Classification of handwriting kinematics in automated diagnosis and monitoring of Parkinson's disease

Zham, P 2019, Classification of handwriting kinematics in automated diagnosis and monitoring of Parkinson's disease, Doctor of Philosophy (PhD), Engineering, RMIT University.


Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Zham.pdf Thesis application/pdf 2.43MB
Title Classification of handwriting kinematics in automated diagnosis and monitoring of Parkinson's disease
Author(s) Zham, P
Year 2019
Abstract Parkinson's disease is one of the most prevalent neurodegenerative conditions. Currently, there is no standard clinical tool available to diagnose PD. One of the research priorities is to come up with biomarkers which will improve the diagnostic process and can be used for the clinical test. At present, the only way to assess this disease is by visually observing the symptoms of the patient which is performed only by expert neurologists. As of now, there is no treatment to prevent the progression of PD. However, there is an elemental drug `Levodopa' (L-dopa) available to control the disease by increasing dopamine cells in the brain. It is important to detect PD and start treatment in the early stages as it helps to control the symptoms and significantly delays the development of motor complications.

In this study fine motor symptoms handwriting has been studied. As a first objective I have conducted the experiments on the significant number of patients and age-matched control (112 Participants:56 PD and 56 controls), and thus completed the task of data collection. The system developed extracts the dynamic features of the handwriting/drawing, reports the possible strength of dynamic features providing a basis for automated analysis. The advantage of this approach is that patients are not required to follow complex commands, and the analysis can be fully automized. I anticipate that following appropriate clinical tests already planned, the system will be able to detect early disease symptoms remotely outside hospitals or clinics. It could also be used for self-evaluation by patients with neuromuscular and motor neuron disorders. This device can be used without compromising on the comfort level of Patients who may still prefer writing with an ink pen on plain paper.

This study proposes a new feature `Composite Index of Speed and Pen-pressure' (CISP) to distinguish between different stages of Parkinson's disease. The experiment also demonstrated a method which can be used with guided spiral drawing to improve classification results to predict Parkinson's disease. Further, I recommend using a panel of writing tasks which might prove to be an effective biomarker for cell loss in the substantia nigra and the associated dopamine deficiency. Thus, models developed can be used in designing an automated application for predicting and monitoring Parkinson's disease
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Engineering
Subjects Biomedical Instrumentation
Keyword(s) Parkinson's disease
Levodopa
Dynamic features
Kinematic features
Machine learning
Micrographia
Versions
Version Filter Type
Access Statistics: 98 Abstract Views, 87 File Downloads  -  Detailed Statistics
Created: Mon, 30 Sep 2019, 16:04:49 EST by Adam Rivett
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us