Cortical morphometric network alterations in new daily persistent headache: a structural MRI study

New daily persistent headache (NDPH) is defined by a distinctly remembered onset followed by continuous daily headache, yet its cortical structural network basis remains poorly understood. We examined whether NDPH is associated with alterations in individual-level cortical morphometric similarity and assessed the spatial neurochemical and transcriptomic context of these changes. High-resolution T1-weighted magnetic resonance imaging was acquired from 35 patients with NDPH and 37 healthy controls. Individual Morphometric INverse Divergence (MIND) networks were constructed across 360 cortical parcels using cortical thickness, surface area, gray matter volume, mean curvature, and sulcal depth. Group differences in global and regional node-level MIND were tested using covariate-adjusted generalized additive models, and regional node-level MIND values were additionally summarized according to the Cole–Anticevic Brain-wide Network Partition (CAB-NP). Regional node-level effects were further characterized by seed-to-cortex edge decomposition. The unthresholded regional effect map was related to normative positron emission tomography receptor/transporter maps and Allen Human Brain Atlas gene-expression profiles using spin-based spatial testing and partial least squares regression. Global MIND was lower in NDPH (Cohen’s d = − 0.56, P = 0.011). Twelve parcels showed false discovery rate (FDR)-corrected reductions, involving frontal, cingulate, posterior insular–opercular, posterior cingulate, and visual association cortices. CAB-NP–based summaries showed lower mean regional node-level MIND for parcels assigned to the cingulo-opercular, Visual2, language, frontoparietal, dorsal-attention, and default networks. For the 12 parcels identified in the regional node-level analysis, 303 of 4,242 unique seed-related edges survived false discovery rate correction, and 302 were lower in NDPH, indicating broadly reduced seed-to-cortex morphometric similarity. The regional effect map was negatively associated with normative distributions of the µ-opioid receptor, metabotropic glutamate receptor 5 (mGluR5), and cannabinoid receptor type 1 (CB1). The fourth partial least squares component showed the strongest transcriptomic association but did not survive component-level false discovery rate correction; its negatively weighted genes were enriched for mitochondrial and respiratory-chain processes. In exploratory clinical analyses, no association survived FDR correction across the complete family of 12 regions × 8 clinical variables. NDPH is associated with distributed alterations in individual-level cortical morphometric organization, with edge-level decomposition showing that the regional reductions were characterized by broadly decreased seed-to-cortex similarity profiles. The neurochemical and transcriptomic findings are atlas-based and hypothesis-generating rather than evidence of patient-specific molecular abnormalities. ClinicalTrials.gov, NCT05334927. Retrospectively registered on April 19, 2022.