With fixed source selection, mean 80 to 150 Hz peak overlap was higher at seizure onset pairs in 21 SEEG patients

We tested whether relative power peaks overlap more at seizure onset pairs on matched electrode-name groups. We used fixed public SEEG features from OpenNeuro ds004100 version 1.1.3. Twenty-one patients supplied two 60-second clips each. Each signal had six peak windows. Mean 80 to 150 Hz peak overlap was 22.40 percent at seizure onset pairs and 16.12 percent at other pairs. The difference was 6.28 percentage points, with a 95 percent patient bootstrap interval of 1.73 to 11.66 points. The exact patient sign-reversal p value was 0.016968 under independent patients and symmetric null contrasts. Fourteen patients had positive contrasts and seven had negative contrasts. Beta peaks gave a larger difference of 11.27 points. Thus the result does not establish a high-frequency-specific marker or clinical prediction value. The result depends on a fixed source pair set with an unresolved coordinate-unit rule. The package preserves all input features, source versions, retrieval recipes, peak indices, patient scores, control draws, checked code, and figures.

stereoelectroencephalography · epilepsy · power peaks · peak overlap · intracranial EEG · matched comparison · signal processing