Heartbeat-locked SEEG activity persists after ECG regression, but pulse artifact remains unresolved

In 31 of 37 epilepsy patients, 5-minute interictal bipolar SEEG recordings from OpenNeuro ds004100 showed heartbeat-locked activity at 200–600 ms after the ECG R peak. Across 3,099 adjacent-contact bipolar pairs, the median fraction with nominal HER p < 0.05 was 17.5% (29/31 patients above 5%; Holm p = 3.2×10⁻⁶), but a post hoc pseudo-beat control in 12 patients showed that the random-time null was liberal: 9.4% of pairs were significant under control timing. Real heartbeat timing exceeded the pseudo-beat control by about 6 percentage points (15.5% vs 9.4%; 9/12 patients; p = 0.020), and the residual HER fraction after regression on the ECG average and its derivative also exceeded control (13.8% vs 6.8%; 10/12; p = 0.0098). ECG regression did not reduce the full-sample fraction (16.9% vs 17.5%; p = 0.66). The HER–QRS correlation also appeared under control timing, and the bipolar QRS window showed no excess. The data therefore support a modest heartbeat-locked excess over the pseudo-beat control, but do not distinguish a neural response from a pulse artifact. The study used one ECG lead and one 5-minute clip per patient; the control covered 12 patients, and few pairs passed within-patient FDR (58 pairs in 4 patients).

heartbeat-evoked response · SEEG · cardiac field artifact · pseudo-beat control · epilepsy · electrophysiology