A fixed spectral adjustment changed the sign of theta power change estimates in 2 of 8 valid SEEG rest sessions

We tested whether resting hippocampal theta power changes kept their sign after a fixed spectral adjustment. We used public SEEG records from OpenNeuro ds006065 version 1.0.0. The primary model used 1.5–3 Hz and 13–30 Hz to predict the background in 4–10 Hz. Eight of nine planned sessions passed the quality rule. These sessions came from six patients. Two point estimates changed sign after adjustment. Both sessions had wide sampling intervals that included zero. The mean raw change per patient was +0.857 dB. The mean adjusted change was −0.121 dB. Alternate models changed the sign of this mean. The result shows measurement sensitivity in these short rest records. It does not establish a stimulation effect or isolate true oscillation power. The package contains source code, 18 signal excerpts, source versions and hashes, all planned checks, figures, and repeat instructions. All 18 source ranges were checked again. Reconstructed signal arrays matched the archive exactly. A second analysis run reproduced the numerical results.

theta power · spectral adjustment · SEEG · stimulation · power spectrum · sensitivity analysis