A detectable within-person memory effect occurred in 8 of 48 FR1 patients under a full-sensitivity model

In archived FR1 stereo-EEG (SEEG) features, 8 of 48 patients had a significant within-person subsequent memory effect (SME) for tilt, defined as z-scored 70–150 Hz power minus z-scored 3–8 Hz power. The Bayesian prevalence estimate was 0.12 (95% highest posterior density interval [HPDI] 0.03–0.25), and the count exceeded chance after Holm correction (p = 0.0073). Although 33 of 48 patients had a positive tilt statistic (sign test p = 0.013), the effect was small in most patients. HFA alone was significant in 5 of 48 patients (Holm p = 0.091); a theta decrease was significant in 7 of 48 (Holm p = 0.019). Permutation-based null checks were calibrated. Under the model's assumed within-person sensitivity of 1, the pre-registered rule classified the detectable effect as a minority effect. When assumed sensitivity was 0.5 or 0.3, the tilt prevalence HPDI widened to 0.07–0.53 or 0.15–0.90. Thus, the finding concerns effects detectable in one session and does not imply that most patients have no SME. The primary analysis used 48 eligible patients from 51 in the archived primary set; an exploratory analysis of all 54 archived subjects included 51 eligible patients and gave 9 of 51, with prevalence MAP 0.13 (95% HPDI 0.04–0.26). The archived features came from OpenNeuro ds004789, but the exact dataset release used to produce them is not established by the available provenance.

subsequent memory effect · stereo-EEG · SEEG · population prevalence · FR1 · permutation test · memory encoding