engleski

Contribution to differential electromyographic diagnostics of low back pain and radiculopathy

Low back pain is and will continue to be one of the leading causes of disability, absence from work and thus costs for individuals and societies. It can be expected that lifetime prevalence of the low back pain will continue to increase with increasing sedentary lifestyle across the globe in developing and already developed countries. Diagnosing low back pain patients or their differentiation from healthy subjects is thus an interesting task. One of the means for diagnostics arise from properties of surface EMG signals which provide information about complex changes occurring within low back muscles during fatiguing contractions. So far, studies of the surface EMG based classification models of low back pain patients have been directed only to their differentiation from healthy subjects. This thesis deepens the knowledge on surface EMG based classification models and introduces differentiation of low back pain patients with radiculopathy from nonspecific chronic low back pain patients and healthy subjects. The measurement protocol is simplified, and number of classification features is reduced only to one. Surface EMG signals were measured above low back muscles: m. erector spinae at L1-L2 interspace and m. erector spinae at L4-L5 interspace. A variant of the Roman chair was used to perform static contractions and subject’s own upper body weight was used to induce muscle fatigue in low back muscles. The time-frequency method called Hilbert-Huang transformation (HHT) was utilized to estimate power spectrum from recorded surface EMG signals above lower back. As a descriptor of surface EMG spectral changes, the regression line of slope of the median frequency of the power spectrum was calculated to form classification feature for the decision tree classification. Two splitting criterions were evaluated, Gini diversity index and Maximum deviance reduction. The results show that regression line slope of the median frequency derived from HHT based power spectrum estimate is significantly different between low back pain patients with radiculopathy and other two groups, being chronic low back pain patients and control subjects. There was no significant difference between chronic low back pain patients and control subjects. The proposed classification feature and model enabled differential electromyographic diagnostics of low back pain and radiculopathy.

biomedical signal processing ; classification ; electromyography ; Hilbert–Huang transform ; low back pain ; radiculopathy

nije evidentirano