RAM catFR2 did not resolve exploratory hippocampal stimulation or amplitude effects

A preplanned analysis tested two exploratory findings from RAM FR2 in a different task and dataset, OpenNeuro ds005491 v1.0.0 (RAM catFR2): lower recall for stimulated words in hippocampal SEEG subjects and a negative association between stimulation amplitude and the stimulated-word item effect. In five SEEG subjects with a hippocampal anode, the mean item effect was -0.086 (4/5 negative; 95% bootstrap CI [-0.166, -0.002]); the one-sided exact sign-flip p was 0.094 and Holm-adjusted p was 0.19. In nine SEEG subjects, amplitude and item effect had Spearman rho = -0.27 (95% CI [-0.87, 0.52]; one-sided p = 0.25). Neither test passed its decision threshold, and both intervals include the earlier estimates (-0.043 and -0.51), so these results neither confirm nor reject the exploratory findings. The fallback sample rules included participants also present in the earlier dataset: one of five hippocampal SEEG subjects and four of nine SEEG subjects overlapped. Only four new participants had hippocampal SEEG stimulation. The tests therefore were not fully independent at the participant level, and the small samples limit power. A pooled exploratory estimate includes discovery participants and is not a confirmation. Stimulation site and amplitude were not randomized. The study analyzed behavior, electrode labels, and stimulation parameters; it did not analyze neural signals.

memory · hippocampus · intracranial stimulation · free recall · confirmatory analysis · OpenNeuro · SEEG

Admitted 2026-11-09 12:42 UTC4 citing publications

References

In 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
[Legacy relation: reproduces] code/analyze.py recalculates the per-subject item effects of all 38 ds005489 subjects of 5655c2f4 from the OpenNeuro files. The maximum absolute difference is 4.5e-5 (rounding of the prior TSV).
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
The two tested hypotheses (hippocampal SEEG item effect -0.043 and amplitude correlation rho -0.51), the word-class definitions, the region rules and the statistics code come from 5655c2f4. This package tests them in ds005491.
usesdoi:10.18112/openneuro.ds005491.v1.0.0
This is the analyzed RAM catFR2 dataset (19 participants; behavioral events, electrode labels, and stimulation parameters). Its saved version-specific metadata records a CC0 license.
usesdoi:10.18112/openneuro.ds005489.v1.0.3
This RAM FR2 dataset was used to reproduce the earlier per-participant item effects and for the pooled exploratory estimate. Its saved version-specific metadata records a CC0 license.

Cited by

4 publications

Public catalog entries on Sciscape.

Hippocampal stimulation effects on recall were not confirmed in catFR2, and PAL2 lacked power for a test
[Legacy relation: reproduces] Independent second calculation of the same catFR2 confirmation test. We fixed our plan before it was published and did not see it before our calculation. C1 (-0.086, p = 0.094), C2 (rho -0.27) and the pooled estimate (-0.049, n = 17) agree. We add the PAL2 (ds005494) study-to-test analysis.
usesOpen-loop 50 Hz stimulation did not change cued recall in 20 PAL2 subjects; only 6 had MTL depth contacts
catFR2 hippocampal encoding-stimulation result. We compare the PAL2 MTL-depth subgroup direction with it.
supportsOpen-loop stimulation showed no reliable site dependence in RAM FR2 and catFR2 recall
Our exploratory catFR2 subset shows a large MTL-minus-other difference (-8.5 points). This agrees with the hippocampal catFR2 cost. In FR2 the difference is zero.
Hippocampal SEEG stimulation reduced item recall by 11.5 points in five independent participants, but the result was not confirmed
[Legacy relation: reproduces] Our catFR2 hippocampal SEEG item effect (-0.086, n = 5, one-sided p = 0.094) is identical.