Kees-Jan Kan

The Young Netherlands Twin Register (YNTR): Longitudinal Twin and Family Studies in Over 70,000 Children

The Netherlands Twin Register (NTR) began in 1987 with data collection in twins and their families, including families with newborn twins and triplets. Twenty-five years later, the NTR has collected at least one survey for 70,784 children, born after 1985. For the majority of twins, longitudinal data collection has been done by age-specific surveys. Shortly after giving birth, mothers receive a first survey with items on pregnancy and birth. At age 2, a survey on growth and achievement of milestones is sent. At ages 3, 7, 9/10, and 12 parents and teachers receive a series of surveys that are targeted at the development of emotional and behavior problems. From age 14 years onward, adolescent twins and their siblings report on their behavior problems, health, and lifestyle. When the twins are 18 years and older, parents are also invited to take part in survey studies. In sub-groups of different ages, in-depth phenotyping was done for IQ, electroencephalography , MRI, growth, hormones, neuropsychological assessments, and cardiovascular measures. DNA and biological samples have also been collected and large numbers of twin pairs and parents have been genotyped for zygosity by either micro-satellites or sets of short nucleotide polymorphisms and repeat polymorphisms in candidate genes. Subject recruitment and data collection is still ongoing and the longitudinal database is growing. Data collection by record linkage in the Netherlands is beginning and we expect these combined longitudinal data to provide increased insights into the genetic etiology of development of mental and physical health in children and adolescents.

Genetic and Environmental Stability in Attention Problems Across the Lifespan: Evidence From the Netherlands Twin Register

OBJECTIVE To review findings on attention-deficit/hyperactivity disorder and attention problems (AP) in children, adolescents, and adults, as established in the database of the Netherlands Twin Register and increase the understanding of stability in AP across the lifespan as a function of genetic and environmental influences. METHOD A longitudinal model was fitted on Netherlands Twin Register AP scores from 44,607 child (<12-year-old), adolescent (12- to 18-year-old), and adult (>18-year-old) twins. RESULTS Mean AP showed a downward trend with age. Age-to-age correlations ranged from 0.33 (50-≥60 years old) to 0.73 (10-12 years old). Stability in individual differences in AP was due to genetic and environmental factors, and change was due primarily to environmental factors. Nonadditive genetic influences were present from childhood to adulthood. Total genetic variance decreased slightly throughout aging, whereas environmental variance increased substantially with the switch from maternal to self-ratings at 12 years of age. As a result, heritability coefficients decreased from 0.70 to 0.74 in childhood (maternal ratings) to 0.51 to 0.56 in adolescence (self-ratings), and 0.40 to 0.54 in adulthood (self-ratings). In childhood, male subjects scored higher than female subjects. After the rater switch at 12 years of age, female subjects tended to score higher than male subjects. CONCLUSIONS Stability of AP is the result of genetic and environmental stability. The decrease in estimated heritability at 12 years of age is due to an increase in occasion-specific environmental variance and likely reflects a methodologic effect. Because environmental influences have lasting effects on AP, their early detection is crucial.

On the Nature and Nurture of Intelligence and Specific Cognitive Abilities: The More Heritable, the More Culture Dependent

To further knowledge concerning the nature and nurture of intelligence, we scrutinized how heritability coefficients vary across specific cognitive abilities both theoretically and empirically. Data from 23 twin studies (combined N = 7,852) showed that (a) in adult samples, culture-loaded subtests tend to demonstrate greater heritability coefficients than do culture-reduced subtests; and (b) in samples of both adults and children, a subtest’s proportion of variance shared with general intelligence is a function of its cultural load. These findings require an explanation because they do not follow from mainstream theories of intelligence. The findings are consistent with our hypothesis that heritability coefficients differ across cognitive abilities as a result of differences in the contribution of genotype-environment covariance. The counterintuitive finding that the most heritable abilities are the most culture-dependent abilities sheds a new light on the long-standing nature-nurture debate of intelligence.

