A fixed contact number rule gave higher mean label concordance than 6 dB rise timing at 61 boundaries in 26 SEEG patients with two seizures each

We tested two seizures from each of 26 SEEG patients in public OpenNeuro ds004100 version 1.1.3. The fixed graph had 61 clinical seizure onset label boundaries on 54 parsed shafts. Each comparison used two pure-label bipolar signals with no shared contact. A fixed rule favored the smaller contact number. A signal rule favored the earlier sustained 6 dB fast-to-slow power ratio rise. Mean boundary concordance was 0.848 for contact number and 0.659 for timing. Timing minus number was -0.189, with a 95 percent patient bootstrap interval of -0.313 to -0.046. The labeled pair had the smaller number at 50 of the 61 boundaries. Timing reached 0.723 over all opposite-label pairs on the same shafts. We downloaded all 52 source ranges again and rebuilt all features with exact agreement. The package preserves code, labels, graph edges, source hashes, results, and retrieval instructions. Contact number does not measure tissue depth. Clinical labels can use these seizures. The study does not establish causal seizure origin or surgical benefit.

stereoelectroencephalography · seizure onset zone · epilepsy · spectral power · signal timing · contact number · label concordance · OpenNeuro