight discordance on cognitive ability: A longitudinal infant twin study
1 Department of Psychology, The University of Western Ontario
Address: Department of Psychology Social Science Centre The University of Western Ontario, London,
Ontario, N6A 5C2 Tel: 519-661-4050 Fax: 519-661-3961 Email: erg@julian.uwo.ca
A consistent finding for small, premature children has been demonstrated in regards to cognitive ability such
that prognosis tends to worsen as birth weight and gestational length decrease. As cognitive abilities are highly
heritable, studies looking at multiple birth children who are discordant for birth weight provide an opportunity to
assess the effect of low birth weight on cognitive ability, while holding genetic factors and gestational length
constant. A sample of infant twin pairs who had birth weights more than 15% discordant received a broad range of
cognitive tests at 12 intervals between the ages of 3 months and 6 years. Cognitive ability scores across these 12
assessments were compared using a series of t-tests to determine the effect of birth-weight differences and,
specifically, to see if heavier twins demonstrated cognitive advantages over smaller twins. A longitudinal twin
design represents a unique opportunity to follow the effect of low birth weight on of cognitive abilities across
development, while controlling for other factors. Furthermore, the present study includes a much broader range of
cognitive measures than previous studies including standard psychometric measures of intelligence and language
development, as well as computer-based measures of working memory capacity and information-processing speed.
[Poster]
D. Goldman1, R.A. Kittles1, A.W. Bergen1, M.
Eggert1, M. Virkkunen1, and J. Long1
Role of the Y chromosome in alcohol dependence and related personality traits: a cladistic analysis
with eight-locus haplotypes in Finnish males
1 Lab of Neurogenetics, NIAAA
Address: Lab of Neurogenetics, NIAAA, Park Bldg Room 451, 12420 Parklawn Drive, Rockville MD
20852, Tel 301 443 0059, Fax 301 443 8579, Email dgneuro@box-d.nih.go
The extent to which genes found on the Y chromosome influence behavioral differences between males is
unknown. We used haplotype analysis to evaluate the role of Y chromosome genetic variation in alcohol
dependence and TPQ personality traits. Haplotypes of 359 psychiatrically interviewed Finnish males were
determined using alleles at seven microsatellite loci as well as the nucleotide substitution in the DYZ3 alphoid
satellite locus. An unrooted phylogeny of the 102 observed haplotypes was constructed using parsimony with a
single step mutation model. Using the algorithm of Templeton, associations of these haplotypes to behavior were
tested in a nested design starting at the terminal [0 step] ends of the Y cladogram and then defining progressively
larger groupings of Y haplotypes[1 step connected, 2 step connected, etc]. Significant association with alcohol
dependence was found for three Y haplotype clades, with significance levels of p=0.002, p=0.02 and p=0.01 for
those clades. However, there were no associations to three personality traits: harm avoidance, reward dependence
and novelty seeking, which have proposed to mediate vulnerability to alcoholism. These results, obtained with a
fully objective association design, indicate that a gene located on the Y chromosome may contain a functional
variant influencing behavior, and more specifically, vulnerability to alcoholism.
Julia D. Grant1, Kathleen K. Bucholz1, Wendy S.
Slutske2, Tara L. McLaughlin1, Pamela A.F. Madden1, and Andrew C.
Heath1
Genetic and environmental associations between childhood conduct problems and adolescent
marijuana use3
1Department of Psychiatry, Washington University School of Medicine, 40 N.
Kingshighway, Suite 1, St. Louis, MO 2University of Missouri, Columbia, MO. 3
Supported by AA09022, AA07728, DA07261, DA00272
Address: Department of Psychiatry, Washington University School of Medicine, 40 N. Kingshighway,
Suite 1, St. Louis, MO 63108 Phone: (314) 286-2299 FAX: (314) 286-2213 Email: julie@matlock.wustl.edu
Previous research has indicated that conduct problems in childhood are associated with increased substance use
during adolescence (e.g., S. Miller-Johnson, J.E. Lochman, J.D. Coie, R. Terry, & C. Hyman, 1998, J. Abnorm.
