Hippocampal theta and beta power rose with set size during encoding, while maintenance power showed no load effect in 11 SEEG patients

This independent analysis examined hippocampal stereo-electroencephalography (SEEG) power during a verbal Sternberg working-memory task with set sizes of 4, 6, or 8 consonants. It used the first session of 11 patients from OpenNeuro ds004752 v1.0.1, with 33–50 correct, artifact-free trials per patient and 79 bipolar pairs that had at least one hippocampal contact. For each pair, Spearman correlations related set size to log10 band power in fixation (0–1 s), encoding (1–3 s), and maintenance (3–6 s); patient-level mean correlations were tested with exact two-sided sign-flip tests and Holm correction across five bands. In the primary maintenance analysis, no band from theta through high-frequency activity (4–150 Hz) changed with set size (mean rho from −0.062 to +0.056; all Holm p ≥ 0.49). In the secondary encoding analysis, theta and beta power increased with set size in 10 of 11 patients (mean rho +0.073 and +0.065; Holm p = 0.010 and 0.016). These small associations (rho about 0.07) may reflect the different visual displays for the three set sizes, rather than memory load. Maintenance power exceeded fixation power in theta, alpha, beta, and gamma bands; response time increased with set size. A same-shaft neocortical comparison included only four patients and could not establish a regional difference. Limits include the small partial sample, atlas labels not checked on patient images, no seizure-onset contact exclusion, and possible display-offset response in the first 0.5 s of maintenance. The report, complete analysis code, source-file list, EDF header hashes, and numerical results are preserved with the original files.

stereo-electroencephalography · working memory · hippocampus · spectral power · verbal Sternberg task · OpenNeuro ds004752