Mean alpha recurrence AUC was 0.0044 lower for half support than peak height in paired SEEG files from 33 patients under CAR

We tested whether support in both 30 s halves ranks alpha peak recurrence better than full-minute peak height. We used the first 60 s in each paired SEEG file from OpenNeuro ds004100, version 1.1.3. The fixed rule used common average reference, a 3 dB peak threshold, and a 1 Hz recurrence tolerance. Of 36 patients, 33 had both recurrent and nonrecurrent candidates and could define an AUC. Their mean AUC was 0.6566 for the smaller support in the two halves and 0.6610 for full-minute height. The mean gain was -0.0044. Its 95% patient bootstrap interval was -0.0546 to 0.0521. These data did not show a gain. They did not establish equal performance. The interval assumes independent patients from a common population. All eight fixed descriptive changes had negative mean gains, with different patient coverage. The package preserves full and half spectra, scores, statistical draws, source hashes, code, and a raw-data retrieval recipe. Independent numerical checks and an offline repeat passed. The biological interval between files is unknown.

alpha peak recurrence · stereoelectroencephalography · SEEG · spectral peaks · Welch power · common average reference · paired files · area under the curve · patient bootstrap · signal processing