Child Psychol., 26, 221-232). However, the extent to which genetic and environmental influences on
these measures overlap has not been investigated. In the present analyses we examined the relationship between
self-reported conduct problems and self- reported marijuana use in a sample of adolescent female twins. Data from
916 twin pairs (MZ=530, DZ=386; mean age at follow-up=16.95 years) who participated in both the initial
telephone diagnostic interview and the one-year brief follow-up assessment of the Missouri Adolescent Female
Twin Study were analyzed. Conduct problems were measured on a continuous scale (M=3.43; range=0-40).
Marijuana use was analyzed on a 3 point scale: never tried marijuana (N=1321), tried marijuana once (N=121), used
marijuana multiple times (N=390). Mean age of first marijuana use was 15.68 years. Correlational analyses
indicated familiality for both measures: Pearson correlations for conduct problems were MZ=.75, DZ=.48;
polychoric correlations for marijuana use were MZ=.82, DZ=.63. A bivariate Cholesky model was tested using Mx.
Additive genetic and shared environmental influences were significant for both measures. For conduct problems,
additive genetic influences accounted for 49% of the variance, shared environmental influences accounted for 26%
of the variance, and nonshared environmental influences accounted for 25% of the variance. For marijuana use,
additive genetic, shared environmental, and nonshared environmental influences accounted for 36%, 46%, and 18%
of the variance respectively. The genetic correlation between the measures was .52 and the shared environmental
correlation was .82.
James C. Ha1
Heritability of some common measures of cognitive and reflex development in infant pigtailed
macque monkeys
1 Regional Primate Research Center and Psychology Department, University of
Washington
Address: Regional Primate Center, University of Washington, Box 357330, Seattle WA 98195-7330
phone 206-543-2420 fax 206-685-8606 email jcha@u.washington.edu
The University of Washington's Infant Primate Research Laboratory has maintained an infant primate nursery
since 1972, and has developed a battery of assessments for infant primate development, modified from those used to
monitor child development. The nursery animals are part of a pedigreed breeding colony of pigtailed macaque
monkeys. The pedigree now contains 11,000+ animals, and provides an invaluable tool for quantitative genetic
analysis of the phenotype data maintained in our large computer records system, including the infant primate
development measures. I will present the results of a series of variance-component analyses of heritability on
several data sets, including birthweight (a strong predictor of later psychological delays), reflex development tests
(e.g. sucking, grasping, visual orientation, clasping: h2 = 0.0 to 0.64), and object permanence
development tests (h2 = 0.0 to 0.40). Future plans will also be described.
Norman D. Henderson1, M. Bohl2, C. Cykowski2
and J. C. DeFries2
Albino gene effects on fear-related behaviors of male and female mice 3
1 Department of Psychology, Oberlin College, Oberlin, OH 44074 2
Institute for Behavioral Genetics, University of Colorado, Boulder, CO 80309 3 Supported by NIMH
Grant MH-53480 and by an Oberlin College Research Status Award to first author
Address: Department of Psychology, Oberlin College, Oberlin, OH 44074 Phone: (303)517-2486 FAX:
(440)775-8356 e-mail: fhenders@oberlin.edu
Although DeFries, Hegmann and Weir (1966, Science, 154, 1577-1579) found that albino
mice from segregating F2-F4 generations had lower activity and higher defecation scores than pigmented mice in a
brightly lighted open field, Flint et al. (1995, Science, 269, 1432-1435) were unable to localize a
QTL for these measures at or near the albino locus. We assessed coat color and sex differences in fear-related
behaviors in 577 F2 mice derived from strains originally selected for high and low open-field activity prior to
inbreeding (DeFries, Gervais, and Thomas, 1978, Behavior Genetics, 8, 3-13). All animals were
tested in a battery consisting of an open field, a light-dark box, a mirror chamber, an elevated plus maze and an
elevated square maze. These tests vary substantially in environmental conditions and each generates several putative
measures of fear or anxiety in mice and rats. As expected, significant coat color differences were found for several
open field behaviors. Albino mice had significantly less total activity and spent less time in the center of the brightly
lit field, and defecated significantly more than pigmented animals. In addition, however, albino mice spent less time
and exhibited less locomotor activity and less scanning over the open arms of the plus and square mazes, which are
both tested under low illumination. In contrast, albino mice were slightly more active in the enclosed arms of both
the plus and square mazes as well as in the light-dark box and mirror chamber box. In all cases albino effects were
highly significant but too small to account for the magnitude of the behavioral differences observed between the
High- and Low-selected parent strains. Estimates of h2 due to segregation at the albino locus ranged from .02 to .06
for activity- and time-based measures of fear/anxiety and .01 or less for defecation measures. Although female mice
were significantly less active and had lower defecation and urination scores than males in most tests, no significant
Sex by Albino interactions were observed.
