The Assessment of Upper-Limb Spasticity Based on a Multi-Layer Process Using a Portable Measurement System

• Published in: IEEE transactions on neural systems and rehabilitation engineering Vol. 29; pp. 2242 - 2251
• Main Authors: Wang, Chen, Peng, Liang, Hou, Zeng-Guang, Zhang, Pu
• Format: Journal Article
• Language: English
• Published: New York IEEE 2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Abnormalities; Biomechanics; Damping; Elbow; Elbow (anatomy); Electromyography; ensemble empirical mode decomposition (EEMD); Feature extraction; Genetic algorithms; Immune system; Kinematics; Mechanical impedance; modified genetic algorithm; Monolayers; Motor neurons; multi-layer fusion; Multilayers; Muscles; Neurological diseases; Patients; portable assessment device; Reflexes; Regression models; Spasticity; Spasticity quantification; Stiffness; Stretch reflex; Stroke; Stroke (medical condition); Support vector machines; Viscoelasticity
• ISSN: 1534-4320; 1558-0210; 1558-0210
• DOI: 10.1109/TNSRE.2021.3121780

Spasticity is a common disabling complication caused by the upper motor neurons dysfunction following neurological diseases such as stroke. Currently, the assessment of the spastic hypertonia triggered by stretch reflexes is manually performed by clinicians using perception-based clinical scales, however, their reliability is still questionable due to the inter-rater and intra-rater variability. In order to objectively quantify the complex spasticity phenomenon in post-stroke patients, this study proposed a multi-layer assessment system based on a novel measurement device. The exoskeletal device was developed to synchronously record the kinematic, biomechanical and electrophysiological information in sixteen spastic patients and ten age-matched healthy subjects, while the spastic limb was stretched at low, moderate and high velocities. The mechanical impedance of the elbow joint was identified using a modified genetic algorithm to quantify the alterations in viscoelastic properties underlying pathological resistance. Simultaneously, the time-frequency features were extracted from the surface electromyography (sEMG) signals to reveal the neurophysiological mechanisms of the spastic muscles. By concatenating these single-layer decisions, a support vector regression (SVR)-based fusion model was developed to generate a more comprehensive quantification of spasticity severity. Experimental results demonstrated that the stiffness and damping components of the spastic arm significantly deviated from the nonspastic baseline, and strong correlations were observed between the proposed spasticity assessment and the severity level measured by clinical scales (<inline-formula> <tex-math notation="LaTeX">{R = {0.86},\;{P} = {1.67}{e} - {5}} </tex-math></inline-formula>), as well as the tonic stretch reflex threshold (TSRT) value (<inline-formula> <tex-math notation="LaTeX">{R = - {0.89},\;{P} = {3.54}{e} - {6}} </tex-math></inline-formula>). These promising results suggest that the proposed assessment system holds great potential to support the clinical diagnosis of motor abnormalities in spastic patients, and ultimately enables optimal adjustment of treatment protocols.