A six-subject pilot found no corrected association between hippocampal band power and verbal memory load

We examined whether hippocampal oscillatory power during verbal working-memory maintenance changed with set size in session 1 of six participants from the 15-participant OpenNeuro ds004752 SEEG dataset. We analyzed 41 bipolar pairs and a median of 40–45 correct, artifact-free trials per participant. Welch log10 power was calculated in theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), low gamma (30–70 Hz), and high gamma (70–145 Hz) during maintenance (3–6 s) and fixation (0–1 s). For each pair, we calculated Spearman correlation between set size and power; participant summaries used mean Fisher z, with exact sign-flip tests and Holm correction across bands. Response time increased with set size in all six participants (mean rho 0.35, p = 0.031), the smallest possible p with six participants. No band passed Holm correction. Maintenance theta showed a small positive trend (mean rho +0.106, 5/6 participants, p = 0.094; Holm p = 0.469); high gamma showed a small negative trend (mean rho −0.107, 2/6, p = 0.188; Holm p = 0.750). Fixation effects were near zero. This pilot does not show that a load effect is absent: it includes six of 15 participants, one session each, and has low power. Atlas labels were not checked, a bipolar pair could include a contact outside the hippocampus, and epileptic spikes were not detected. Cross-frequency and hippocampal-cortical coupling were not tested.

hippocampus · working memory · SEEG · verbal memory · memory load · oscillatory power · theta · high gamma · pilot study