J. Dee Higley1, and Allyson Bennett 1
Impaired CNS Serotonin Functioning, Excessive Alcohol Intake, and Aggression: A Nonhuman
Primate Model of Genetic and Environmental Influences
1 Laboratory of Clinical Studies - Primate Unit, NIH Animal Center, NIAAA,
Poolesville, MD 20837
Address: Address NIH Animal Center, PO Box 529, Building 112, Poolesville, MD 20837, Telephone
301-496-9550; FAX 301-496-0630, email- higleyd@exchange.nih.gov
Studies show that rhesus macaques with low CNS serotonin functioning, as measured by low concentrations of
cerebrospinal fluid (CSF) 5-hydroxyindoleacetic acid (5-HIAA), are impulsive, consume alcohol excessively, are
socially ostracized, and exhibit high levels of aggression (Higley et al., 1998, Annals of the New York Academy of
Sciences, 836, 39-56). One of the most replicated studies in our laboratory is that interindividual differences in CSF
5-HIAA concentrations are trait-like, showing stability across time and situations. We will present data showing that
maternal and paternal genetic influences play major roles beginning early in life in producing low CSF 5-HIAA
concentrations. Such differences are further exacerbated by early deleterious rearing experiences, particularly
parental deprivation. We have recently analyzed molecular genetic data to investigate the relationship between
serotonin transporter (5-HTT) genotypes, CSF 5-HIAA concentrations, and alcohol-related aggression. Our results
show that monkeys with the short (s) allele 5-HTT variant have low CSF 5-HIAA concentrations, consume alcohol
in excess, and are particularly aggressive when intoxicated. However, the phenotypic expression of this s variant is
environmentally-dependent, with the s variant affecting CSF 5-HIAA concentrations and alcohol consumption only
in subjects reared by age-mates, without adult influence.
