Leveraging Surface Electromyography for Accurate Low Back Pain Classification Using Hybrid Deep Learning

Low Back Pain (LBP) is one among the familiar musculoskeletal disorders and techniques that are able to identify LBP patients from healthy people by utilizing balance performance data which is very rare. Surface Electromyography (sEMG) is a renowned method that allows non-invasive recordings of muscle activities. Increasing interest and the developments in Artificial Intelligence (AI) and Machine Learning (ML) techniques, with their application in various fields, are also reflected in several studies which manage LBP issues. The myoelectric action will reflect the neuromuscular adaptation in LBP patients, therefore allowing co-activation strategies among muscles and more perceptions into motor control coordination. Therefore, this paper develops hybrid deep learning based accurate LBP classification (HDL-ALBPC) model using sEMG signals. The presented HDL-ALBPC model majorly focuses on the recognition and classification of LBP on sEMG. To accomplish this, the HDL-ALBPC model initially derives slime mold optimization (SMO) based feature selection technique to elect features from the extracted ones. For LBP classification, teaching and learning based optimization (TLBO) with deep neural network (DNN) method is utilized in this research. The optimal hyperparameter selection of the DNN method is carried out by the implementation of the TLBO algorithm. The experimental authentication of the HDL-ALBPC model is verified using a series of simulations and the comparison study demonstrates the betterment of the HDL-ALBPC model.