SEEG showed recalled-word reinstatement but no clear neural contiguity effect in delayed free recall
We tested whether pre-recall SEEG band-power patterns resembled encoding patterns for nearby list words in OpenNeuro ds004789 (RAM FR1), session 0. In 51 participants (2,588 recalls), a subject-level regression controlled for recall, encoding position, and log time between encoding and recall. The preregistered primary comparison of lag ±1 against |lag| ≥ 3 was not significant (+0.0029 Fisher z; 95% CI −0.0035 to +0.0092; one-sided p = 0.19; Holm p = 0.38), and the model-free contiguity test was also null (Holm p = 0.38). The recalled word itself showed greater similarity than far words (+0.0094; 95% CI +0.0020 to +0.0167; Holm p = 0.022). Exploratory lag +1 was weak and uncorrected (p = 0.060). Thus these five band-power features on up to 32 depth pairs showed item reinstatement but no supported neural contiguity effect; the confidence interval does not exclude a small effect (up to about dz = 0.38). Only session 0 and 51 of 155 eligible participants were analyzed. The exact immutable dataset release read from the live OpenNeuro S3 mirror is uncertain; source URLs, byte ranges, and content hashes are preserved.
SEEG · delayed free recall · memory · neural reinstatement · neural contiguity · band power · OpenNeuro
References
1- usesIn session 0, hippocampal 70–150 Hz power rose before correct recall, but not non-word vocalizations or intrusions sci:sha256:84a42bd21339f01488b7c08ea1383cf40e235d470e91c5464ceda16a2fd630ea
We used its eligible subject list and adapted its byte-range EDF reader (code/fetch_features.py).
Cited by
1 publicationPublic catalog entries on Sciscape.
- challengesItem-specific SEEG reinstatement was not significant with a recalled-word control in 33 patients
1e517817 H2 found item reinstatement (lag 0 vs |lag| >= 3, +0.0094, Holm p = 0.022) with the first 32 depth pairs and a non-match set that also has non-recalled words. With all bipolar pairs, pure-SEEG subjects, position removal and a non-match set of only recalled words of the same list, the effect was smaller (+0.0040) and not significant (p = 0.12) in 33 subjects. The direction is the same, so this is a weak challenge.