From mild cognitive impairment to normal cognition in Parkinson's disease: The role of brain network architecture

Introduction: The state of mild cognitive impairment related to Parkinson's disease (PD-MCI) is unstable. However, the mechanisms underlying reversion to normal cognition in patients with early PD-MCI are still not fully understood. Objectives: To investigate whether reversion to normal cognition in PD-MCI is associated with the architecture of brain structural covariance networks (SCNs). Methods: Graph-theoretical analyses of SCNs were conducted using MRI-derived cortical thickness and subcortical volume measures from 119 drug-naïve PD patients and 51 healthy controls (HCs) at baseline. Among patients, 16 reverted from PD-MCI to normal cognition (Reverters) at the 1-year follow-up, 39 showed persistent PD-MCI (Non-Reverters), and 64 remained cognitively normal (PD-CN). Hub identification and global (efficiency, path length, clustering, small-worldness) and regional (clustering, efficiency) measures of SCNs were compared across groups. Results: Higher educational level and better cognitive performance were associated with reversion to normal cognition at the 1-year follow-up. At baseline MRI assessment, 15 hubs were identified in HCs and PD-CN, 8 in Reverters, and 5 in Non-Reverters. Reverters exhibited higher global efficiency than the other groups. Non-Reverters showed higher regional efficiency and clustering than PD-CN and HCs in the precentral regions. Conclusions: The loss of brain regions as network hubs within SCNs reflects cognitive impairment in early PD, suggesting a close link between the organization of brain SCNs and clinical outcomes. Cognitive resilience appears to be associated with a distinct SCN architecture at the earliest PD stages. These findings may contribute to understanding of the mechanisms underlying PD-MCI reversion and inform future treatments.