Analysis of cognitive dysfunction in parkinson s disease using morphometric and machine learning techniques

Abstract

Parkinson s disease (PD) is the second most common and progressive newlineneurodegenerative disorder. The loss of dopamine generating neurons in the newlinesubstantia nigra causes disturbances in the message receptors of the striatum. newlineThis results in a multitude of motor and non-motor symptoms. Beyond the newlineperception of PD as a movement disorder, non-motor symptoms appear at the newlineearly stages of the disease before the predominant motor symptoms appear. newlineCognitive dysfunction is a significant clinical non-motor symptom that newlineincludes impairment in executive functions, visuospatial reasoning, memory newlineand language processing. The phase of cognitive impairment is determined by newlineits severity from mild to end stage dementia. Clinical assessments used for newlineidentification of PD, could not capture the anatomical variations of the brain. newlineThe neuroimages such as Single Photon Emission Computed Tomography newline(SPECT) and Magnetic Resonance Imaging (MRI) are able to capture the newlinestructural and pathophysiological changes in PD. Early diagnosis of cognitive newlineimpairment in PD enables effective treatment measures that could slow or halt newlinethe progression of PD. newlineIn this work an attempt is made to analyse the cognitive impairment in newlinePD using morphometric and machine learning techniques. The Magnetic newlineResonance (MR) and SPECT images of Healthy Control (HC) and PD newlinesubjects obtained from the PPMI database are utilized in this work. The PD newlinesubjects are categorized based on the MoCA scores as No Cognitive newlineimpairment Parkinson s Disease (NC-PD), PD subjects with Mild Cognitive newlineimpairment (PD-MCI) and PD-Dementia (PD-D) subjects. Binarisation is newlineused to segment the striatal region from SPECT images of HC and PD subject newlinegroups newline