Wavelet-crosscorrelation analysis can help predict whether bursts of pulse stimulation will terminate afterdischarges

Objective: Extraoperative cortical localizing stimulation (LS) is a standard clinical tool used to assess brain function before epilepsy surgery. However, LS can produce unwanted afterdischarges (ADs). We previously have shown that brief pulses of electrical stimulation (BPS) can terminate ADs caused by cortical stimulation. Our objective was to assess whether wavelet-crosscorrelation analysis could help predict the conditions under which BPS would be most likely to terminate ADs. Methods: We used wavelet-crosscorrelation analysis to get wavelet-correlation coefficients (WCC), and determine time lag (TL) and absolute value of TL (ATL) between two electrodes. For Analysis-1, we compared WCC and ATL in epoch 1 which was before LS, epoch 2 which was after LS but before BPS, and epoch 3 which was after BPS. For Analysis 2, we compared WCC and ATL during epoch 1 under 4 conditions: epochs when ADs subsequently terminated within 2 s after the end of BPS (1A), terminated within 2-5 s (1B), did not terminate within 5 s (1C), and when ADs did not appear (1D). Results: We found that BPS efficacy in terminating ADs was predicted by (1) low correlation and (2) slow propagation speed between electrode pairs in the 2-10 s period before stimulation. Conclusions: Wavelet-crosscorrelation analysis can help predict conditions during which BPS can abort ADs. It is possible that similar analyses could help predict when BPS or other interventions could abort clinical seizures.