Electronic Thesis and Dissertation Repository


Master of Engineering Science


Biomedical Engineering


Dr. Ana Luisa Trejos


Musculoskeletal conditions are the most common cause of severe long-term pain and physical disability, accounting for the highest disability costs of about $17 billion yearly. To provide better rehabilitation tactics, the knowledge gap between injuries and their healing mechanisms needs to be addressed. The use of electromyography (EMG) is very popular in detecting neuromuscular diseases or nerve lesions; however, there is limited knowledge available for quantifying healing patterns of EMG in orthopedic patients who have injured their joints, muscles, or bones. In order to quantify the progress of orthopedic patients and assess their neuromuscular health and muscle synergy patterns, EMG signals were collected from 16 healthy individuals and 15 injured patients as they underwent rehabilitation. Subjects performed a series of standard motions such as flexion–extension of elbow and pronation–supination of the arm. Different metrics were used to process and analyze the EMG data collected using MATLAB. The metrics were as follows: root mean square, average rectified signal, mean spike amplitude, zero crossings, median power frequency, and mean power frequency. A normal range across the muscle groups has been identified and to which the patient population was compared. This comparison showed statistically significant differences in the magnitudes of muscle recruitment and activation between the two groups. Furthermore, a comparison within the patient population at the beginning of their therapy versus at the end of the therapy was conducted. Statistical differences arose in this second analysis, further proving that patients’ signals tend to change and showing trends closer to those of the healthy population. The time domain metrics showed the greatest significant differences between the groups, specifically the root mean square and average rectified signal. This analysis was successful in showing a general trend of increased mean in the patient population compared to healthy individuals. The frequency domain metrics did not show statistical significance. The work presented successfully used several EMG metrics in order to distinguish an injured person from a healthy person and to determine if an injured patient is healing. Additionally, a database of EMG signals to be fed into the control system of the mechatronics rehabilitative brace was created. This work has advanced the use of EMG beyond the scope of nerve damage. The experiments conducted showed that EMG could be used as method to assess musculoskeletal health.