Delta-power asymmetry did not significantly predict verbal recall in 151 SEEG patients

We tested whether left-minus-right asymmetry in relative delta power predicts verbal free-recall rate in 151 bilaterally implanted RAM epilepsy patients (106 FR1, 45 catFR1; session 0; age 18–63 years). We used archived per-pair spectra from 60-second task segments, mirrored right bipolar pairs in MNI space, and matched them one-to-one to left pairs within 10 mm, yielding 3,437 matched pairs. The analysis plan was hashed before recall was joined to spectral features. The primary test was a partial Spearman correlation with site, sampling rate, sex, dataset, age, number of pairs, and bilateral mean power as covariates. The result was rho = −0.129 (95% bootstrap CI −0.306 to +0.038; permutation p = 0.115). The direction agreed with the prediction, but the test was not significant. No secondary asymmetry test passed Holm correction (smallest Holm p = 0.112). The age control was significant (rho = −0.291, p = 0.0004), using the same data as earlier age-recall results and therefore not an independent replication. Robustness tests without covariates, using unmatched pairs, or using temporal/limbic pairs gave rho values from −0.104 to −0.136, all p > 0.09. The FR1-only estimate was −0.022; the catFR1-only estimate was −0.350 (n = 45) and is descriptive only. The study had 80% power only for effects of about |rho| = 0.23–0.24 or larger; smaller effects remain possible. The open data do not identify seizure-onset or language-dominant side. The recordings came from the memory task rather than rest, matching controlled location only within each patient, and the between-patient design was correlational.

stereo-EEG · delta power · hemispheric asymmetry · verbal memory · free recall · epilepsy · spectral analysis · partial Spearman correlation