Across 53 patients, encoding HFA decline did not predict primacy in free recall

We tested whether between-patient differences in word-encoding high-frequency activity (HFA, 70–150 Hz) tracked free-recall behavior in 53 patients with pure SEEG from session 0 of the FR1 task. Inputs were archived per-word band-power features for 54 patients (bipolar pairs; 0–1.6 s after word onset, timed from events.tsv sample/fs); one patient with no recalls was excluded. HFA decreased with serial position in 46 of 53 patients, and mean behavioral primacy was +0.20, but HFA slope and primacy did not correlate (Spearman rho +0.25, 95% CI −0.05 to +0.52, permutation p = 0.075, Holm p = 0.15). This direction was opposite to the attention-gradient prediction; the interval excludes rho below −0.05. HFA subsequent memory effect (SME) did not track recall rate (rho −0.004, 95% CI −0.29 to +0.29, Holm p = 0.97). In a secondary, uncorrected theta analysis, SME and recall rate were positively correlated (rho +0.32, 95% CI +0.05 to +0.55, p = 0.024); mean theta SME was negative (d = −0.12, p = 0.002). These between-patient estimates may be attenuated by electrode coverage differences and few lists; participants were epilepsy patients taking antiepileptic drugs, with one session per patient.

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