Demographic, Familial, Social, and Mood Data
The groups were matched for age, verbal intelligence, and duration of formal education (), but they differed with regard to gender, as the majority of drug-dependent individuals were men. The shared familial environment in the sibling pairs, which lasted in most cases into adolescence, was notably different from the childhood familial environment reported by healthy volunteers; the sibling pairs had larger household sizes, a higher incidence of parental divorce and of parents with addiction problems, and more frequent experiences of childhood abuse.
Task performance differed significantly over the four cognitive domains (executive function, visual memory, attention control, and response control) across the three groups, as reflected by a highly significant domain-by-group interaction (F=6.7, df=2.6, 371.7, p<0.001), refuting the null hypothesis that all cognitive impairments were attributable to deficits in general intelligence. No significant main effects of domain were found, but there was a significant main effect of group (F=9.6, df=2, 142, p<0.001). To investigate the nature of the interaction, we compared the groups separately on each domain, revealing significant results for the domains of executive function (F=16.0, df=2, 144, p=0.004) and response control (F=5.9, df=2, 144, p=0.012). shows that the difference between groups in terms of visual memory and attention did not reach significance. Post hoc comparisons revealed that both siblings and drug-dependent individuals differed significantly from healthy volunteers in executive function (tsibs=2.6, p=0.012; tdrug=8.8, p<0.001) and from each other (t=5.9, p<0.001). In response control, the sibling pairs differed significantly from healthy volunteers (tsibs=4.0, p<0.001; tdrug=3.6, p<0.001) but not from each other. Parental drug or alcohol abuse was associated with aggravated performance in executive function and visual memory only in drug-dependent individuals but not in their siblings.
Cognitive and Emotional Profiles of Participants in a Study of Endophenotypes for Drug Dependencea
Affective and Personality Domains
We observed a significant domain-by-group interaction for affective and personality domains (F=23.0, df=2, 45, 356, p<0.001), providing support for the hypothesized specificity of the domains. We observed a significant main effect of group (F=10.2, df=2, 45, 356, p<0.001) but not a significant main effect of domain. Group comparisons conducted separately for each domain revealed significant group differences in all four domains: emotional (F=37.3, df=2, 145, p<0.001), psychosocial (F=8.7, df=2, 146, p<0.001), impulsivity-compulsivity (F=36.7, df=2, 146, p<0.001), and self-evaluation (F=11.8, df=2, 146, p<0.001). As shown in , both siblings and drug-dependent individuals differed significantly from healthy comparison volunteers in emotional functioning (tsibs=2.8, p=0.007; tdrug=8.4, p<0.001), impulsivity-compulsivity (tsibs=−2.6, p=0.012; tdrug=−8.9, p<0.001), and self-evaluation traits (tsibs=2.0, p=0.044; tdrug=5.3, p<0.001). On psychosocial functioning, the siblings were not different from healthy volunteers, but data from drug-dependent individuals revealed significant psychosocial impairment (t=4.4, p<0.001). further illustrates that the impairments in all four domains were exacerbated in the drug-dependent individuals compared with their siblings (emotional: t=−5.5; psychosocial: t=−3.4; impulsivity-compulsivity: t=6.5; self-evaluation: t=3.2; p<0.005 in all cases). Parental drug or alcohol abuse did not affect emotional function and personality traits in the sibling pairs.
Familial Relatedness and Traumatic Childhood
The difference between sibling pairs and healthy volunteers reached significance in executive function and response control. We compared the within-pair variance between the biological sibling pairs in these two measures with the variance between random sibling pairs (46
). The observed variance on response control was significantly smaller within the biological pairs compared with the randomly permutated distribution (p=0.006), indicating that impairment in response control is a shared trait between family members. For executive function, variances in biological pairs did not significantly differ from those in random sibling pairs, suggesting that impairments in executive function are not familial. Biological sibling pairs also differed significantly from unrelated healthy volunteers in emotional function, impulsivity-compulsivity traits, and self-evaluation, but the variances within these biological pairs did not significantly differ in any of these domains from variances within randomly permutated pairs. Comorbid dependence on opioids was not related to a different cognitive or personality profile in the current sample.
The sibling pairs reported significantly higher levels of childhood trauma compared with unrelated healthy volunteers (). We used a composite variable across all of the three types of abuse (49
) to explore the relationship within the sibling pairs’ traumatic childhood and the cognitive and personal domain measures. Executive function (r=−0.3, p<0.05) and impulsive-compulsive traits (r=0.3, p<0.05) were significantly associated with the degree of childhood abuse in the sibling pairs. The significant relationship between childhood abuse and emotional function, however, did not survive correction for multiple comparisons.
Cognitive Profiles and Indexes of Drug Abuse
Correlational analyses between last stimulant use, duration of stimulant abuse, and cognitive or personality domain measures did not reach significance. Compulsive pattern of stimulant use was significantly associated with emotional functioning (r=−0.5, p<0.05) and self-evaluation (r=0.5, p<0.05). Relationships between alcohol use (as reflected by the AUDIT score) and cognitive/personality measures did not survive correction for multiple comparisons. No statistically significant relationships were observed between cognitive performance and the duration of cannabis use or the age at onset. Trait impulsivity and anxiety traits were associated with each other in all volunteers (r=0.5, p<0.05). Cognitive and personality domains correlated with each other; a full correlation matrix can be found in the online data supplement.