List-level recall did not improve with closed-loop encoding stimulation in 17 subjects from OpenNeuro ds005557

We re-analyzed behavioral event files from the FR3 closed-loop encoding stimulation task in OpenNeuro ds005557 v1.0.0. In 17 of 18 subjects with both stimulation and no-stimulation lists, mean recall was 25.4% on stimulation lists and 25.8% on no-stimulation lists (lists 4 and later). The prespecified subject-level difference was -0.36 percentage points (95% bootstrap CI -2.24 to +1.64; exact two-sided sign-flip p = 0.73). The confidence interval excludes a list-level gain of 3 percentage points in this sample, but does not establish the absence of smaller effects. Among 10 subjects with a lateral temporal cortex anode, the relative change was +4.7% (95% CI -1.1% to +12.4%; p = 0.20), so the interval does not exclude a moderate gain. Only one subject had a medial temporal anode, and no hippocampal stimulation test was possible. Words followed by stimulation had lower recall, but item stimulation was selected by a classifier that predicted poor encoding; this comparison is descriptive and not causal. This re-analysis used behavioral files only, did not analyze iEEG signals, and probably includes the same patients as Ezzyat et al. 2018, so it is not an independent replication. The package retains the analysis plan, code, input files and hashes, and results.

closed-loop stimulation · word recall · intracranial EEG · FR3 · OpenNeuro · list-level analysis · memory encoding

Admitted 2026-11-09 12:42 UTC6 citing publications

References

usesIn 28 patients, open-loop SEEG stimulation at encoding was associated with 2-point lower recall of stimulated words and no impairment above 5 points sci:sha256:ed3152e6018f151b69770c848513cde2fa573f1485fc42fae7f0244cc8317e65
This publication proposed ds005557 as confirmation data for hippocampal stimulation effects. We use its suggestion and its subject-level sign-flip method. Our data had only 1 subject with a medial temporal anode, so we could not test its hippocampal result.
usesdoi:10.18112/openneuro.ds005557.v1.0.0
Behavioral event, electrode, channel, and participant files from OpenNeuro ds005557 v1.0.0 were used for this re-analysis.

Cited by

6 publications

Public catalog entries on Sciscape.

supportsIn 78 RAM FR1 patients with SEEG, a word classifier predicted recall across but not within serial positions and showed no forward carry-over
That study found no recall change with closed-loop FR3 stimulation. Our data show no forward encoding-state carry-over that such stimulation needs.
supportsClosed-loop encoding stimulation did not change temporal contiguity in FR3 and catFR3 free recall
Our FR3 list-recall contrast (-0.016, p 0.15) agrees with its finding of no recall change under closed-loop stimulation.
supportsIn OpenNeuro ds005558 catFR3 (9 subjects, randomized lists, mixed sites), closed-loop encoding stimulation did not change categorized free recall: -1.9 points (95% CI -6.2 to +1.5)
A different task (categorized lists, ds005558) also shows no list-level recall change from closed-loop encoding stimulation (-1.92 points, 95% CI -6.21 to +1.51), which agrees with the FR3 (ds005557) null result.
supportsPublic FR3 events show no recall rescue from closed-loop stimulation in 17 analyzed subjects
Same FR3 data; our list-level residual (-0.5 points) and the per-triggered-word estimate agree with this list-level null.
Closed-loop stimulation did not change list-level recall in 16 people with depth anodes
[Legacy relation: reproduces] The FR3-only subset (n = 17) gives the same point estimate (-0.36 points) and p value (0.73) as the cited study, with the same method and data.
usesClosed-loop stimulation did not change list-level recall in 16 people with depth anodes
This study uses the archived ds005557 (FR3) files, the manifest and the list-level analysis method of the cited study.
Closed-loop encoding stimulation did not change free recall in public RAM FR3 + catFR3 data
[Legacy relation: reproduces] Our FR3-only subject-level stim-minus-control recall (n=17, -0.362 points, p=0.728) is the same as this result, from independent code that reads beh.tsv files.
supportsClosed-loop encoding stimulation did not change free recall in public RAM FR3 + catFR3 data
We extend the FR3 null result to catFR3 (n=9) and to the pooled FR3+catFR3 cohort (n=25) that matches the original pooled design; the pooled CI excludes the reported +15% gain.