Automated ECG and EEG signal interpretation using Machine Learning Accuracy, Sensitivity, and Specificity Analysis for Early Diagnosis of Neurological Disorders

The early diagnosis of neurological disorders is often hindered by subtle and overlapping electrophysiological abnormalities present in electrocardiogram (ECG) and electroencephalogram (EEG) signals. The application of machine learning (ML) and deep learning (DL) algorithms offers a promising avenue for automated, accurate, and scalable interpretation of such biosignals. This study aimed to quantitatively evaluate the performance of automated ECG and EEG interpretation using ML and DL models, emphasizing accuracy, sensitivity, and specificity for early neurological disorder detection. A retrospective analytical design was adopted using two large open-source datasets: PTB-XL for ECG signals and CHB-MIT for EEG recordings, encompassing over 50,000 labeled samples. Signal preprocessing included artifact removal, normalization, and segmentation into temporal frames. Feature extraction was performed using time–frequency domain and statistical measures. Five algorithms—Support Vector Machine (SVM), Random Forest (RF), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and a hybrid CNN–LSTM model—were evaluated alongside a Transformer-based architecture for EEG data. Model performance was assessed using fivefold cross-validation, and multimodal fusion was implemented through decision-level integration. Deep learning models significantly outperformed traditional ML methods. The CNN–LSTM hybrid achieved an accuracy of 97.5%, sensitivity of 96.8%, and specificity of 98.1% in ECG classification. The Transformer model for EEG analysis attained 97.2% accuracy and an AUC of 0.99. Combining both modalities through fusion yielded the highest diagnostic accuracy of 98.5%, demonstrating that multimodal signal integration enhances early detection of neurological abnormalities. The study confirms the efficacy of deep learning, particularly CNN–LSTM and Transformer models, in automated biosignal interpretation for early neurological disorder diagnosis. The results highlight the diagnostic potential of multimodal ECG–EEG analysis and establish a reproducible framework for evaluating ML-based medical diagnostic systems. Future research should focus on clinical validation, interpretability enhancement, and integration into real-time, wearable healthcare platforms.