Abstracts

Rationale: Whether sphenoidal electrodes (SpE) provide clinically useful information above that provided by true temporal scalp electrodes (T1/T2, here called TE) in the localization of temporal ictal onsets has not been clearly established. The few studies published to date have yielded conflicting results. Methods: We reviewed up to 4 ictal recordings per patient in a consecutive series of telemetry patients who had SpE and TE placed for possible temporal lobe epilepsy. All tracings (117 seizures in 36 patients) were digitally reviewed by two primary readers (SM, JN) who were unblinded to patient identity but blind to the other's reading. Each seizure was reviewed by each reader twice, once in a montage using SpE and once using TE (T1/T2), and scored as either meeting (SpE+, TE+) or not meeting (SpE -, TE -) mesial temporal onset criteria (5Hz or faster unilateral SpE or TE discharge within 30 seconds of electrographic onset). Any seizure scored as SpE+/TE- by only one primary reader was evaluated by a third reader (JE) for consensus. Results: Reader #1 found 36 seizures were SpE+ and 46 were TE+. Reader #2 found 34 seizures SpE+ and 45 seizures TE+. Interreader agreement was 83% for SpE (?=0.59) and 77% for TE (?=0.51). Four seizures in 3 patients were felt by consensus (at least 2 of 3 readers) to be SpE+ and TE-. In 3 of these, TE were not recording properly prior to seizure onset. With regard to the fourth, both primary readers felt that other similar seizures in that patient were TE+, and thus were as localizing as the SpE+/TE- seizure. Six seizures in 5 patients were felt by both primary readers to be TE+ and SpE- (these were not reviewed by a third reader). Such occurrences were due to SpE not recording properly prior to seizure onset (in 5 of 6) or blocking from muscle activity during the event. The overall rates of electrode dysfunction were similar for SpE (9/117) and TE (7/117) . Discussion: SpE and TE are similar in their ability to detect seizures which meet mesial temporal onset criteria. In the presence of well-functioning TE, SpE provided additional localizing information in less than 1% of seizures in this series, and no additional clinically useful data was obtained.