In determining the effects of regular marihuana use on neurocognition, abilities within specific relevant cognitive domains prior to regular drug use have not been available. The present study examined effects of current and past regular use of marihuana in subjects for whom pre-drug performance had been ascertained in a prospective, longitudinal fashion. A total of 113 young adults, assessed since infancy, were evaluated using neurocognitive tests for which commensurate measures were obtained prior to the initiation of marihuana smoking. Marihuana users, determined by urinalysis and self-report, were categorized as light (< 5 joints per week) and heavy (> or = 5 joints per week) current users and former users, the latter having used the drug regularly in the past (> or = 1 joint per week) but not for at least 3 months. A third of the subjects were using marihuana on a regular basis at the time of assessment with half being heavy users. Among former, regular users, approximately half had been smoking 5 or more joints per week. Overall IQ, memory, processing speed, vocabulary, attention, and abstract reasoning were assessed. After accounting for potentially confounding factors and pre-drug performance in the appropriate cognitive domain, current regular heavy users did significantly worse than non-users in overall IQ, processing speed, immediate, and delayed memory. In contrast, the former marihuana smokers did not show any cognitive impairments. It was concluded that residual marihuana effects are evident beyond the acute intoxication period in current heavy users after taking into account pre-drug performance but similar deficits are no longer apparent 3 months after cessation of regular use, even among former heavy using young adults.