Altered brain network properties in patients with idiopathic normal pressure hydrocephalus: a diffusion spectrum imaging study

Idiopathic normal pressure hydrocephalus (iNPH) is a treatable dementia syndrome, yet the relationship between cerebrospinal fluid dynamic disturbance and large-scale white-matter network disruption remains incompletely understood. This study used diffusion spectrum imaging (DSI) and graph-theoretical analysis to characterize structural connectome alterations in definite iNPH and to explore their potential clinical relevance. We acquired DSI data from 23 patients with definite iNPH and 20 normal controls (NCs). Whole-brain structural connectomes were reconstructed to evaluate global and nodal topological properties. Network-based statistics (NBS) were used to identify altered subnetworks, and exploratory logistic regression and ROC analyses assessed the apparent diagnostic performance of selected network metrics. Compared with NCs, patients with iNPH showed increased network segregation, including a higher weighted clustering coefficient ( p = 0.027) and higher transitivity ( p < 0.01), together with impaired global integration, including prolonged characteristic path length ( p = 0.036). NBS identified weakened structural connectivity primarily involving interhemispheric frontal, parietal, and cingulate pathways. A same-cohort exploratory model combining binary transitivity and weighted assortativity coefficient yielded an apparent AUC of 0.867, with sensitivity of 0.955 and specificity of 0.750. Patients with iNPH exhibited structural connectome alterations characterized by increased segregation, impaired integration, hub reorganization, and disrupted interhemispheric connectivity. DSI-derived network metrics may provide complementary exploratory markers of iNPH-related network disruption, but their diagnostic utility requires validation in larger independent cohorts.