Stability in symptoms of anxiety and depression as a function of genotype and environment: a longitudinal twin study from ages 3 to 63 years.

BACKGROUND The influence of genetic factors on major depressive disorder is lower than on other psychiatric disorders. Heritability estimates mainly derive from cross-sectional studies, and knowledge on the longitudinal aetiology of symptoms of anxiety and depression (SxAnxDep) across the lifespan is limited. We aimed to assess phenotypic, genetic and environmental stability in SxAnxDep between ages 3 and 63 years. METHOD We used a cohort-sequential design combining data from 49 524 twins followed from birth to age ⩾20 years, and from adolescence into adulthood. SxAnxDep were assessed repeatedly with a maximum of eight assessments over a 25-year period. Data were ordered in 30 age groups and analysed with longitudinal genetic models. RESULTS Over age, there was a significant increase during adolescence in mean scores with sex differences (women>men) emerging. Heritability was high in childhood and decreased to 30-40% during adulthood. This decrease in heritability was due to an increase in environmental variance. Phenotypic stability was moderate in children (correlations across ages ~0.5) and high in adolescents (r = 0.6), young adults (r = 0.7), and adults (r = 0.8). Longitudinal stability was mostly attributable to genetic factors. During childhood and adolescence there was also significant genetic innovation, which was absent in adults. Environmental effects contributed to short-term stability. CONCLUSIONS The substantial stability in SxAnxDep is mainly due to genetic effects. The importance of environmental effects increases with age and explains the relatively low heritability of depression in adults. The environmental effects are transient, but the contribution to stability increases with age.

Nonlinear epigenetic variance: review and simulations

We present a review of empirical evidence that suggests that a substantial portion of phenotypic variance is due to nonlinear (epigenetic) processes during ontogenesis. The role of such processes as a source of phenotypic variance in human behaviour genetic studies is not fully appreciated. In addition to our review, we present simulation studies of nonlinear epigenetic variance using a computational model of neuronal network development. In each simulation study, time series for monozygotic and dizygotic twins were generated and analysed using conventional behaviour genetic modelling. In the results of these analyses, the nonlinear epigenetic variance was subsumed under the non-shared environmental component. As is commonly found in behaviour genetic studies, observed heritabilities and unique environmentabilities increased with time, whereas common environmentabilities decreased. The fact that the phenotypic effects of nonlinear epigenetic processes appear as unsystematic variance in conventional twin analyses complicates the identification and quantification of the ultimate genetic and environmental causes of individual differences. We believe that nonlinear dynamical system theories provide a challenging perspective on the development of individual differences, which may enrich behaviour genetic studies.

Intelligence and the brain: a model-based approach

Various biological correlates of general intelligence (g) have been reported. Despite this, however, the relationship between neurological measurements and g is not fully clear. We use structural equation modeling to model the relationship between behavioral Wechsler Adult Intelligence Scale (WAIS) estimates of g and neurological measurements (voxel-based morphometry and diffusion tensor imaging of eight regions of interest). We discuss psychometric models that explicate the relationship between g and the brain in a manner in line with the scientific study of g. Fitting the proposed models to the data, we find that a MIMIC model (for multiple indicators, multiple causes), where the contributions of different brain regions to a unidimensional g are estimated separately, provides the best fit against the data.

Assessing genetic influences on behavior: informant and context dependency as illustrated by the analysis of attention problems.

Assessment of genetic influences on behavior depends on context, informants, and study design: We show (analytically) that, conditional on study design, informant specific genetic variance is included in the genetic variance component or in the environmental variance component. To aid the explanation, we present an illustrative empirical analysis of data from the Netherlands Twin Register. Subjects included 1,571 monozygotic and 2,672 dizygotic 12-year-old twin pairs whose attention problems (AP) were rated by their parents, teachers, and themselves. Heritability estimates (h2) of AP were about ~0.75 for same informant ratings (mother, father, and same teacher ratings) and ~0.54 for different informants’ ratings (different parents’, different teachers’, and two twins’ self-ratings). Awareness of assessment effects is relevant to research into psychiatric disorders. Differences in assessment can account for age effects, such as a drop in heritability of ADHD symptoms. In genome-wide association studies, effects of rating specific genetic influences will be undetectable.

