A fixed spectral line model gave a mean later beta residual change of -0.115 dB at selected SEEG pairs in eight patients

We tested whether beta loss remained after a fixed spectral line adjustment. We used public SEEG data from OpenNeuro ds004473, version 1.0.2. Earlier work selected 87 of 822 adjacent bipolar pairs from eight patients. We kept those sites and used 965 trials from the next instruction block. The model fit log power at 8 to 12 Hz and 35 to 45 Hz. It predicted the mean log power at 13 to 30 Hz. The raw mean log beta change was -0.681 dB. The line prediction change was -0.566 dB. The response minus baseline residual was -0.115 dB at selected pairs. Its 95 percent patient t interval was [-0.563, 0.332] dB. The selected minus all-pair residual change was -0.077 dB, with interval [-0.416, 0.261] dB. Two exact patient sign tests gave Holm p values of 1.0000 and 1.0000. We checked all used source byte ranges. All new integrated beta powers matched the earlier powers. The package preserves code, trial features, source hashes, raw examples, results, and repeat instructions. This is a follow-up on a known cohort. The line is a fixed spectral control. It does not establish a true aperiodic component or prove an oscillation.

stereoelectroencephalography · beta power · spectral control · movement task · intracranial EEG · patient-level analysis