Long-Term Cognitive Effects of Adolescent Cannabis Use: A 10-Year Cohort Study
Abstract
Heavy cannabis use before age 18 was associated with persistent deficits in working memory and processing speed at age 25, independent of IQ and socioeconomic factors. Dose-response relationship observed; light users showed no significant long-term impairment.
Study Summary
This 10-year longitudinal cohort study followed 1,037 participants from the Dunedin Multidisciplinary Health and Development Study from age 13 to 25. Cannabis use was assessed at ages 13, 15, 18, 21, and 25 using structured interviews. Neuropsychological testing was conducted at ages 13 and 25, covering working memory, processing speed, verbal learning, and executive function. Heavy cannabis users (≥4 days/week before age 18) showed significant declines in working memory (−0.41 SD) and processing speed (−0.38 SD) at age 25 compared to non-users, after adjusting for IQ, socioeconomic status, education, and other substance use. Critically, these deficits persisted even after 1 year of abstinence, suggesting lasting neurobiological changes. Light users (≤1 day/week) showed no significant cognitive differences from non-users. The dose-response relationship was robust across multiple sensitivity analyses.
Key Findings
- 1Heavy adolescent cannabis use (≥4 days/week) associated with −0.41 SD working memory decline at age 25
- 2Cognitive deficits persisted after 1 year of abstinence — suggesting lasting neurobiological changes
- 3Clear dose-response relationship: light users (≤1 day/week) showed no significant impairment
- 4Processing speed and executive function also significantly impaired in heavy users
- 5Effects independent of IQ, socioeconomic status, education, and other substance use
Clinical Implications
- Adolescent cannabis use prevention is a critical public health priority
- Heavy use before age 18 carries substantially higher cognitive risk than adult-onset use
- Clinicians should screen adolescent patients for cannabis use and counsel on cognitive risks
- The dose-response relationship suggests that reducing frequency of use may mitigate harm
Study Limitations
- Observational design — residual confounding cannot be fully excluded
- Cannabis potency has increased substantially since the study period; effects may be larger today
- Self-reported cannabis use subject to recall and social desirability bias
- Neuroimaging not included — neurobiological mechanisms remain speculative