Prenatal cannabis exposure associated with altered brain connectivity: neural correlates of cognitive and mental health variability in offspring

Emerging evidence from human and preclinical studies suggests that prenatal cannabis exposure (PCE) may be associated with offspring neurodevelopmental outcomes. Although prior studies have linked PCE to cognitive and mental health differences in offspring, the large-scale functional brain network organization associated with these outcomes remains incompletely understood. In this study, we leveraged baseline data from the ongoing Adolescent Brain Cognitive Development (ABCD) Study, which enrolled more than 11,000 children aged 9–10 years. We first examined associations between PCE, occurring before and after maternal awareness of pregnancy, and offspring psychopathology and cognitive performance. Second, we evaluated associations between PCE and resting-state functional network connectivity (FNC). Third, we examined associations between FNC and behavioral outcomes. Fourth, spatial correlation analyses were performed to quantify the correspondence between PCE-related and outcome-related FNC patterns. Resting-state functional MRI data were analyzed using the NeuroMark framework to identify individualized networks and estimate FNC among them. Linear mixed-effects models were used, controlling for familial, pregnancy-related, and child-specific covariates. Among 10,838 children (female/male = 5,200/5,638; mean age = 9.96 ± 0.62 years), 751 (6.93%) were prenatally exposed to cannabis. Compared with non-exposed peers, exposed children exhibited higher levels of psychopathology and significant differences across multiple cognitive domains, including lower performance on most cognitive measures but higher composite fluid cognition scores (Cohen’s d = –0.1393 to 0.2451, FDR-corrected p < 0.05). PCE was also associated with altered FNC patterns. Importantly, these PCE-related FNC patterns significantly corresponded with FNC patterns associated with mental health symptoms (positive spatial correlations between t -statistic maps, r = 0.0641 to 0.5993, FDR-corrected p < 0.05) and cognitive performance (inverse spatial correlations, r = –0.5438 to –0.6665, FDR-corrected p < 0.05). Collectively, our findings indicate that PCE is associated with alterations in large-scale functional brain network connectivity in late childhood. PCE-associated connectivity alterations showed concordant relationships with adverse offspring outcomes, identifying candidate connectivity features that warrant further investigation as potential neurobiological pathways linking PCE with offspring neurodevelopment. Not all connectivity patterns associated with offspring behavioral outcomes were related to PCE, suggesting that only a subset of brain–behavior relationships may represent candidate pathways linking PCE with offspring neurodevelopment.