Public ds006065 files showed theta phase locking in 1 of 4 closed-loop files, with no stronger locking than controls at atlas-nearest pairs

We tested whether released SEEG files from OpenNeuro ds006065 show theta-phase targeting of closed-loop hippocampal stimulation. We analyzed all nine stimulation files (four closed-loop, five phase-blind control) from seven people, using one 60-s segment per file. The files have no pulse markers, so we detected pulses from stimulation artifacts (19–57 analyzed pulses per file). For each file, we used the bipolar pair nearest to a hippocampal MNI point, blanked each pulse, filtered at 4–8 Hz, and measured Hilbert phase 10 ms before each pulse. We compared phase-locking value (PLV) with 2,000 random-time sets blanked in the same way. In the closed-loop files, PLV exceeded the null only in sub-p17 (PLV 0.42, null 0.14, Holm p = 0.004); it did not exceed the null in sub-p16, sub-p18, or sub-p19. Locking was not stronger in closed-loop than control files (mean z 0.74 vs 0.93; one-sided Mann–Whitney p = 0.72). Results at the second- and third-nearest pairs were also not significant (p = 0.72 and 0.21). Some control files also showed locking, and the strongest locking depended on the pair. The dataset does not identify the sensing contact, reference, or band used by the closed-loop system. Each file provides only 60 s, and the comparison includes four versus five sessions. Thus, this result does not show that phase targeting failed. It shows that these public files, at atlas-chosen pairs, do not provide clear evidence of the targeting.

stimulation · theta phase · stereo-EEG · phase locking · hippocampus · closed-loop stimulation · OpenNeuro ds006065