In four SEEG patients, left/right key information was not better aligned to key presses than to cues

The arrow cue and pressed key usually agree in OpenNeuro ds004473, so left/right decoding may reflect visual cue or motor information. We compared response-locked and cue-locked SEEG band-power decoding windows with the same mean latency in four of eight patients. Using bipolar signals, log high-gamma (70–150 Hz) and beta (13–30 Hz) envelopes, and shrinkage LDA, response-locked AUC did not exceed cue-locked AUC consistently: the median response-minus-cue difference was −0.046, with one of four patients positive (exact Wilcoxon p = 0.25). Patient 2 decoded above chance (permutation p = 0.01); patient 6 was at the limit (p = 0.05), while patients 1 and 5 were at chance. These data do not show a motor, key-press-locked origin for left/right information. Interpretation is limited by four patients, low response-time variance in two patients, only eight cue/key mismatch trials, and unknown hand or finger use. The reduction to four patients was not pre-specified and resulted from a 45-minute time limit.

SEEG · intracranial EEG · key decoding · cue alignment · response alignment · high gamma · beta power · OpenNeuro · linear discriminant analysis