R. Hitzemann1, K. Demarest1, J. Koyner1, L.
Cipp1, B. Hitzemann1, and J. McCaughran1
Identification of QTLs for Saline and Ethanol-Induced Locomotor Responses 2
1 Departments of Psychiatry and Neurobiology, SUNY at Stony Brook, Stony
Brook, NY 11794-8101 and Research and Psychiatry Services, VAMC, Northport, NY 11768
2Supported in part by MH-51372, AA-11043 and the Department of Veterans Affairs
Address: Robert Hitzemann, Ph.D., Department of Psychiatry, SUNY at Stony Brook, Stony Brook,
NY 11794-8101; Tel. (516) 444-2903; E-mail: rhitzemann@mail.psychiatry.sunysb.edu
We have phenotyped 25 strains of the BXD recombinant inbred (RI) series (N = 13/strain) and 1825 C57BL/6J
x DBA/2J F2 intercross animals for the saline and the ethanol (1.5 g/kg) induced (difference score) responses. Data
were collected as the distance traveled, under normal laboratory lighting, in four 5 min blocks beginning
immediately after drug administration. The RI data confirm that both the saline and ethanol responses have a high
split halves reliability (> 0.8) and significant heritability (0.38 and 0.32, respectively). There was no significant
correlation between the saline and ethanol responses (p > 0.2); however, there was a significant correlation between
the ethanol response and both chlordiazepoxide-induced activity (r = 0.60, p < 0.01) and acute ethanol-withdrawal
(r = 0.58, p < 0.01). Analysis of the RI strain means revealed significant (p < 0.01) saline QTLs on chromosomes 1,
3, 5, 10, 13, and 18 and significant ethanol QTLs on chromosomes 2, 4, 6, and 9. Between 300 and 600 of the F2
animals were phenotyped in a pseudo-random fashion for microsatellite markers spaced at approximately 20 cM
across the genome. QTLs exceeding the LOD threshold of 4.3 were found on chromosomes 1 (saline) and
chromosomes 2 (ethanol), complimenting the RI data. Numerous suggestive QTLs (LOD 2.5 to 3.5) were also
detected for both phenotypes; these QTLs largely exhibited dominant allele effects. Heterogeneous stock (HS) mice
(8-way cross, G24) (N = 200) were used for fine mapping. A chromosome 1 saline QTL was mapped to 57 +/- 1.5
cM (EUCIB panel) and accounted for 11% of the phenotypic variance; a chromosome 2 ethanol QTL mapped to 56
+/- 2 cM and accounted for 8% of the phenotypic variance. The saline QTL compliments previous work (Flint et al.
1995; Gershenfeld et al. 1997) demonstrating open-field activity QTLs on chromosome 1. The ethanol QTL is in the
same general region of chromosome 2 where QTLs have been identified for seizure sensitivity (Frankel et al. 1995)
and acute ethanol withdrawal (Buck et al. 1997). Using a variety of behavioral genetic techniques, we have
identified the central nucleus of the amygdala (CeA) as a locus for the variation in ethanol response (e.g. Hitzemann
and Hitzemann, 1997). Congenic strains (under construction) will be used to determine what are the relationships
among ethanol response, chromosome 2 QTL(s) and the CeA.
Cathleen B. Hunt1, Alexander Weiss1, J. E. King1
Is malevolence or proteanism heritable in Chimpanzees (Pan troglodytes)?2
1 Department of Psychology, The University of Arizona, Tucson, AZ 85721.
2 Supported by ChimpanZoo and the Jane Goodall Institute
Address: Cathleen B. Hunt, Department of Psychology, The University of Arizona, Tucson, AZ 85721.
(520) 319-1087 huntc@u.arizona.edu
Recent research shows that two chimpanzee personality traits, dominance and dependability, are heritable (A.
Weiss and J.E. King, 1998, BGA Meeting, Stockholm). However, further examination of the dependability factor
indicates that it may be composed of two sub-factors, malevolence and proteanism. Four item descriptors of
dependability: jealous, defiant, aggressive and irritable, describe a malevolent personality dimension. Three other
items that describe dependability: erratic, (un)predictable, and disorganized, capture what is meant by "proteanism."
Some evolutionary theorists suggest that disorganized and unpredictable (protean) behavior can be adaptive in
primate courtship and competition and that different levels of proteanism may be maintained in a population by
means of frequency dependent selection (G.F. Miller, 1997, in A. Whiten and R.W. Byrne, eds., Machiavellian
Intelligence II, Cambridge University Press, Cambridge, UK). If this is true, then proteanism should be
heritable. As expected, principal factor analysis with promax rotation of the nine dependability items found two
correlated factors, malevolence and proteanism. We created factor scores by unit-weighting items that loaded on the
factor at .57 or greater. Two items, impulsive and reckless, did not meet this criteria and were removed from final
analyses. The Symmetric Differences Squared (SDS) method was used to estimate genetic, shared zoo, and
non-shared environmental variance components of malevolent and protean traits among 145 zoo chimpanzees (L.W.
