No Desikan–Killiany region showed a corrected HFA memory effect after serial-position control in sample-timed FR1 data

This analysis mapped word-level subsequent memory effects across Desikan–Killiany (DK) cortical regions in session 0 of the OpenNeuro ds004789 FR1 SEEG cohort. It used archived per-word log band-power features for 54 subjects, cut from word onset using events.tsv sample/fs and spanning 0–1.6 s. For each bipolar pair and frequency band, list and serial-position effects were removed with OLS before comparing recalled with not-recalled words using Welch t statistics. Subject-region means were tested by sign flips, with Holm correction across regions in each family. No region passed Holm correction for the primary 70–150 Hz high-frequency activity (HFA) test; the largest uncorrected mean was fusiform (+0.245 t; p = 0.013; Holm p = 0.38). In the secondary 3–8 Hz theta family, recalled words had lower theta in parahippocampal (mean −0.669 t; Holm p = 0.0015) and inferior temporal regions (−0.403 t; Holm p = 0.041). The secondary HFA-minus-theta tilt was positive in parahippocampal (+0.912; Holm p = 0.0015) and fusiform regions (+0.552; Holm p = 0.032). The exploratory association between regional HFA effect and evoked index was not significant (rho = +0.35, permutation p = 0.069). DK labels cover cortex, not hippocampus or amygdala; clinical coverage differed between subjects, and the HFA null does not rule out effects of about ±0.2 t. The package preserves analysis code, source hashes, and result tables.

subsequent memory effect · SEEG · free recall · high-frequency activity · theta power · serial position · Desikan–Killiany atlas