Classification of Cervical Disc Herniation Disease using Muscle Fatigue based surface EMG signals by Artificial Neural Networks

Abstract

This study presents the classification of cervical disc herniation patients and healthy persons by using muscle fatigue information. Cervical disc herniation patients suffer from neck pain and muscle fatigue in the neck increases these aches. Neck pain is the most common pain type encountered after back pain. The discomforts that occur in the neck region affect the daily quality of life, so the number of researches done in this area is increasing. In this study surface Electromyography (EMG) signals were used to examine muscle fatigue. EMG signals were obtained from Trapezius and Sternocleidomastoid (SCM) muscles in the cervical region of 10 control subject and 10 cervical disc herniation patients. Surface EMG was preferred because it is a noninvasive method. In the first step of this study, EMG signals were filtered and adapted for analysis. In the second step, muscle fatigue was determined using Median and Mean frequency values obtained by Fourier Transform and Welch methods. Feature extraction was the third step which was performed by Short Time Fourier Transform (STFT), Discrete Wavelet Transform (DWT) and Autoregressive method (AR). Finally, Artificial Neural Network (ANN) was used for classification. Training and test data were created by using feature vectors to classify patients with ANN. According to the results, the superior feature extraction method was investigated on patient classification using muscle fatigue information. The best results were obtained by AR method with %99 classification accuracy. Also, the best results were obtained by DWT with %100 classification accuracy for SCM muscle. This study has contributed that AR and DWT are a suitable feature extraction methods for surface EMG signals by providing high accuracy classification with artificial intelligence methods for cervical disc herniation disease. Besides, it is shown that muscle fatigue distinguishes cervical disc herniation patients from healthy people.

Frekans Alanı Öznitelik Çıkarımına Dayalı Makine Öğrenme Teknikleri ile Trapezius Yüzey EMG Sinyallerini Kullanarak Servikal Disk Fıtığı Hastalığının Tahmini

“…To answer this question, an experimental study was performed. Detailed information about the experimental study can be found in [22], but the present study only addresses the resting state data recorded from the trapezius muscle. In this prospective study, sEMG data were collected by a physician and a technician in the neurology department of Selcuk College Faculty of Medicine using surface electrodes and a Neoropack Nihon Kohden EMG device in 10 CDH patients (8 males and 2 females, aged 17 to 67 years) and 10 healthy volunteers (4 males and 6 females, aged 19 to 48 years).…”

“…In the literature, some studies classify neuromuscular diseases [10][11][12], detect muscle activity [13][14], muscle fatigue [15][16], and classification of low back pain [17][18] and neck pain [19][20][21] using sEMG. To our knowledge, no study classifies CDH patients using surface EMG, except [22]. In addition, some studies draw attention to the upper trapezius muscle during semi-static activities that require repetitive movements of the upper extremity.…”