Extensive evidence indicates that current and recently abstinent cocaine abusers compared to drug-naïve controls have decreased grey matter in regions such as the anterior cingulate, lateral prefrontal and insular cortex. Relatively little is known, however, about the persistence of these deficits in long-term abstinence despite the implications this has for recovery and relapse. Optimized voxel based morphometry was used to assess how local grey matter volume varies with years of drug use and length of abstinence in a cross-sectional study of cocaine users with various durations of abstinence (1–102 weeks) and years of use (0.3–24 years). Lower grey matter volume associated with years of use was observed for several regions including anterior cingulate, inferior frontal gyrus and insular cortex. Conversely, higher grey matter volumes associated with abstinence duration were seen in non-overlapping regions that included the anterior and posterior cingulate, insular, right ventral and left dorsal prefrontal cortex. Grey matter volumes in cocaine dependent individuals crossed those of drug-naïve controls after 35 weeks of abstinence, with greater than normal volumes in users with longer abstinence. The brains of abstinent users are characterized by regional grey matter volumes, which on average, exceed drug-naïve volumes in those users who have maintained abstinence for more than 35 weeks. The asymmetry between the regions showing alterations with extended years of use and prolonged abstinence suggest that recovery involves distinct neurobiological processes rather than being a reversal of disease-related changes. Specifically, the results suggest that regions critical to behavioral control may be important to prolonged, successful, abstinence.