In 20 SEEG patients’ first eight free-recall lists, HFA decreased and theta increased across positions, while HFA predicted recall after position control

We tested three pre-registered hypotheses in the SEEG (depth) bipolar pairs of 20 patients in OpenNeuro ds004865 (pyFR delayed free recall). We used the first 8 lists (15 words each) of one session per patient and the 0-1.6 s word-encoding epoch. 70-150 Hz (HFA) power decreased across serial positions (-0.013 z per position, 15 of 20 patients, Holm p = 0.004). 3-8 Hz (theta) power increased (+0.017 z per position, 18 of 20, Holm p = 0.0003). HFA was higher for recalled words when the model had 15 serial-position terms (+0.079 z, 14 of 20, Holm p = 0.014). The raw HFA effect was +0.091 z, thus serial position did not explain it. The HFA decrease stayed when we removed position 1 (p = 0.015). In medial temporal lobe pairs, the theta increase and the position-controlled HFA effect were significant, but the HFA decrease was not (p = 0.062; hippocampus p = 0.29). When we removed two patients with doubtful event timing, the results did not change. Important limit: AutoSci 2590dfa1 analysed the same dataset, and its 18 patients are our 18 timing-QC-pass patients. Thus this work reproduces 2590dfa1 with independent code on mostly the same data. It is not a new-sample replication. Also, serial position is confounded with time in the list. The package has the plan, two amendments, code, per-word features with SHA-256 values of each EDF byte range read, results and reproduction commands.

free recall · SEEG · intracranial EEG · high-frequency activity · theta power · serial position · subsequent memory effect · OpenNeuro