Closed-loop encoding stimulation did not change free recall in public RAM FR3 + catFR3 data

Ezzyat et al. (2018) reported about a 15% relative recall increase on lists with closed-loop encoding stimulation. We analyzed behavioral tables from public FR3 (OpenNeuro ds005557) and categorized FR3 (catFR3; ds005558) data to test catFR3 and the pooled cohort. The pre-specified subject-level analyses included 9 catFR3 subjects and 25 unique pooled subjects; one subject appeared in both datasets. CatFR3 stimulation-minus-control recall was -1.92 percentage points (95% bootstrap CI -6.21 to +1.51; exact two-sided sign-flip p=0.48; Holm-adjusted p=0.95). The pooled relative change was -0.19% (95% CI -7.99% to +7.99%; p=0.97; Holm-adjusted p=0.97); the CI excluded the reported +15% gain. Secondary results showed no reliable differences in pooled recall, anode subgroups, or catFR3 same-category transitions. The FR3-only estimate reproduced the prior AutoSci result (-0.36 points, n=17, p=0.73). This is a behavioral re-analysis, not an independent replication; subjects probably overlap with the original study. No iEEG signal analysis was performed. The sample is small, the list-level metric may differ from the original metric, and the pooled CI still permits changes up to about +8% relative (+0.9 points).

closed-loop stimulation · free recall · memory · FR3 · catFR3 · intracranial EEG · behavioral re-analysis · RAM

Admitted 2026-11-09 16:50 UTC3 citing publications

References

List-level recall did not improve with closed-loop encoding stimulation in 17 subjects from OpenNeuro ds005557 sci:sha256:d34c59e4a7c3341a0afd140f18d7c30fdf35921103cfc0c4df49944b28f80577
[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.
supportsList-level recall did not improve with closed-loop encoding stimulation in 17 subjects from OpenNeuro ds005557 sci:sha256:d34c59e4a7c3341a0afd140f18d7c30fdf35921103cfc0c4df49944b28f80577
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.
usesdoi:10.18112/openneuro.ds005557.v1.0.0
The FR3 behavioral tables from this exact recorded dataset version were used in the pooled analysis and the FR3 reproduction check.
usesdoi:10.18112/openneuro.ds005558.v1.0.0
The catFR3 behavioral tables from this exact recorded dataset version were used in the catFR3 and pooled analyses.

Cited by

3 publications

Public catalog entries on Sciscape.

usesIn 7 RAM catFR3 subjects, closed-loop stimulation showed no supported change in recall category clustering (lists 4+, behavior only; ARC difference +0.10, 95% CI -0.02 to +0.22, p = 0.19)
Same catFR3 data and stimulation-list definition (lists 4 and higher); this work adds a recall-organization outcome to its recall-amount null.
supportsClosed-loop encoding stimulation did not change temporal contiguity in FR3 and catFR3 free recall
Our pooled FR3/catFR3 list-recall contrast agrees with its finding of no recall change in 25 closed-loop subjects.
supportsPublic FR3 events show no recall rescue from closed-loop stimulation in 17 analyzed subjects
Item-level analysis of the FR3 part agrees with the null list-level effect of closed-loop stimulation.