Identification of Influential Brain Regions in Parkinson’s Disease through Connectome Analysis of Neuro Images using Graph Theoretic Measures

Parkinson’s Disease (PD) exhibits complex neurodegenerative changes that manifest across structural and functional brain connectivity. The analysis of this region-to-region brain connectivity and the study of deviations in the signals emitted by them can point out Parkinson’s Diseasecausing neuro-biomarker. Existing detection systems rely on this connectome construction using diffused MRI or DTIs, and use a standard atlas to extract regions of interest to classify Parkinson’s pathology amidst normal. This work has used all three major standard atlases available, like Harvard-Oxford Atlas (HOA), Automated Anatomical Labelling atlas (AAL), and Shaefer – 400 or 1000 parcellations to label the regions of interest. And construct connectomes using Pearson correlation to compute local and global graph measures like degree, strength, closeness, and betweenness centralities. Also, have constructed node-level graph theoretic measures and compared the deviations in those measures between Parkinson’s and healthy controls. Then a thorough analysis with respect to each graph measure has been done, and the highly altered regions for each graph theoretic measure. Taken those specific regions which showed greater alterations in multiple measures computed, especially the thalamus, palladium, putamen, subcallosal cortex, etc., along with 30 more regions. The unnoted Heschl’s gyrus(bilateral) also stood out among the others. This is an initiative to early detect this neurodegenerative disorder that is becoming increasingly prevalent in elderly people, making their later years of life traumatic.