In OpenNeuro ds006065, theta-synchronized pulses increased late monopolar responses over phase-blind pulses inside and outside the hippocampal region, but not in bipolar signals

We tested again a claim by Kragel et al. 2025 (Nat Commun 16:4061) on OpenNeuro ds006065 v1.0.0 (CC0). The claim: lateral temporal pulses that are synchronized to hippocampal theta (TS) increase the late (50-250 ms) hippocampal evoked potential (EP) more than phase-blind pulses (PB). We wrote the plan before we calculated outcomes. We found the pulses from the artifact on a fixed 2.002 s grid. Hippocampal channels come from an MNI box (no anatomical labels) and must pass a sham-based responsiveness test. Per session, the outcome is the mean log(post/pre) of the EP peak-to-peak amplitude. The test is an exact permutation of the TS/PB labels over 9 sessions (126 splits). Results: monopolar hippocampal late EP, TS +0.38 vs PB +0.05, difference 0.34 (95% bootstrap CI 0.08 to 0.59), one-sided p = 0.040. Thus the primary criterion of replication is met. The early EP (15-50 ms) did not differ (p = 0.64), as in the paper. But responsive monopolar channels outside the hippocampus had almost the same difference (0.28, p = 0.032). With bipolar signals, the hippocampal difference was -0.06 (CI -0.29 to 0.15, p = 0.67). This agrees with an earlier bipolar re-analysis in this library (b1f3943a). Thus the TS-related increase is in a signal that is common to many monopolar channels (reference or widespread state). The data do not show a local hippocampal change. Limits: 9 sessions, two people in both conditions, MNI-box localization, unknown recording reference, few bipolar hippocampal channels (low power), and a measure that is different from the paper's mixed model. The package has the plan, numpy-only code, per-channel results, average EPs, figures, S3 keys with ETags, and checksums.

theta-synchronized stimulation · hippocampus · evoked potential · SEEG · monopolar reference · bipolar reference · OpenNeuro