"How Stiff Is Your Arm?"... Robot Enables Quantitative Diagnosis of Spasticity

Identifying and Correcting the Root Cause of Measurement Reliability Issues in Spasticity Assessment
UNIST Professor Sanghoon Kang's Team Publishes in Top 3% Rehabilitation Journal,
Paving the Way for Improved Rehabilitation in Stroke and Occupational Injury Patients

The method of diagnosing upper limb spasticity, which has traditionally relied on a physician’s tactile sense to assess arm stiffness, is expected to change.

This is because a domestic research team has developed a robotic technology that enables the degree of spasticity to be diagnosed with more precise numerical values.

The team led by Professor Sanghoon Kang from the Department of Mechanical Engineering at UNIST has newly developed a technique that quantifies spasticity by applying a subtle force to the patient’s arm and measuring its movement response. The system is designed so that even non-experts can make quantitative assessments within a few minutes, which is expected to help in designing personalized rehabilitation therapies and establishing criteria for industrial accident compensation.

Professor Sanghoon Kang, UNIST. Photo by UNIST

The research team validated this technology using a two-degree-of-freedom direct-drive robot. According to the study, even in direct-drive structures, there remained small but significant joint friction that affected measurement results. This demonstrates that even in representative rehabilitation robot models known for low friction, such as MIT’s Manus, there may be factors that reduce measurement reliability.

The team also experimentally confirmed for the first time that such residual friction is a major cause of responses that appear nonlinear, similar to those of the human arm. Previous research attributed low linearity?and consequently, low reliability?to the inherent nonlinearity of the human arm.

The team applied the Internal Model Based Impedance Control (IMBIC) strategy and was able to compensate for nearly 100% of the robot system’s residual nonlinear friction. Their experiments demonstrated that this allowed the arm’s movement to behave linearly, resulting in high reliability.

Researcher Sungil Hwang explained, “Existing robot-based spasticity measurement technologies have not been widely used due to issues of reliability and nonlinearity, but this study identified that the root cause of these issues is not the human arm but the residual friction inside the robot system. By correcting for this, we have been able to dramatically improve the reliability and accuracy of spasticity measurements.”

Researcher Sungil Hwang. Provided by UNIST

Muscle spasticity is a representative upper limb motor disorder caused by conditions such as stroke or occupational nerve injury. Traditionally, medical personnel have assessed the degree of spasticity by manually moving the patient’s arm and relying on their sense of touch. However, this method has significant variability depending on the examiner’s skill and makes it difficult to distinguish movement characteristics between joints or in different directions.

Method for Measuring Upper Limb Spasticity and Reliability Improvement Results Based on This Study.

Professor Sanghoon Kang stated, “Being able to quantify and track the patient’s condition will help in designing rehabilitation treatments and establishing criteria for industrial accident compensation. We also plan to increase the potential for clinical application through collaboration with the Ulsan Public Hospital for Occupational Accidents, which is scheduled to open in 2026.” Professor Kang is also currently serving as an adjunct professor at the University of Maryland School of Medicine in the United States.

This study was led by Sungil Hwang, a researcher at UNIST, as the first author, with Dr. Hyunah Kang as a co-author.

The research was supported by the Ministry of Science and ICT’s National Research Foundation Pioneer Project for Future Convergence Technology, the Pan-Ministry Medical Device Development Project Group, and the National Rehabilitation Center’s Rehabilitation Robot Translational Research Service.

The results were published on March 26 in the IEEE Transactions on Neural Systems and Rehabilitation Engineering, a journal ranked in the top 3% in the field of rehabilitation medicine.