Abnormal cerebellar connectivity within the motor subnetwork in MSA with cerebellar dysfunction
Apurva Shah1, Shweta Prasad2, Santosh Dash3, Jitender Saini4, Pramod Kumar Pal3, and Madhura Ingalhalikar1

1Symbiosis Centre of Medical Image Analysis, Symbiosis International University, Pune, India, 2Department of Clinical Neurosciences and Neurology, National Institute of Mental Health and Neurosciences, Bangalore, India, 3Department of Neurology, National Institute of Mental Health and Neurosciences, Bangalore, India, 4Department of Neuroimaging & Interventional Radiology, National Institute of Mental Health and Neurosciences, Bangalore, India


Multiple system atrophy with cerebellar features (MSA-C) is a distinct subtype of MSA characterized by predominant cerebellar symptoms. Neuroimaging studies have demonstrated cerebellar abnormalities; however, abnormality of structural connectivity of the motor subnetwork has not been studied and this study aims to investigate this aspect. We observed impairment in the structural segregation, integration and network resilience with significantly reduced nodal strength and connectivity in several cerebellar as well as non-cerebellar regions that correlate with UMSARS scores. Our findings provide definitive evidence of abnormalities that may be causally implicated in the motor features of cerebellar dysfunction and Parkinsonism observed in MSA-C.


Multiple system atrophy (MSA) is a sporadic, neurodegenerative disorder characterized by available combination of progressive autonomic dysfunction, parkinsonism, cerebellar and pyramidal features1. MSA can be categorized based on the predominant motor symptoms as MSA with predominant parkinsonism (MSA-P) and MSA with predominant cerebellardysfunction (MSA-C)2. MSA-C represents a distinct motor subtype of MSA, and is characterized by gaitataxia, limbataxia, scanning dysarthria, and cerebellar oculomotor dysfunction in addition to autonomicdysfunction and parkinsonism3. This work focuses on the motor features of MSA-C and hypothesizes that patients with MSA-C would demonstrate significantly higher abnormality in the cerebellar network in comparison to the basal ganglia network. This subnetwork comprising of the precentral cortex, basal ganglia, thalamus, ventral diencephalon and cerebellum has been previously utilized in connectivity-based studies in Parkinson’s disease4,5.


All subjects (23 MSA-C, 25 HC, age and gender matched) were scanned on a 3T Philips Achieva MRI scanner using a 32-channel head coil. 3D-T1 structural image with high resolution were obtained with echo-time(TE)=3.7ms, repetition-time(TR)=8.1ms, field-of-view(FOV)=256x256x155mm, slice-thickness=1mm, voxel-size=1x1x1mm, sense-factor=3.5, acquisition matrix=256x256, flip-angle=80. Diffusion imaging(DWI) were acquired using a single-shot spin-echo, echo-planner sequence in axial section with TE=62ms. TR=8783ms, FOV=224x224mm, voxel-size=1.75x1.75x2mm; diffusion gradient direction=15, b-value=1000s/mm2 and single b=0s/mm2 image. To obtain complete connectome and determine the connectivity from various cerebellar nodes, two atlases Desikan atlas from Freesurfer-6.0(http://surfer.nmr.mgh.harvard.edu) and AAL atlas were fused to delineate the complete brain and multiple region of interest (ROIs) within cerebellum (from AAL). Using FSL-5.0.9(https://fsl.fmrib.ox.ac.uk), probabilistic tractography6 was implemented to track fibers from each node to every other node. Based on hypothesis of differences in the motor subnetwork, we extracted 38 ROIs to create a 38*38 subnetwork as shown in Figure-1. The ROIs selected included bilateral pre-central cortex, basal ganglia, thalamus, ventral diencephalon and cerebellum. Global, nodal measures were computed and analyzed for group difference using MANCOVAN model while co-varying for age and gender and performed False Discovery Rate(FDR) with significant q-value<0.05. Edge-wise analysis was performed using NBS for 10000 permutation with significant p-value<0.01. Further, global, nodal and edge strength were correlated with duration of illness, total-UMSARS and motor component scores. Pearson’s correlation was computed with age, gender and cerebellum volumes as co-variates and the significance threshold was maintained at p-value<0.05.


