Intelligent Parkinson's Disease Classification and Progress Monitoring using Non-invasive Techniques

• Author / Creator

  Sara Soltaninejad

• Parkinson's disease (PD) is the second major neuro-degenerative disorder caused by dopaminergic loss in the brain region known as the Substantia Nigra (SN). The major symptoms of this disease are motor and non-motor abnormalities, which may show at early stages of PD. Physical exam, demographic characteristics, and neuroimaging analysis have been commonly used to diagnose PD. However, such assessments by clinicians are subjective, time-consuming, expensive, and prone to error. In my research, the goal is studying smart, non-invasive, and practical approaches to help clinicians assess PD, intending to diagnose and monitor the disease progression.

  The first section of this research is gait analysis for PD. We propose a non-invasive approach using movement data captured by Kinect to monitor the motor deficits of PD patients. The motion data of standard exercises, normally performed in rehabilitation sessions by physiotherapists, are collected including Step Length (SL), tremor, and Time up & Go (TUG). The standard medical Unified Parkinson's Disease Rating Scale (UPDRS) is used by a physical therapist to determine the level of severity as the ground truth. The general framework after obtaining the motion data includes two steps; feature extraction from the kinematic motion data, and classification using Random Forest (RF) and K-means. Freezing of Gait (FOG) is a short absence or reduction in the ability to walk for PD patients that may result in a fall, decrease in patients’ quality of life, and even death. Existing FOG assessments by doctors are based on a patient's diaries and experts' manual video analysis, which give subjective, inaccurate, and unreliable results. In the present research, an automatic FOG assessment system (Kin-FOG) is designed for PD patients to provide objective information to neurologists about the FOG condition and the symptoms’ characteristics. The proposed FOG assessment system utilizes Kinect for capturing data.

  Clinical characteristics of the patients have essential information for the specialist in PD diagnosis. However, assessment of this information by doctors can also be subjective, vulnerable to human errors. Therefore, the second section of my research targets automatic, early and non-invasive assessment of PD using clinical properties with machine learning.

  Neuroimaging has been successfully used for diagnosing the neurological disease. Magnetic Resonance Imaging (MRI) is one of the most popular methods due to its non-invasive nature and high resolution images. The most common MRI sequences are T1-weighted and T2-weighted scans. In the third section of this research, data analysis is conducted for PD diagnosis using MR images with machine learning. The proposed method follows four steps; preprocessing, feature extraction, feature selection, and classification. Preprocessing pipeline is performed using different medical libraries; Freesurfer, SPM/CAT12, and FSL. Thereafter, subcortical and region-based features are extracted from the preprocessed MR images. Feature selection or dimensionality reduction is performed by Principal Component Analysis (PCA). In the classification, two important machine learning algorithms are used, Support Vector Machine (SVM) and RF. Furthermore, we assess four deep-learning based models that classify patients based on the biomarkers in MRI data. In the last part of the neuroimaging analysis, SN region in MRI T2 and T1 are evaluated using Local Binary Pattern (LBP) and Histogram Oriented Gradient (HOG) features.

• Subjects / Keywords

  • Machine Learning
  • Clinical Information Analysis
  • Neuroimaging Analysis
  • Parkinson's Disease
  • Gait Analysis

• Graduation date

  Spring 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-41sk-s535

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.