Machine Learning-Based Stroke Prediction and Heartbeat Classification Using ECG and PPG Signals in MRI Environments

Timely stroke detection is essential for effective treatment and improved patient outcomes. Traditional methods like MRI and CT scans, while informative, are often expensive and time-consuming, limiting their use in real-time applications. This study explores a machine learning-driven approach to enhance stroke prediction using ECG and PPG signals, which are more suitable for continuous monitoring. Due to signal distortions caused by electromagnetic interference in MRI settings, a self-supervised learning framework is utilized to learn robust features from unlabelled ECG data. A hybrid deep learning model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks is proposed to analyze real-time multi-modal signals. This approach improves the accuracy of heartbeat classification and stroke risk assessment, offering a practical solution for wearable health monitoring and early stroke prevention in dynamic clinical environments. The results indicate strong potential for real-time healthcare integration.