Twin and family studies on the development of cognitive and externalizing problems

Abstract

Cognitive and externalizing problems are responsible for much personal suffering, as well as large monetary costs for society. Intervention and prevention efforts have often failed in reduction of unwanted behaviors, perhaps due to lack of understanding of the development of these complex traits. Studies on risk factors often treat the associations naively by not considering potential unmeasured common causes of the risk factor and the outcome. These common causes may be shared within families; i.e., the association is subject to familial confounding. Analyses informed of family belonging can help to further the knowledge regarding causality. Therefore, in this thesis, I used existing, and developed novel, methodologies to assess familial confounding. Models to adjust for familial confounding, as well as models identifying sources of familial confounding, were implemented. Further, I developed a genetically sensitive longitudinal design with multiple raters and time-points.

In study I, advancing paternal age was associated with offspring violent offending. Advancing paternal age was found to increase the incidence of violent criminal convictions among re-offending offspring when siblings were compared; a result congruent with causal inference. Contrary, and congruent with non-causal inference, advancing paternal age did not increase the probability to ever be convicted a violent criminal offence.

Study II identified an association between maternal smoking during pregnancy (SDP) and offspring stress coping in late adolescence. However, the association did not persist when exposure-discordant siblings were compared. This result is compatible with anon-causal interpretation, and suggests that the association is due to familial confounding; other factors shared between siblings accounts for the association. In a quantitative genetic analysis these factors were found to be of genetic origin.

Study III continued the investigation of SDP, this time as a risk factor for cognitive outcomes (general cognitive ability and poor academic achievement), externalizing outcomes (criminal convictions, violent criminal convictions, and drug misuse), and pregnancy related outcomes (birth weight, preterm birth, and born small for gestational age). SDP was associated with all outcomes, but within-sibling analyses found that the association persisted only for pregnancy related outcomes, and disappeared for cognitive and externalizing outcomes. Quantitative genetic analyses found that genetics accounted for the majority of the association between SDP and cognitive and externalizing outcomes.

In study IV externalizing traits in mid-childhood were found to predict ADHD-like traits in adolescence when pre-existing associations were adjusted for. Further, ADHD- like traits in late adolescence predicted externalizing behavior in young adult age. The two traits were correlated in mid-childhood (age 8-9), and become even more correlated through early (age 13-14) and late (age 16-17) adolescence and young adult age (age 19-20). Stable and new factors accounted for approximately half of the correlation between the traits each, throughout development. Genetic variation explained two thirds of the correlations.

In conclusion, advancing paternal age might increase the rate of violent offending among violent re-offenders. SDP does not seem to be causing offspring cognitive and externalizing problems in adolescence and adulthood; the observed associations are better explained by shared genetic factors. Externalizing behavior predicts ADHD early in development, and ADHD predicts externalizing behavior late in development. And, although new sources of covariance arise throughout development, ADHD-like and externalizing traits become even more correlated from childhood to young adulthood.