Object alternation (OA) is a well-established measure of perseveration and orbitofrontal function in non-human primates. Although several studies have used OA to examine orbitofrontal system dysfunction in humans with neurological and psychiatric disease, this task itself has not been validated as a bona fide measure of frontal dysfunction in humans. To address this issue, six patients with bilateral frontal lobe lesions documented by computerized tomography (CT) or magnetic resonance imaging (MRI) and 15 healthy controls were given OA, as well as other measures of frontal system dysfunction, delayed alternation (DA), delayed response (DR), and the Wisconsin Card Sorting Test (WCST). The CT and MRI scans were interpreted blindly. The patients with bilateral frontal lesions were significantly impaired on OA, DA, DR and the WCST. Analyses of the CT/MRI lesions suggest that the neuroanatomical regions involved in the deficits on OA include Brodmann areas 10, 24, 32 and 47, as well as possibly 11, and that OA is a sensitive measure of ventrolateral-orbitofrontal and medial frontal dysfunction in humans. Our findings lend further support for the use of experimental paradigms adopted from animal models to study the functional neuroanatomy and neuropsychological mechanisms underlying cognitive functions in humans with neurological and psychiatric disease.