Independent brain cortical signatures of risk for adolescent cannabis use and consequences of such use are moderated by sex.

There is an accumulation of evidence linking adolescent cannabis use with variations in brain structure and function, however it remains poorly understood whether cannabis-associated variations in brain structure represent pre-existing risk factors or consequences of cannabis use. We investigated whether cannabis use propensity and within-person variations of cannabis use were associated with cortical thickness during adolescence. Adolescents (n = 136, 74 female) completed three neuroimaging sessions and annual assessments from 12 until 17 years of age (with 90% follow-up). Cannabis use was disaggregated into between- (vulnerability) and within-person (time-varying) components using longitudinal multi-level modelling, controlling for age, sex and alcohol use. Across the whole sample, cortical thickness was lower in years when participants' cannabis use exceeded their own average level of cannabis use (F1,25663.3 = 3.96, p = 0.047; mean: -0.0023 mm/once-per-week increase). This effect was stronger in males (F1,11447.7 = 9.83, p = 0.0017), such that each once-per-week increase in cannabis use was associated with a 0.005 mm reduction in cortical thickness, comparable to 17.9% of the annual rate of cortical thinning (-0.028 mm/year). The strongest within-person effects of cannabis were observed in regions with the greatest expression of CNR1, the gene that codes for the CB1 receptor (sample: rho = -0.33, pspin = .025; males: rho = -0.5, pspin = .005). At the between-person level, males (but not females) also exhibited a stable cortical thickness signature associated with propensity towards cannabis use and this signature was present before cannabis exposure. These results highlight the importance of longitudinal analyses using multi-level modelling to disaggregate potential risk factors from potential consequences of substance use.