Altered cortical network profile of across-window ALFF dispersion in trigeminal neuralgia

Dynamic resting-state functional magnetic resonance imaging (fMRI) has revealed temporal variability in local activity and functional connectivity in trigeminal neuralgia (TN). Prior analyses focused on voxels, regions of interest, or connectivity states, leaving unclear whether group differences in the across-window dispersion of the amplitude of low-frequency fluctuation (ALFF) vary among large-scale cortical networks. Thirty-eight patients with TN and 31 healthy controls (HCs) were included. Blood oxygen level–dependent time series from Schaefer-400 parcels were converted to percent signal change (PSC), subjected to confound regression, and analyzed using sliding windows of 50 repetition times. Across-window ALFF dispersion (dALFF-SD) was quantified as the standard deviation of window-level ALFF estimates. Parcel values were spatially z-standardized within participants and averaged within seven Yeo networks. Generalized estimating equations tested the group × network interaction, adjusting for age, sex, and mean framewise displacement. NeuroQuery term maps and normative receptor/transporter maps provided exploratory spatial annotation. The cortical network profile of spatially standardized dALFF-SD differed between groups (Wald χ²(6) = 16.50, P = 0.011). After false discovery rate (FDR) correction, TN showed greater relative limbic-network prominence (adjusted TN–HC difference: β = 0.265, 95% CI 0.081–0.450; q = 0.033). The limbic-versus-other-networks contrast was also greater in TN (β = 0.305, 95% CI 0.075–0.535; Holm-adjusted P = 0.020) and remained positive across alternative window, atlas, head-motion, signal-quality, and resampling analyses. Unstandardized PSC dALFF-SD showed no group × network interaction (Holm-adjusted P = 0.137). The strongest positive NeuroQuery correlations were with “behavioral inhibition” ( r = 0.429), “frustration” ( r = 0.423), and “anxiety” ( r = 0.415); the strongest receptor/transporter correlation was with the normative µ-opioid receptor map (ρ = 0.375). None remained significant after within-family FDR correction. The cortical network profile of dALFF-SD differed between patients with TN and HCs, with relative limbic-network prominence after spatial standardization. These findings extend dynamic fMRI research in TN from local regions to large-scale cortical networks. Not applicable.