OBJECTIVE The study seeks to evaluate impairments of implicit learning and executive function in chronic polydrug abusers. It was hypothesized that implicit learning and executive function correlate with anhedonia. METHODS A cross-sectional group comparison. METHODS Department of Psychiatry, University of Tübingen, Germany. Participants A total of 25 male polydrug abusers with opiate dependence, n = 26 polydrug abusers abstinent for more than 3 months and n = 26 non-drug-using healthy males. METHODS Abstinent polydrug abusers were recruited from a community treatment centre, current polydrug abusers from local drug counselling services and controls through advertisements. METHODS A psychological battery assessing implicit learning (serial reaction-time task), various executive functions (latent inhibition, delayed matching-to-sample, Trail Making Test, acquisition and modification of conditioned responses, figural reasoning) and verbal logic memory was administered. Hedonic thoughts and activities as well as depressive symptoms were assessed through questionnaires. RESULTS In chronic polydrug abusers, there were moderate impairments of implicit learning, of acquisition, reversal and extinction of conditioned responses, of latent inhibition as well as anhedonia, while working memory was spared. In the abstinent group, cognitive performance was normal except for latent inhibition and more anhedonia and depression than in controls. CONCLUSIONS The findings suggest that current polydrug abusers suffer from impairment of many cognitive functions and from anhedonia. During abstinence, there is near normal cognitive function but still anhedonia. Anhedonia was correlated with implicit learning but not with executive function.