Brain Regions Related to Impulsivity Mediate the Effects of Early Adversity on Antisocial Behavior.

BACKGROUND Individual differences in impulsivity and early adversity are known to be strong predictors of adolescent antisocial behavior. However, the neurobiological bases of impulsivity and their relation to antisocial behavior and adversity are poorly understood. METHODS Impulsivity was estimated with a temporal discounting task. Voxel-based morphometry was used to determine the brain structural correlates of temporal discounting in a large cohort (n = 1830) of 14- to 15-year-old children. Mediation analysis was then used to determine whether the volumes of brain regions associated with temporal discounting mediate the relation between adverse life events (e.g., family conflict, serious accidents) and antisocial behaviors (e.g., precocious sexual activity, bullying, illicit substance use). RESULTS Greater temporal discounting (more impulsivity) was associated with 1) lower volume in frontomedial cortex and bilateral insula and 2) greater volume in a subcortical region encompassing the ventral striatum, hypothalamus and anterior thalamus. The volume ratio between these cortical and subcortical regions was found to partially mediate the relation between adverse life events and antisocial behavior. CONCLUSIONS Temporal discounting is related to regions of the brain involved in reward processing and interoception. The results support a developmental imbalance model of impulsivity and are consistent with the idea that negative environmental factors can alter the developing brain in ways that promote antisocial behavior.

The Computerized Neurocognitive Battery: Validation, aging effects, and heritability across cognitive domains.

OBJECTIVE The Computerized Neurocognitive Battery (CNB) enables efficient neurocognitive assessment. The authors aimed to (a) estimate validity and reliability of the batterys Dutch translation, (b) investigate effects of age across cognitive domains, and (c) estimate heritability of the CNB tests. METHOD A population-representative sample of 1,140 participants (aged 10-86), mainly twin-families, was tested on the CNB, providing measures of speed and accuracy in 14 cognitive domains. In a subsample (246 subjects aged 14-22), IQ data (Wechsler Intelligence Scale for Adults; WAIS) were available. Validity and reliability were assessed by Cronbachs alpha, comparisons of scores between Dutch and U.S. samples, and investigation of how a CNB-based common factor compared to a WAIS-based general factor of intelligence (g). Linear and nonlinear age dependencies covering the life span were modeled through regression. Heritability was estimated from twin data and from entire pedigree data. RESULTS Internal consistency of all tests was moderate to high (median = 0.86). Effects of gender, age, and education on cognitive performance closely resembled U.S. SAMPLES The CNB-based common factor was completely captured by the WAIS-based g. Some domains, like nonverbal reasoning accuracy, peaked in young adulthood and showed steady decline. Other domains, like language reasoning accuracy, peaked in middle adulthood and were spared decline. CNB-test heritabilities were moderate (median h2 = 31%). Heritability of the CNB common factor was 70%, similar to the WAIS-based g-factor. CONCLUSION The CNB can be used to assess specific neurocognitive performance, as well as to obtain a reliable proxy of general intelligence. Effects of aging and heritability differed across cognitive domains.

Network Models for Cognitive Development and Intelligence

Cronbach’s (1957) famous division of scientific psychology into two disciplines is still apparent for the fields of cognition (general mechanisms) and intelligence (dimensionality of individual differences). The welcome integration of the two fields requires the construction of mechanistic models of cognition and cognitive development that explain key phenomena in individual differences research. In this paper, we argue that network modeling is a promising approach to integrate the processes of cognitive development and (developing) intelligence into one unified theory. Network models are defined mathematically, describe mechanisms on the level of the individual, and are able to explain positive correlations among intelligence subtest scores—the empirical basis for the well-known g-factor—as well as more complex factorial structures. Links between network modeling, factor modeling, and item response theory allow for a common metric, encompassing both discrete and continuous characteristics, for cognitive development and intelligence.