Performance on the Twenty Questions task has been assumed to be deficient in those with frontal lobe damage, although few studies have specifically examined this. In the present study, 88 participants with frontal lobe epileptic dysfunction were assessed on this task, as were 57 participants with temporal lobe epileptic disturbance and 28 participants with no evidence of any neurological damage. Results indicated that those with either left or bifrontal lobe damage were more impaired on several indexes of test performance. Analysis also revealed that participants with orbitofrontal damage were impaired on the measure of 'first guess' as compared to groups with damage to other regions of the prefrontal cortex (p < .05). Overall, the results support the utility of the Twenty Questions task in assessing frontal lobe damage and may suggest that qualitatively different deficits may be recorded depending on the region of frontal lobe dysfunction.