A pre-registered replication found no significant association between implant laterality and verbal or spatial memory in three RAM SEEG/ECoG tasks

This pre-registered replication tested whether the association between more left-sided SEEG/ECoG implants and lower free recall generalizes across verbal and spatial memory tasks. Using the same laterality index (LI) as the original study, we analyzed session 0 of PAL1 cued recall (n = 55), RepFR1 free recall (n = 38), and YC1 object-location memory (n = 48), adjusting for age, sex, site, and log lateral-contact count. Partial Spearman correlations were negative for the verbal tasks (PAL1 rho = -0.183, Holm p = 0.373, 95% CI [-0.464, 0.115]; RepFR1 rho = -0.132, Holm p = 0.418, 95% CI [-0.479, 0.211]) and positive for the spatial task (YC1 rho = +0.243, Holm p = 0.276, 95% CI [-0.123, 0.586]). The specificity contrast was +0.40 (95% CI [-0.03, +0.82]). No primary test was significant and all confidence intervals included zero, meeting the pre-registered failure-to-replicate rule. The signs agreed with a left-verbal/right-spatial pattern, but the small samples had only about 0.22 to 0.31 power for rho = -0.20. The results neither reject nor confirm the original effect. They also do not support a general memory decrease with left-sided implants. LI describes contact placement and is only a proxy for dysfunction laterality; no seizure-onset or language-dominance data were used. Many participants overlap with the original sample, handedness was not a covariate, only session 0 was analyzed, and YC1 error is in raw virtual-arena units.

implant laterality · SEEG · ECoG · verbal memory · spatial memory · free recall · replication · OpenNeuro · RAM