Hybrid Quantum Convolutional Neural Network for Myopathy Detection from EMG Signals

Abstract

Muscle weakening and degeneration are hallmarks of myopathy, a neuromuscular ailment that is difficult to identify since its symptoms can be confused with those of other disorders including neuropathy. Despite being a key diagnostic tool, electromyography (EMG) interpretation is frequently subjective and necessitates expert analysis. New developments in quantum computing allow for hybrid quantum-classical models that improve the processing of biomedical signals. This study introduces a system for automated myopathy identification using EMG signals that is based on Quantum Convolutional Neural Networks (QCNNs). Combining the durability of deep learning with quantum parallelism, the suggested design uses a conventional neural network for classification and quantum circuits for effective feature extraction. The QCNN outperforms both classical CNN and purely quantum baselines in accuracy, precision, recall, F1-score, and ROC-AUC, according to experimental results, which demonstrate that it reaches 96.99% accuracy. The viability of real-Time clinical implementation was also assessed by analyzing inference time and qubit use. Although encouraging, issues with scalability-due to qubit constraints-model interpretability for clinical trust, and generalization to a variety of patient populations still exist. In the era of Noisy Intermediate-Scale Quantum (NISQ), this work shows how QCNNs can be used for medical diagnostics. It also lays the groundwork for further studies into multi-class classification, cross-modal learning, and deployment on actual quantum hardware. © 2025 IEEE.