Jingye, Yee (2022) Clinically compliant smart diagnosis system for upper limb spasticity. Doctoral thesis, Universiti Tun Hussein Onn Malaysia.

Preview	Text 24p YEE JINGYE.pdf Download (872kB) | Preview
	Text (Copyright Declaration) YEE JINGYE COPYRIGHT DECLARATION.pdf Restricted to Repository staff only Download (500kB) | Request a copy
	Text (Full Text) YEE JINGYE WATERMARK.pdf Restricted to Registered users only Download (116MB) | Request a copy

Abstract

Spasticity is a velocity dependent increase in muscle tone triggered by the increased excitability of the muscle stretch reflex, which could bring about huge physical and emotional impacts on the patients and the family members. The clinical management of spasticity is linked to the severity of the spasticity. Thus, an accurate and reliable assessment of severity is important. The Modified Ashworth Scale (MAS) is one commonly used clinical scale for spasticity assessment based on the resistance against passive movement about a joint with varying degrees of velocity. Due to the ambiguity of the qualitative description of the MAS, the spasticity assessment can be subjective and adequate training is required to ensure interrater reliability. The outcome of this research is a smart diagnosis system for upper limb spasticity based on MAS, which integrates mechatronics system and data-driven classification model. A 3C Framework is designed as the development blueprint for the medical computer-assisted diagnosis system, which includes the Conceptualisation, Creation, and Pre-Clinical Validation phases. In this work, a clinical database of upper limb spasticity for the Malaysian population is developed, and the collected clinical data is used for the development of the data-driven classification model. The classification model is supported by quantitative measurement data acquired with the data acquisition system from 50 recruited subjects. Important features as discussed with the physicians were extracted from the clinical data, and a Logical SVM-RF classifier is developed. This classifier combines the decision-making logic of the physicians and the prowess of different machine learning classifiers, achieving a 91% accuracy in diagnosing the subjects of 6 different MAS levels. This work digitalises the clinical assessment of spasticity as a first step towards the integration of artificial intelligence (AI) and data analytics into the Malaysian clinical setting

Actions (login required)

View Item