Wong, Patrick Qi Han (2023) Development of eeg classification model for post-stroke patients of different recovery stages using artificial neural network. Masters thesis, Universiti Tun Hussein Onn Malaysia.

Preview	Text 24p PATRICK WONG QI HAN.pdf Download (487kB) | Preview
	Text (Copyright Declaration) PATRICK WONG QI HAN COPYRIGHT DECLARATION.pdf Restricted to Repository staff only Download (324kB) | Request a copy
	Text (Full Text) PATRICK WONG QI HAN WATERMARK.pdf Restricted to Registered users only Download (22MB) | Request a copy

Abstract

Stroke rehabilitation is a therapeutic process aimed at maximising the patient's physical and social potential. Although patients can recover full consciousness from stroke, most of them cannot perform activities of daily living with the affected limbs. Impairment of the limbs significantly limits the level of activity as well as social and physical interactions of a post-stroke patient. The performances of patients are then clinically assessed using standardised scales such as the Barthel Index and Functional Independence Measure to determine if aforementioned patients are able to move on to the next stage of recovery. An EEG sub-band PSD dataset is used to develop an artificial neural network predictive model to classify the post-stroke patients into their respective recovery stages. Due to the small dissimilarity in intensity of certain sub-bands, classification between intermediate and advanced stages is predicted to be more difficult. At the end of the study, an ANN predictive model is developed with a satisfactory performance of 74.1%. When comparing classification results with an actual physiotherapist, an agreement rate of 58. 22% is achieved. This research contributes to helping medical professionals in evaluating the recovery progress of a post-stroke patient, therefore easing the rehabilitation process of a patient towards full recovery

Actions (login required)

View Item