Investigating Muscle Electrophysiological Profiles after Cervical Spinal Cord Injury Through Surface Electromyography Clustering Analysis

Abstract

Objective:

Cervical spinal cord injuries (SCIs) result in significant neurological and functional impairments. Current clinical assessments, such as the International Standards for Neurological Classification of Spinal Cord Injury, provide essential diagnostic and prognostic insights but have limited sensitivity in detecting residual motor control. This study aims to investigate whether surface electromyography (sEMG) signals can reveal distinct electrophysiological profiles that complement clinical information, potentially enhancing the assessment of SCI.

Methods:

sEMG signals were recorded from 184 upper extremity muscle groups across 22 adult individuals with cervical SCI. Time and frequency domain features were extracted. Multiple clustering algorithms, including k-means, k-medoids, density-based spatial clustering of applications with noise, and hierarchical clustering, were applied to identify distinct sEMG profiles. Internal validation metrics (Silhouette scores) and resampling-based robustness assessments were used to confirm the reliability of the clusters. Identified clusters were evaluated for their associations with clinical variables, including neurological level of injury (NLI), American Spinal Injury Association Impairment Scale scores, myotome levels, manual muscle testing scores, and lower motor neuron injury status.

Results:

Distinct and reproducible clusters were identified, and significant associations were found between the sEMG clusters and clinical variables, particularly the myotome level and NLI. However, the clusters were not fully explained by clinical variables, indicating that sEMG may capture additional physiological nuances, such as residual motor pathways or compensatory mechanisms, that are not readily assessed through standard clinical evaluations.

Conclusions:

This study demonstrates that in individuals with cervical SCI, sEMG-based clustering identified distinct muscle electrophysiological profiles. These profiles are partially aligned with clinical variables. Yet the additional dimensions captured by sEMG may have the potential to enhance neurological assessments and improve the clinical management of SCI. These findings underscore the need for further research with larger and more diverse datasets to validate the clinical relevance of sEMG clusters and explore their implications for rehabilitation strategies.