Fangda Leng ^1,2 Yue Gao ² LI Fan ² Luhua Wei ² Yun Chuang Sun ² Fang Liu ³ Ying Zhu ⁴ Jianxing Qiu ⁴ Zhaoxia Wang ^2* Yiwei Zhang ^4*

• ¹ University of California, San Francisco, San Francisco, United States
• ² Department of Neurology, First Hospital, Peking University, Beijing, Beijing Municipality, China
• ³ Department of Neurology, First Hospital of Tsinghua University, Beijing, Beijing Municipality, China
• ⁴ Department of Radiology, First Hospital, Peking University, Beijing, Beijing Municipality, China

Background: Iron deposition has been observed in Parkinsonism and is emerging as a diagnostic marker for movement disorders. Brain functional network disruption has also been detected in parkinsonism, and is believed to be accountable for specific symptoms in parkinsonism. However, how iron deposition influences brain network remains to be elucidated.Methods: We recruited 16 Parkinson's disease (PD), 8 multiple system atrophy (MSA) and 7 progressive supranuclear palsy (PSP) patients. T1-weighted, susceptibility weighted images and resting-state functional MRI (rs-fMRI) were acquired. Quantitative susceptibility mapping (QSM) analysis was performed to quantify iron deposition in substantia nigra, putamen and dentate nucleus. Cerebellar network, sensorimotor network, default mode network and language networks were segregated using independent analysis. Network and iron deposition status were evaluated in relation to diagnostic groups, motor and non-motor symptoms. The relationship between quantitative iron deposition and brain network status was further interrogated. To further validate the findings, 13 healthy controls and 37 PD patients who had available T1 and rs-fMRI scans were selected from Parkinson's progression markers initiative (PPMI) database, and network analysis was performed.Results: In local cohort, compared to PD, MSA patients showed greater iron deposition in putamen, while PSP patients had greater iron deposition in caudate nucleus and thalamus. Cerebellar and language networks showed significant difference across diagnostic groups, while default mode network and sensorimotor network did not. MSA patients had significantly impaired cerebellar network and language networks compared to PD patients. Cerebellar network was positively associated with motor symptom scores while language network was positively associated with MoCA scores in the patients. Iron deposition was negatively associated with both networks' activity in the patients. In PPMI cohort, impairment was found in both cerebellar and language networks in PD. Cerebellar and language networks correlated with motor and cognitive impairment respectively. Conclusions: Cerebellar network and language networks are differently influenced in MSA, PD and PSP, which can server as another diagnostic marker. Impairment of cerebellar network and language network are associated with motor symptoms and cognitive impairment, respectively. Moreover, dysfunction of the networks is associated with iron deposition in deep nuclei (SN, DN, Putamen).