Open-loop encoding stimulation reduced item memory across four RAM tasks, with no significant task heterogeneity or site effect

Question: is there one consistent item-level memory cost of open-loop encoding stimulation across four RAM stimulation tasks? Do the task or the stimulation site (MTL vs other) change the cost? Data: behaviour only. Events.tsv files of OpenNeuro ds005489 (FR2), ds005491 (catFR2), ds005494 (PAL2) and per-trial scores of ds005523 (YC2), from archived copies in prior publications. Method: pre-registered plan (SHA-256 hashed). Unit = subject x task. Effect = memory for stimulated items minus memory for non-stimulated items. Per-task standardized effect dz, DerSimonian-Laird random-effects pool (H1), Cochran Q (H2), MTL-minus-other permutation within task (H3), Holm correction. Results: 98 units, 64 subjects. H1 pooled dz = -0.32 (95% CI -0.52 to -0.12, Holm p = 0.006). Pooled raw effect = -1.7 points (-2.8 to -0.6). H2: Q = 0.26, df = 3, I2 = 0 %; task dz values are -0.23 to -0.37. H3: MTL minus other = -0.21 SD (Holm p = 0.57). One unit per subject (n = 64): -0.31 SD, p = 0.020. Only FR2 is significant alone (p = 0.030). Conclusion: the single-task null results for catFR2, PAL2 and YC2 agree with one small common cost; they do not show the absence of an effect. Limits: 21 subjects contribute to more than one task; sites are clinically selected; YC2 outcome is a spatial score, not recall. Package: code, plan, all input files with SHA-256, per-unit table and results.

open-loop stimulation · item memory · free recall · paired-associate learning · spatial memory · hippocampus · memory encoding · electrical stimulation