THERAPEUTIC EFFECTS OF CONVOLUTIONAL NEURAL NETWORK-BASED REHABILITATION PLATFORM ON ACTIVITY DAILY OF LIVING, PULMONARY ENDURANCE, BALANCE FUNCTION, AND GAIT FUNCTION IN INDIVIDUALS WITH CHRONIC STROKE

Patients with chronic hemiplegic stroke undergo traditional physiotherapy (TPT) to improve walking ability, activity of daily living (ADL), balance, and pulmonary endurance. To mitigate these disabilities, we technologically advanced a convolutional neural network (CNN)-based rehabilitation application (CRA) that provides no physical therapy stress or contact to the therapist and offers sufficient repetition. The CRA evaluates patients’ balance, walking ability, ADL, and cardiopulmonary function like a suitable beneficial intervention, such as TPT. We compared the effects of CRA and TPT on the Berg Balance Scale (BBS), Rating of Perceived Exertion Scale, 6-min walking test, Functional Ambulation Category (FAC), and modified Barthel Index (MBI) in patients with stroke. A sample of 30 participants with hemiplegic stroke was randomized into either the CRA or TPT intervention group. CRA and TPT were managed for the participants, with each intervention session lasting 30 min and allocated three per week for five weeks. Analysis of variance was accomplished, and the level of significance was set at [Formula: see text] 0.05. The ANOVA showed that CRA showed effects on the 6-min walking test result, BBS score, and MBI compared with TPT. Our novel results suggest that CRA can expand pulmonary endurance, balance, and ADL more effectively than TPT in patients with stroke.