Optimized Deep Learning Models Using Particle Swarm Intelligence for MindMend, Stroke Rehabilitation System

Stroke is a leading cause of long-term disability, significantly affecting patients' motor skills and daily activities. Traditional rehabilitation methods, while beneficial, often lack the precision and adaptability required for optimal recovery. This paper explores the integration of deep learning models optimized with Particle Swarm Optimization (PSO) to enhance stroke rehabilitation outcomes using brain-computer interface (BCI) technology. We employed a dataset from the BCI Competition IV, which includes EEG data from multiple participants engaged in motor imagery tasks. Various deep learning models, including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), EEGNet, and Multi-layer Perceptrons (MLP), were optimized using PSO to improve classification accuracy. The results demonstrate that PSO significantly enhances the performance of these models, providing a robust framework for developing advanced rehabilitation systems. This approach not only improves the accuracy of motor imagery classification but also offers a personalized rehabilitation experience for stroke patients,