In the first 180 s of four closed-loop and five phase-blind ds006065 runs, hippocampal phase locking was not higher in closed-loop runs; strong locking occurred only with control-signal crosstalk

We examined whether control-channel events in the first 180 s of four closed-loop (cl) and five phase-blind control (clcontrol) runs from OpenNeuro ds006065 aligned with hippocampal 3–8 Hz phase. Events were sign changes in a plateau control channel. For each hippocampal bipolar pair, phase at each event was estimated from pre-event data using an autoregressive forecast and Hilbert phase, and compared with a sham-time null. Among pairs classified as clean by a control-channel correlation threshold, 4/34 closed-loop pairs and 7/72 control pairs exceeded the sham 95th percentile; about 5% was expected by chance. No closed-loop run had a clean-pair phase-locking value (PLV) above 0.19. Strong locking occurred only on pairs correlated with the control channel, including p19 D3-D4 (PLV 0.60, r = −0.73) and all hippocampal pairs in p18 (|r| 0.37–0.76). The audit therefore found no greater hippocampal phase locking in closed-loop than control runs; the strong locking was consistent with crosstalk. Median estimator error was 39–47 degrees, so small locking could be hidden. The analysis covers only the first 180 s per run, and the events were not confirmed as stimulation pulses.

intracranial EEG · hippocampus · phase locking · closed-loop stimulation · crosstalk · phase estimation · OpenNeuro · electrophysiology