Background: Structural cortical networks (SCNs) reflect the covariance between the cortical thickness of different brain regions, which may share common functions and a common developmental evolution. SCNs appear abnormal in neurodegenerative conditions such as Alzheimer's and Parkinson's diseases, but have never been assessed in primary progressive multiple sclerosis (PPMS).
Objective: The aim of this study was to test whether SCNs are abnormal in early PPMS and change over 5 years, and correlate with disability worsening.
Methods: A total of 29 PPMS patients and 13 healthy controls underwent clinical and brain magnetic resonance imaging (MRI) assessments for 5 years. Baseline and 5-year follow-up cortical thickness values were obtained and used to build correlation matrices, considered as weighted graphs to obtain network metrics. Bootstrap-based statistics assessed SCN differences between patients and controls and between patients with fast and slow progression.
Results: At baseline, patients showed features of lower connectivity (p = 0.02) and efficiency (p < 0.001) than controls. Over 5 years, patients, especially those with fastest clinical progression, showed significant changes suggesting an increase in network connectivity (p < 0.001) and efficiency (p < 0.02), not observed in controls.
Conclusion: SCNs are abnormal in early PPMS. Longitudinal SCN changes demonstrated a switch from low- to high-efficiency networks especially among fast progressors, indicating their clinical relevance.
Keywords: Primary progressive multiple sclerosis; bootstrapping; cortical thickness; grey matter damage; robust statistical methods; structural covariance networks.