OBJECTIVE: Experimental data indicate inhibitory effects of the cerebellum on seizure activity. Structural damage such as cerebellar atrophy, which is a common finding in patients with chronic epilepsy, may reduce these effects. METHODS: Outcome after temporal lobectomy was studied in 78 consecutive patients, with or without cerebellar atrophy diagnosed by MRI. RESULTS: Thirty five patients (45%) showed cerebellar atrophy. At a mean follow up of 14.6 (range, 6-40) months, 50 patients (64%) had no postoperative seizures. In these patients, the frequency of cerebellar atrophy was significantly lower (34%) than in patients who relapsed (64%, p < 0.01). Occurrence of generalised tonic-clonic seizures (GTCS) within two years before surgery, occurrence of GTCS at any time preoperatively, long duration of epilepsy, and older age at surgery were also associated with recurrence of seizures. Multiple logistic regression analysis suggested occurrence of GTCS within two years before surgery and cerebellar atrophy as the main predictive indicators. When both factors were present, the percentage of patients remaining seizure free since surgery fell to 30%, compared with 60% when only GTCS were present, 78.6% when only cerebellar atrophy was present, and 87.5% when both factors were absent. CONCLUSIONS: Cerebellar atrophy shown by MRI was a frequent finding in surgically treated patients with temporal lobe epilepsy. The presence of cerebellar atrophy seems to worsen the prognosis after temporal lobe resection.