Abstract Background: Functional brain imaging studies of major depression have consistently revealed hypometabolism or hypoperfusion in specific regions of the prefrontal cortex and basal ganglia. Studies of cognitive functioning in major depression have suggested that some but not all subjects exhibit cognitive deficits that are consistent with frontal-subcortical dysfunction, although the reasons for this heterogeneity are unclear. In this study, we explored this heterogeneity among depressed subjects by examining the relationship between cognitive functioning and treatment outcome. Method: Subjects with major depression were administered a complete neuropsychological test battery prior to treatment with fluoxetine. Results: There were no significant differences between responders and nonresponders to fluoxetine in terms of age, educational achievement, number of past episodes of depression, and estimated premorbid IQ. However, nonresponders performed significantly worse than responders on several pretreatment measures of executive functioning, after controlling for baseline group differences in depression severity. Limitations: The results are based on a small sample of primarily female subjects, resulting in low statistical power and less generalizability to samples of male subjects with depression. Conclusions: The findings suggest that subtle prefrontal dysfunction in subjects with major depression may be predictive of poor response with particular medications. Assessment of the executive functions may play a particular role in pretreatment identification of subjects likely to respond to specific medications.