Significant decrease in density (corrected p-value<0.01), transitivity (corrected p-value<0.01), characteristic path length (corrected p-value<0.01) and clustering coefficient (corrected p-value<0.01) was observed in motor subnetwork of patients with MSA-C when compared to HC. Several regions with significantly reduced nodal strength within the motor subnetwork were observed in patients with MSA-C in comparison to HC (Figure-1). Edge-wise analysis revealed significantly lower connectivity in patients with MSA-C involving 40 connections of the subnetwork (Figure-2). A significant negative correlation was observed between connectivity of the thalamus and total UMSARS scores and UMSARS-II score (Figure-3). At the edge-level, a few connections showed significant correlations with the total UMSARS score, and motor severity score (part-II). The connections involved were between the left cerebellum_8 to vermis_7, left thalamus to left ventral diencephalon, left-cerebellum_10 to vermis_4,5, and left-cerebellum_8 to vermis_7. Duration of illness showed a negative correlation between vermis_6 to vermis_7. Only one connection from left-cerebellum_4,5 to right-cerebellum_4,5 demonstrated a positive correlation with duration of illness.


Our results provide definitive evidence of damage to connections which may be causally implicated in the motor features of cerebellar dysfunction and Parkinsonism observed in MSA-C. Abnormal global metrics of the motor subnetwork in the MSA-C are suggestive of impairment in the structural segregation, integration and network resilience of the motor subnetwork in MSA-C. At a nodal level, we observed significantly reduced nodal strength in several regions (Figure 1), indicating poor connectivity with other nodes. In edge-wise analysis, a higher proportion of abnormal connections involved nodes within the cerebellum and this observation probably substantiates the predominance of cerebellar dysfunction in MSA-C. The observed negative correlation between the UMSARS scores and thalamus connectivity suggests a definitive involvement of the motor subnetwork in MSA-C.


Significant alterations in the structural connectivity of the motor subnetwork at the global,nodal and edge levels are observed in MSA-C. The higher degree of abnormality observed in inter and intra-cerebellar connectivity corroborate with the predominant motor symptom of cerebellar dysfunction. Future studies exploring the structural connectivity of non-motor subnetworks are crucial to aid in a better understanding of the overall disease process in MSA-C.


Department of Science and Technology – Science Education and Research Board (DSTSERB)(ECR/2016/000808) provided partial funding for setting up the computing facility.


  1. Fanciulli, A. and G.K. Wenning, Multiple-system atrophy. N Engl J Med, 2015.372(3): p. 249-63.
  2. Gilman, S., et al., Second consensus statement on the diagnosis of multiple systematrophy. Neurology, 2008. 71(9): p. 670-6.
  3. Ciolli, L., et al., An update on the cerebellar subtype of multiple system atrophy.Cerebellum Ataxias, 2014. 1: p. 14.
  4. Barbagallo, G., et al., Structural connectivity differences in motor network between tremor-dominant and nontremor Parkinson's disease. Hum Brain Mapp, 2017. 38(9):p. 4716-4729.
  5. Lewis, M.M., et al., Differential involvement of striato- and cerebello-thalamocortical pathways in tremor- and akinetic/rigid-predominant Parkinson's disease.Neuroscience, 2011. 177: p. 230-9.
  6. Behrens, T.E., et al., Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? Neuroimage, 2007. 34(1): p. 144-55.


Figure-1: Node Strength of patients with MSA-C versus healthy controls. Reduces nodalstrength was observed in the left thalamus, left ventral diencephalon, right ventraldiencephalon, left cerebellum crus 1,2, left cerebellar lobule 4,5, right cerebellar lobule 4,5left cerebellar lobule 6, right cerebellar lobule 6, vermis 4,5 and vermis 7. Cereb:cerebellum; L: Left; R: Right; Ventral DC: Ventral diencephalon

Figure-2: Network based statistics based edge-wise analysis. Forty abnormal edge-wereconnections were observed in patients with MSA-C in comparison to healthy controls.Abnormalities were observed in connections between the cortex and basal ganglia structures,basal ganglia and thalamus, thalamus and cerebellum, and within the cerebellum i.e. lobulesand the vermis. Cereb: cerebellum; L: Left; R: Right; Ventral DC: Ventral diencephalon

Figure-3: Correlation graphs for significant correlations between connectivity metrics andtotal UMSARSscore, motor component score and duration of illness(DoI).(A) Negative correlations between the total UMSARSscore and the left thalamus, and between the total UMSARS score and edge connection between the left-cerebellum_8 to vermis7, left-thalamus-proper to ventral-diencephalon and left-cerebellum_10 to vermis4,5.(B) Negative correlations between the motor component score and the left-thalamus, and between the motorcomponent score and the edge connection between the left-cerebellum_8 and vermis7.(C) Negative correlation between the DoI and vermis6 to 7, and positive correlation between duration ofillness and left-cerebellum_4,5 and right-cerebellum_4,5.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)