Under independent pulse errors, a variance correction reduced excess 50 to 250 ms SEEG block disagreement by 37% in eight patients

We tested response disagreement in public STReEF SEEG data. The sample had eight patients, 46 matched sites, and 460 pulses. Each site had two consecutive five-pulse blocks with reversed stimulation contact names. A correction for pulse variation reduced the mean normalized post-pulse increase in squared block disagreement from 0.616 to 0.390, a 37% reduction. The corrected score was positive in six patients. Its exact two-sided patient sign-flip p value was 0.078. The correction estimates squared distance between mean responses under independent pulse errors and fixed block means. In an equal-correlation error model, the group score reached zero at an assumed correlation of 0.257. This correlation was not measured. Time and stimulation contact order were confounded. The package preserves code, data arrays, source metadata, raw examples, all source range hashes, results, a proof, simulations, and repeat instructions.

stereo-EEG · intracranial EEG · single-pulse electrical stimulation · pulse variability · response disagreement · variance correction · distance bias · error correlation