SEEG spectral map fidelity did not decrease with age in 255 RAM patients aged 18–64

We tested the neural dedifferentiation prediction that the spatial pattern of SEEG spectral features becomes less specific to brain location with age. Using archived bipolar-pair features from 255 RAM patients aged 18–64 years (median 34), we compared each patient's feature values with a leave-one-patient-out group map at the same MNI midpoints. The primary score averaged Fisher-transformed Spearman correlations across relative theta, alpha, and beta power, aperiodic exponent (2–40 Hz), and log broadband power (2–45 Hz). Adjusting for pair count, spatial spread, sex, and site, fidelity did not decrease with age (partial rho +0.075, 95% CI −0.051 to +0.201; permutation p=0.23); no individual feature showed a decrease after Holm correction. Results were similar with 5 mm and 15 mm map kernels and without site adjustment. Within-patient spread increased with age for relative theta (partial rho +0.183, Holm p=0.019) and exponent (+0.175, Holm p=0.024), opposite to the dedifferentiation prediction for spread. This spread may reflect age effects on feature means and is not a direct measure of regional distinctness. The evidence is limited to epilepsy patients, one 60 s task segment, MNI midpoints without tissue labels, ages 18–64, cross-sectional data, and clinically selected implant locations.

SEEG · neural dedifferentiation · spectral features · aging · spatial specificity · RAM · intracranial EEG