Hippocampal theta and high-frequency activity did not predict object-placement accuracy in 24 YC1 patients

This pre-registered analysis tested whether hippocampal power during passive encoding predicted later object-placement accuracy in session 0 of the YC1 virtual navigation task. Intracranial EEG from 24 SEEG patients (1,048 valid test trials) was read from OpenNeuro ds005522 v1.0.0 using HTTP byte ranges. For each object, log power in six bands was averaged over full 1-s records in its two 5-s learning periods, z-scored within hippocampal bipolar pairs, and averaged across pairs. Placement accuracy was scored against random locations in each patient's arena; subject-level Spearman correlations were combined with a sign-flip test and bootstrap confidence intervals. Neither primary measure predicted accuracy: theta (3–8 Hz) mean Fisher z = +0.053 (95% CI −0.012 to +0.121; 15/24 positive; p = 0.13; Holm p = 0.27) and high-frequency activity (70–150 Hz) mean z = −0.021 (95% CI −0.076 to +0.036; 11/24 positive; p = 0.50; Holm p = 0.50). The exploratory 30–60 Hz median split was positive (p = 0.022, uncorrected). Left-hippocampus theta was +0.070 (n = 18, p = 0.080); right-hippocampus theta was +0.013 (n = 15, p = 0.73). The primary tests do not support group-level prediction by theta or high-frequency activity. The theta confidence interval remains compatible with a small positive effect (rho up to about 0.12). Limits include session 0 only, about 44 trials per patient, dataset region labels, and no interictal-spike rejection beyond an amplitude rule. A non-hippocampal control was dropped in a pre-statistics amendment because of network bandwidth.

spatial memory · hippocampus · intracranial EEG · SEEG · theta · high-frequency activity · object-location memory · subsequent-memory effect · YC1 · OpenNeuro