Objective: Despite widespread drug testing in sports and warnings about the potential risks of using anabolic-androgenic steroids (AAS), non-medical use is prevalent among athletes, non-athletes, and disturbingly among adolescents. To date, most research has focused on the anabolic properties and short-term health risks of AAS use. In contrast, studies investigating the effects on cognitive function in men using high doses of multiple exogenous steroids are lacking. The primary purpose of this naturalistic study was to examine the effects of non-medical steroid use on sex-related cognitive abilities in male bodybuilders. The secondary purpose of the study was to evaluate the psychological functioning of male bodybuilders who use AASs. Methods: Eight male bodybuilders who used high doses of AASs were matched with bodybuilding and aerobic controls who had never used AASs, according to age, education, and estimated verbal intelligence. AAS use of the bodybuilders appeared similar to reports in the literature of self-administered AASs regimens used by strength athletes. All groups underwent a battery of cognitive tests and self-report psychological inventories, and had serum total testosterone and binding proteins measured immediately after testing. Cognitive measures selected were those that have previously shown sex differences. The study examined four psychological domains: aggression, personality, body image, and eating-disordered attitudes/behaviours. Results: Male bodybuilders who used AASs scored significantly lower than controls on mental rotations and on the WAIS-III Digit-Symbol Coding subtest. There were no other significant group differences on the cognitive tasks. A curvilinear (inverted U) relationship was identified between spatial ability and total testosterone in men who did not use AASs. As there were only a few AAS users in the current study, there was little power to demonstrate a linear or nonlinear relationship. Overall, there were no significant differences between groups on the psychological variables. AAS users exhibited elevated levels of antisocial personality traits, with 38% scoring in the clinically significant range. Bodybuilders reported some body weight concerns, specifically a drive for muscularity combined with a drive for a well-toned body, with no difference between AAS users and bodybuilding controls. Three AAS users and one bodybuilding control exhibited psychological disturbances, as evidenced by elevated scores on multiple psychological measures. Conclusions: The results of this preliminary study provide some evidence that high doses of AASs in men might influence certain aspects of cognition, specifically reducing complex visuospatial skills and perceptual speed. The data also suggests that endogenous testosterone influences spatial ability in healthy men in a curvilinear fashion. Further research with larger samples of AAS users is required to quantify the cognitive effects of non-medical AAS regimens. The study also contributes to the growing literature on the psychological effects of bodybuilding and AAS use. Although many AAS users and bodybuilders might display minimal psychopathology, there is likely a subgroup of individuals who exhibit clinically significant psychological disturbances. Further research is necessary to identity the nature and severity of psychological symptomatology in this population, and effective modes of treatment.