In two SEEG patients, a band-power LDA separated key presses from rest but also detected fixation before the go cue

In two patients performing a two-choice key task, a shrinkage linear discriminant analysis (LDA) using bipolar stereo-electroencephalography (SEEG) band power separated bins around a key press from inter-trial rest (high-gamma AUC 0.95 and 0.92; 20 trials per patient). The high-gamma detector also distinguished the pre-movement fixation period from rest (AUC 0.79 and 0.63), and press from fixation (AUC 0.87 in both patients). With a threshold set at the 95th percentile of training-rest scores, 55% and 50% of high-gamma detections occurred before the go cue. Held-out rest false-positive rates were 2.2% and 5.5% for high gamma, and 11.9% and 17.1% for high gamma plus beta. About eight selected features gave AUCs similar to all features. These results indicate that a rest-versus-press detector can respond to task engagement or preparation as well as movement. The analysis covers two of eight patients, 20 trials per patient, approximately 20 seconds of rest per patient, and linear band-power features; it does not establish performance on continuous recordings or across a patient population.

stereo-electroencephalography · SEEG · brain-computer interface · movement onset · high gamma · beta power · linear discriminant analysis · cross-validation · task engagement · OpenNeuro ds004473