Grimes and W.R. Harvey, 1980, Journal of Animal Science, 50, 634-644). Non-shared
environmental effects accounted for almost all of the variance in malevolence. As hypothesized, proteanism was
highly heritable with negligible shared zoo effects. These results indicate that malevolent characteristics in
chimpanzees are primarily defined by their unique experiences with their zoo environment and conspecifics. The
heritability of proteanism supports Miller's assertion that varying levels of unpredictability can be adaptive and may
be maintained by frequency dependent selection in primate populations. [Poster]
Kristen C. Jacobson1, and D.C. Rowe1
Sibling influences on adolescent antisocial behavior
1Department of Family Studies, University of Arizona, Tucson, Arizona 85721
Address: P.O. Box 210033; The University if Arizona; Tucson, AZ 85705. Phone: (520) 621-7127.
Fax: (520) 621-3401. Email: kjacobso@u.arizona.edu
The present study investigated whether the amount of time siblings spend together moderated the heritability of
adolescent antisocial behavior. Data are from the Sibling Pairs sample of the National Longitudinal Study of
Adolescent Health (Add Health). Only same-sex siblings less than 2 years apart in age were used in these analyses.
The average age of siblings was 16. Adolescent self-reports of antisocial behavior were divided into two scales:
aggression (4 items) and non-violent delinquency (11 items). The time together variable was created by averaging
sibling reports of how much time they spent with one another and how much time they spent with the same group of
friends. Standard regression analyses indicated that time together moderated sibling similarity for delinquency, but
not aggression. Siblings who spent more time together were more alike in delinquent behavior. Based on results
from DeFries-Fulker regressions (J.C. DeFries, and D.W. Fulker, 1985, Behav. Genet.15,
467-473), the heritability of aggression was .20, and the heritability of delinquency was .18. Shared environmental
influences accounted for 18% of the variation in aggression, and 24% of the variation in delinquency. An
augmented DeFries-Fulker model indicated that time together moderated the heritability of delinquency, but not
aggression. Time together did not moderate the shared environmental influence on either aggression or delinquency.
Follow-up analyses using Mx revealed that the heritability of delinquent behavior was .34 among
adolescents who spent more time together; genetic influences did not account for any of the variation in delinquency
among siblings who spent less time together. This model fit the data well (
2 = 29.8, df = 24,
p > .10). Shared environmental influences in both groups were estimated at .21. Equating the parameters
across the two groups resulted in a significantly poorer fit (
2 = 17.5, df = 3, p <
.001).
Kerry L. Jang1, W. John Livesley1, and Philip A.
Vernon2
Antisocial Personality, Family Environment, and Alcohol Misuse
1Department of Psychiatry, University of british Columbia, Vancouver, Canada
2Department of Psychology, University of Western Ontario, London, Canada
Address: Dr. Kerry L. Jang Department of Psychiatry University of British Columbia 2255 Wesbrook
Mall Vancouver, BC Canada. V6T 2A1 Voice: (604) 822-7895 Fax: (604) 822-7756 e-mail: kjang@unixg.ubc.ca
The extent to which specific facets of antisocial personality pathology, family environment and alcohol misuse
share a common genetic and environmental aetiology was estimated. Participants were 347 adult monozygotic twin
pairs (209 sister & 138 brother pairs) and 346 dizygotic twin pairs (170 sister, 82 brother, & 94 sister-brother pairs).
All twin pairs completed self-report measures of alcohol misuse and personality pathology contained in the
Dimensional Assessment of Personality Pathology (Livesley, W.J. & Jackson, D.N.,in press, Manual for the
Dimensional Assessment of Personality Problems-Basic Questionnaire. Port Huron, MI, Sigma)and family
environment (Family Environment Scale: Moos, R.H. & Moos, B.S.,1986, Manual: Family Environment Scale, Palo
Alto, Consulting Psychologists Press). Bivariate genetic correlations between the scales identified a number of the
specific aspects of antisocial personality, such as recklessness, impulsivity, and interpersonal hostility that shared a
common genetic basis with alcohol misuse. Other characteristics typical of antisocial personalit