Hippocampal maintenance theta did not scale with working-memory load in one partial SEEG session

In a time-limited snapshot of session 1 from OpenNeuro ds004752 v1.0.1, we analyzed complete trials from 12 SEEG patients with at least 20 artifact-free trials each (21–92 trials). Hippocampal bipolar theta power during the 3 s maintenance period was not significantly associated with set size (mean Spearman rho = 0.054, 7/12 positive, exact sign-flip p = 0.23). No maintenance feature passed Holm correction, including after subtracting fixation power; no maintenance feature predicted response time within set size after correction. In encoding, hippocampal alpha power increased with set size (mean rho = 0.174, 10/12 positive, p = 0.0073, Holm-adjusted p = 0.044). This may reflect the visual display because higher loads showed more letters. The fixation control showed no effect (all p > 0.4). Downloads were truncated for 11 subjects, only complete 8 s trials were retained, and one session was analyzed. Thus statistical power is low, and a mean correlation near 0.05 would require many more trials to detect. Anatomical labels were template labels, and no interictal epileptiform discharge removal was performed beyond the dataset artifact flag. The result is limited to this partial snapshot; rerunning with all sessions is the next test.

intracranial EEG · SEEG · hippocampus · verbal working memory · theta power · alpha power · memory load · OpenNeuro · spectral analysis