Occipitotemporal epilepsies: evaluation of selected patients requiring depth electrodes studies and rationale for surgical approaches.

In 8 patients in whom it was uncertain whether they had occipital or temporal lobe (TL) epilepsy, clinical, scalp EEG, and radiologic features were correlated with the sites of seizure onset as determined by depth EEG. The 8 patients were selected from > 40 with occipital epilepsy because they had (a) an aura considered to be of occipital lobe (OL) origin, (b) an occipital interictal epileptic focus, (c) an OL lesion, or (d) a combination of all of these. Scalp EEG and clinical patterns suggested temporal involvement in all, however. Extracranial EEG recordings were often misleading, showing multilobar interictal epileptic abnormalities, and seizure onset was of poor localizing value and did not clarify the problem sufficiently. Intracranial EEG recordings showed that seizure onset could be ordered along an occipitotemporal gradient. Consistent OL seizure onset was observed in patients who had only elementary visual auras. Those who had inconsistent aura or no aura, suggesting OL origin, had onset of most attacks in the TL. All patients had a seizure spread pattern suggesting early TL involvement. To prevent visual field defect, surgical approaches included temporal resection when temporal seizure origin or spread was demonstrated; although occasionally this produced excellent results, it was of limited benefit in most patients, even when some seizures were proven to originate in TL structures. In patients with malignant epilepsy and in those with an occipital lesion, occipital resection should be considered.