The encoding memory effect did not change significantly across early and late lists in 51 SEEG patients

We compared the word-encoding subsequent memory effect (SME) between early and late full 12-word lists in 51 patients with pure SEEG from OpenNeuro ds004789 (FR1, session 0); three of 54 patients were excluded by preregistered rules. In preregistered tests with Holm correction over three measures, no early-to-late change was significant. The tilt SME (z 70–150 Hz minus z 3–8 Hz) changed by −0.037 z (95% bootstrap CI −0.074 to −0.0003; sign-flip p = 0.060; Holm p = 0.18); its raw SME was +0.093 early and +0.056 late. Changes were +0.022 for 3–8 Hz and −0.015 for 70–150 Hz (Holm p = 0.37 for each). After secondary adjustment for serial position, the tilt change was −0.003 z (95% CI −0.042 to +0.036; p = 0.89). Recall probability was 0.276 early and 0.266 late (p = 0.33). Exploratory analysis found increased recency in late lists (+0.051 for positions 10–12 relative to 4–9; p = 0.036, uncorrected), while the primacy change was not significant (−0.020; p = 0.38). The results do not support a decrease in the serial-position-adjusted word-level memory signal. The study covers session 0 and averages all bipolar pairs; list order and session time are confounded, and changes smaller than approximately 0.04 z are not excluded. The analyzed OpenNeuro release is not identified in the archived publication.

subsequent memory effect · SEEG · free recall · serial position · encoding · intracranial EEG · memory