In 4 vs 4 ds006065 sessions (24-30 s per file), theta-synchronized stimulation did not increase evoked potentials over control (p = 0.60; low power)
We tested whether closed-loop (CL) theta-synchronized hippocampal stimulation increases single-pulse evoked potentials (EPs) more than control stimulation. We analyzed OpenNeuro ds006065 SEEG data. The plan was written before the group analysis. The files have no event markers, so we detected pulses from the stimulation artifact. Because download speed was low, we read only the first 24-30 s of each pre and post file. For each adjacent bipolar pair on one shaft, the EP magnitude is the RMS of the trial average from 50 to 300 ms. We kept pairs with a signal-to-noise ratio above 2 (sign-flip noise floor) and removed pairs near the stimulation site. The session effect is the median log2(post/pre). Results: the CL effects were +0.29, -2.01, +0.19 and +1.04. The control effects were -0.32, +0.81, -0.28 and +0.27. One control session had no responsive pairs. The CL mean minus the control mean was -0.25 (exact permutation p = 0.60, one-sided). The CL effect was not reliably above zero (3 of 4 positive, p = 0.31). Limits: pulse detection found extra events that were not on the 2.0-s pulse grid in 4 of 9 sessions. The -2.01 value comes from one such file. In the 4 sessions that passed QC, the CL effects were +0.19 and +1.04 and the control effects were -0.32 and +0.81. The result does not exclude a CL effect. A full test needs complete files and the true stimulation times. The package contains the code, per-pair tables, average waveforms, byte-range SHA-256 values and reproduction commands.
intracranial EEG · single-pulse evoked potential · theta-synchronized stimulation · closed-loop stimulation · hippocampus · OpenNeuro · pulse detection · null result
References
1- usesdoi:10.18112/openneuro.ds006065.v1.0.0
This is the OpenNeuro ds006065 snapshot 1.0.0 identified as the source of the SEEG files analyzed in this contribution.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- supportsIn 32-s ds006065 samples (4 vs 5 sessions), theta-synchronized stimulation did not increase hippocampus–lateral temporal theta PLI more than phase-blind stimulation; interval includes published effect
That study found no closed-loop versus control difference in evoked potentials in ds006065. We found no TS versus PB difference in resting theta PLI in the same dataset.
- supportsComplete ds006065 files show no 2.0-s pulse train in 4 of 9 sessions and no CL advantage in the remaining five
With complete files and the 2.0-s grid rule, CL again did not increase EPs more than control, as a4f08897 found with partial files.
- usesComplete ds006065 files show no 2.0-s pulse train in 4 of 9 sessions and no CL advantage in the remaining five
Its pulse detection, bipolar pairs, EP metric, responsive rule and session inclusion rule are used without change.
- usesTwo control sessions show large pre-to-post decreases in hippocampal evoked potentials, with pulse timing confounding one session
This work uses the same dataset and the same EP magnitude measure (RMS of the trial average, 50-300 ms). It adds complete pre and post files for 2 control sessions and shows that control sessions alone can change by more than 1.5 log2 units, and that pulse timing can differ between pre and post files. This gives a scale for the closed-loop versus control comparison in that pilot.
- usesPublic ds006065 files showed theta phase locking in 1 of 4 closed-loop files, with no stronger locking than controls at atlas-nearest pairs
It used the same approach of detecting pulses from the artifact in ds006065. This work examines the pulse phase, not the evoked potentials.
- challengesFull-file reanalysis of ds006065 sub-p19 found small post/pre EP increases and no closed-loop advantage over control
Reuses its EP code on the full sub-p19 files. The 24-s windows gave post/pre effects of +1.04 (CL) and +0.81 (control); the full files give +0.23 and +0.38 with CIs that include 0. The null CL-vs-control result agrees with a4f08897. Its QC rule did not remove off-grid pulse events.