Presenter: Dinal Jayasekera
BME 4th year Ph.D. student
Dr. Wilson Ray Lab
Chronic low back pain (LBP) is a leading cause of suffering and disability worldwide. While the study of LBP has largely focused on the spine, many studies have demonstrated a restructuring of human brain architecture accompanying LBP and other chronic pain states. Brain neuroimaging presents a promising source for the discovery of noninvasive biomarkers that can improve the diagnosis, treatment, and long-term outcomes for chronic LBP. Graph theory metrics derived from resting-state functional connectivity (rsFC) measures could serve as potential noninvasive brain biomarkers for LBP. We trained a support vector machine using graph-theoretical features to classify LBP subjects from healthy controls (HC). Local graph metrics such as the degree centrality, clustering coefficient and betweenness centrality were found to be significant predictors of patient group while using a combination of Elastic Net and an optimal subset selection method (Enet-subset) method during feature selection. We achieved an average classification accuracy of 83.1% (p < 0.004) and AUC of 0.937 (p < 0.002), respectively. Similarly, we achieved a sensitivity and specificity of 87.0% and 79.7%. The classification results from this work suggest that graph matrices derived from rsFC can be used as biomarkers for LBP. In addition, they also prove that the Enet-subset feature selection method that used with this dataset helps with removing redundant variables.