In 42 analyzed trains, the polarity-invariant SEEG response decreased more than the polarity-flipping component (17 patient units; RMS ratio 0.71, p = 0.036)

Alternating-polarity stimulation separates pulse-locked stereo-electroencephalography (SEEG) into a polarity-invariant component S = mean(pos) + mean(neg), assumed to be neural, and a polarity-flipping component D = mean(pos) - mean(neg), assumed to be artifact. We tested whether S adapted during a stimulation train more than D. We analyzed byte ranges from OpenNeuro ds005169 (visual effects) and ds006519 (negative motor responses), using 128 alternating-polarity trains. Automatic pulse detection accepted 47 trains; 42 had at least one responsive channel and contributed to the analysis. Across 17 dataset-by-subject units (323 responsive of 3,494 valid channels), the mean contrast LR_S - LR_D was -0.348 (95% CI -0.645 to -0.072; 12/17 negative; exact two-sided sign-flip p = 0.036), equivalent to an S-to-D relative RMS ratio of 0.71. S decreased (mean LR_S -0.501; 95% CI -0.795 to -0.268; 15/17 negative; p = 0.00014; last-to-first RMS ratio 0.61). D also decreased, but its change was not significant (mean LR_D -0.202; 95% CI -0.430 to 0.031; 13/17 negative; p = 0.11; RMS ratio 0.82). Across 23 trains with at least five responsive channels, channel changes in S and D were not correlated (median Spearman rho 0.004; 52% positive). Under the plan's decision rule, these results support an adapting polarity-invariant component, consistent with a neural response rather than a residual artifact. Evidence is moderate: the primary test had p = 0.036 with 17 units. The simple detector rejected 81 of 128 trains, so the analyzed sample may be biased. The datasets come from one center, and some people may appear in both. Adaptation, background-state change, and onset of afterdischarges are not distinguished. The analysis did not assess distance, anatomy, intensity, or clinical effect; its fixed 2–15 ms window excludes earlier components. The recordings were read from the OpenNeuro S3 mirror as it existed on 2026-10-10; the exact correspondence to a versioned dataset snapshot is not established.

stereo-electroencephalography · intracranial EEG · alternating-polarity stimulation · stimulation artifact · neural response · adaptation · signal processing · sign-flip test