After single-pulse SEEG stimulation, high-gamma changes at N1 contacts lasted up to 0.75 s in eight patients

Question: a bidirectional BCI that stimulates and reads high gamma must know how long one pulse changes the read signal. Data: OpenNeuro ds005448 v1.0.0, SEEG patients STREEF06-13, 6 random stimulated pairs per patient, all 10-20 pulses per pair, expert N1 labels. Method: byte-range epochs -0.6 to +1.2 s, common median reference, artifact interpolation (-5 to +12 ms), 70-150 Hz Hilbert power in dB against -0.5 to -0.1 s, trial-level signed-rank test per pair-contact combination, one-sided binomial test of the fraction of changed contacts against 5% per patient. Results (total power; 461 N1, 597 near and 3148 far non-N1 combinations): at N1 contacts the median fraction of changed contacts was 0.12-0.16 from 50 to 500 ms, 0.08 at 500-750 ms and 0.05 at 750-1100 ms. The last disturbed bin ended at a median of 0.2 s (maximum 0.75 s, 8 patients). Far non-N1 contacts stayed near chance (0.05-0.09). Patient mean changes were within 0.3 dB. The sign was not consistent, and changes at N1 contacts at 200-500 ms were mostly increases, so we found no consistent high-gamma suppression. Exception: -0.24 dB at far contacts at 50-100 ms in 8 of 8 patients (possible residual artifact). Induced power showed a late excess at far contacts that is probably a method artifact. Limits: 6 pairs per patient because of slow download, 10-20 trials, window ends at 1.1 s, one reference, low power per contact, 1-2 mA only, one center.

stereo-EEG · SEEG · single-pulse stimulation · high-gamma power · brain-computer interface · epilepsy · stimulation artifact · OpenNeuro