Sleep architecture, cocaine and visual learning.

Background: Disturbances in sleep associated with chronic cocaine use may underlie abstinence-related cognitive dysfunction. We hypothesized that sleep-related cognitive function would be impaired in chronic cocaine users, and that this impairment would be associated with abstinence-related changes in sleep architecture.

Methods: Twelve chronic cocaine users completed a 23-day in-patient study that included randomized, placebo-controlled, cocaine self-administration sessions. We report polysomnographic measurement of rapid eye-movement (REM) sleep and slow-wave activity, and performance on a visual texture discrimination task.

Findings: Progressive abstinence from cocaine was associated with characteristic changes in REM sleep. REM sleep was shortest on nights following cocaine use and rebounded in the first week of abstinence before diminishing with progressive abstinence, following a pattern opposite that of slow-wave activity. Overnight visual learning was observed over the first night following 3 consecutive days of laboratory cocaine use; however, learning was not observed at 3 days or 17 days of abstinence. Across all points of abstinence, early-night slow-wave activity was associated strongly with non-deterioration of visual performance overnight. Furthermore, overnight enhancement of visual performance was predicted by the co-occurrence of sufficient early-night slow-wave activity and late night REM sleep, similar to results from studies in healthy subjects.

Conclusions: These results suggest that abstinence-associated sleep-dependent learning deficits are related to characteristic changes in sleep architecture, and promote the idea that treatments directed at sleep ('somno-tropic' treatments) could be helpful in offsetting physiological consequences of cocaine abstinence.