A wealth of clinical data indirectly implicate dysfunction of frontal cortex in schizophrenia, including negative symptoms, the pattern of neuropsychological deficits, and abnormal eye movements. Neuroimaging studies have provided direct evidence of frontal, particularly prefrontal, malfunction, but the results have been inconsistent and controversial. The burning question is whether prefrontal hypofunction is a pathophysiological characteristic of schizophrenia per se or an artifact of the imaging protocol. In studies of patients at rest, 'hypofrontality' has been an inconsistent finding, probably because resting is physiologically and psychologically variable. Cognitive activation paradigms, especially during working memory tasks, have been reliable in showing prefrontal hypofunction in patients, but these results have been challenged as artifacts of poor performance. Performance differences have been addressed by studying patients and controls matched either for poor performance or for normal performance. The former approach, which has the potential of elucidating the specificity of physiological mechanisms associated with poor performance, has shown that prefrontal activity in patients with schizophrenia differs quantitatively and qualitatively from that of normals and of other patient populations who perform at a comparable level. The latter approach, which tends not to find prefrontal differences between patients and controls, may be selecting out important aspects of the disease by focusing on unaffected neural functions. While there are pitfalls to each approach and no single study can address all the potential phenomenological confounds, overall, the functional neuroimaging database in patients with schizophrenia suggests that prefrontal cognitive deficits are because of prefrontal pathophysiology and not the inverse.