Do Differences in Learning Performance Precede or Follow Initiation of Marijuana Use?

Studies examining cross-sectional associations between age at marijuana initiation and memory deficits yield mixed results. Because longitudinal data are sparse, controversy continues regarding whether these deficits reflect premorbid risk factors or sequelae of early marijuana initiation; here, we examine this question in a community sample followed since birth. Masked examiners administered four subtests of the Wide Range Assessment of Memory and Learning (WRAML/WRAML2) from childhood until young adulthood to 119 urban, predominantly African American participants. Multivariable generalized estimated equation models measured longitudinal trajectories of learning. Participants were grouped as never users (n = 26), later initiators (≥16 years old; n = 31), and earlier initiators of marijuana use (n = 62). Marijuana onset groups did not significantly differ on WRAML scaled scores or IQ in childhood, nor did they differ on WRAML scaled scores in adolescence. On most WRAML2 subtests, these groups did not significantly differ in young adulthood after taking into account sex and childhood IQ. However, on Story Memory, later initiators attained higher scaled scores in young adulthood, even after including additional covariates of anxiety, depression, postsecondary education, past-month marijuana use, and past-week high-risk drinking. They showed a significantly more positive trajectory than never users that was driven by within-group improvement after adolescence. Earlier initiators showed within-group decline in Story Memory after adolescence. Differences in learning following earlier initiation of marijuana use may not be solely attributable to premorbid deficits.