Changes of Brain Structure and Structural Covariance Networks in Parkinson's Disease with Different Sides of Onset

Tianqi Xu ¹ Zhihuai Deng ¹ Yinhui Yu ¹ Wenchao Duan ¹ Zeyu Ma ¹ Haoran Liu ¹ Lianling Li ¹ Moxuan Zhang ² Siyu Zhou ² Pengda Yang ² Xueyan Qin ³ Zhenyu Zhang ¹ Fangang Meng ^1,2 Yuchen Ji ^1*

• ¹ First Affiliated Hospital of Zhengzhou University, Zhengzhou, China
• ² Beijing Neurosurgical Institute, Beijing Tiantan Hospital, Capital Medical University, Beijing, Beijing Municipality, China
• ³ Yantai Affiliated Hospital of Binzhou Medical University, Yantai, China

Background: Parkinson's disease (PD) typically presents with unilateral symptoms in early stages, starting on one side and progressing, with the onset side showing more severe motor symptoms even after bilateralization. This asymmetry may reflect complex interactions among multiple brain regions and their network connections. In this study, we aimed to use surface-based morphometry (SBM) and structural covariance networks (SCNs) to investigate the differences in brain structure and network characteristics between patients with left-onset PD (LPD) and right-onset PD (RPD). Methods: A total of 51 LPD and 49 RPD patients were recruited. Clinical assessments included the Unified Parkinson's Disease Rating Scale motor section, Hoehn and Yahr stage, Mini-Mental State Examination, Parkinson's Disease Questionnaire, and Beck Depression Inventory. All participants underwent 3T structural MRI. Freesurfer was used to perform vertex-wise comparisons of cortical surface area (CSA) and cortical thickness (CT), whereas the Brain Connectivity Toolbox was implemented to construct and analyze the structural covariance networks.In patients with LPD, we found reduced CSA in the right supramarginal gyrus (SMG), right precuneus (PCUN), left inferior parietal lobule (IPL), and left lingual gyrus (LING) compared to RPD, while no significant differences in CT were found between the two groups. The CSA of the right PCUN showed a significant positive correlation with MMSE score in LPD patients. In our SCNs analysis, LPD patients exhibited increased normalized characteristic path length and decreased small-world index in CSA-based networks, while in CT-based networks, they showed increased small-world index and global efficiency compared to RPD. No significant differences in nodal characteristics were observed in either CSA-based or CT-based networks between the two groups.In patients with LPD, reductions in CSA observed in the right PCUN, right SMG, left IPL, and left LING may be associated with cognitive impairments and hallucinations among non-motor symptoms of PD. Additionally, the SCNs of LPD and RPD patients show significant differences in global topology, but regional node characteristics do not reflect lateralization differences. These findings offer new insights into the mechanisms of symptom lateralization in PD from the perspective of brain regional structure and network topology.