Clinical Application of Resting-State Functional MRI for Individualized Localization of Core Language Areas in...

This preliminary study aimed to evaluate the clinical utility of resting-state functional magnetic resonance imaging (rs-fMRI) as an individualized tool for localizing core language areas in patients undergoing epilepsy surgery.

A retrospective analysis was conducted on 31 patients with drug-resistant epilepsy who underwent rs-fMRI-based language localization and stereo-electroencephalography (SEEG). The rs-fMRI data were processed using seed-based correlation analysis. Two neurosurgeons independently reviewed the data and reached a consensus on the rs-fMRI localization results. Language function outcomes derived from electrical stimulation of language-related electrodes during SEEG were independently analyzed and documented by two additional physicians. The consistency rate between rs-fMRI-identified language regions and the effects observed during electrical stimulation was subsequently calculated to compare the two techniques.

In the Broca area, left- and right-hemisphere dominance occurred in equal proportions, with an asymmetric subregional distribution (predominantly pars triangularis on the left and pars opercularis on the right). Lateralization in the Wernicke area was less clear, with similar proportions of left, right, and bilateral representation. SEEG validation showed that when rs-fMRI indicated unilateral localization, it reliably identified SEEG stimulation-negative regions (specificity ≥ 94%), but its sensitivity was low (33%-50%), meaning that it frequently missed SEEG stimulation-positive areas. When rs-fMRI showed bilateral Wernicke localization, its diagnostic performance was poor (specificity only 27.3%, with a negative AC1).

Resting-state fMRI identified language reorganization in 54.8% broca and 64.5% wernicke patients with drug-resistant epilepsy, suggesting atypical or non-left-dominant language lateralization. Before surgery involving critical language areas, rs-fMRI scanning can be considered as an auxiliary tool to localize core language regions. When rs-fMRI shows unilateral localization and the area is non-core, it might reduce the need for implanting SEEG electrodes for language mapping in that region. However, in cases of bilateral Wernicke core localization, additional language mapping methods should be combined. These preliminary findings require validation in larger prospective cohorts.