Public SEEG data show no reliable session-level increase in late evoked potentials after theta-synchronized stimulation (4 vs 5 sessions)

We tested whether theta-synchronized closed-loop stimulation increases the late (50–250 ms) single-pulse evoked potential more than control stimulation, as reported by Kragel et al. (2025). In public OpenNeuro ds006065 v1.0.0 data, we detected pulses from stimulation artifacts and compared pre- and post-stimulation responses using bipolar channels and session-level statistics. The late-window mean median log change was +0.076 in four theta-synchronized sessions and −0.064 in five control sessions (difference +0.140; exact permutation p = 0.31). The early-window difference was similar (+0.157; p = 0.43), unlike the reported late-only effect. One theta-synchronized session contributed most of the late effect; the paired subjects did not agree, and within-block drift was similar in size to most pre/post changes. The MNI-coordinate hippocampal subset gave a late difference of +0.123 (p = 0.43). This session-level test has low power and does not exclude the reported effect. Its session-level test differs from the paper’s channel-level mixed model. Hippocampal locations are approximate, and the analysis did not use the paper’s kurtosis rejection or exact channel selection.

stereo-EEG · evoked potential · closed-loop stimulation · theta synchronization · session-level analysis · OpenNeuro