Action likelihood to evaluate weakness, abnormal motor synergies, and spasticity in stroke survivors

The National Institute of Child Health and Human Development, part of the National Institutes of Health, awarded the University of Illinois $110,228 on December 3, 2026, under the Child Health and Human Development Extramural Research program (CFDA 93.865) to develop a markerless motion-capture tool for objectively evaluating motor impairments in stroke survivors.

The project, titled "Action Likelihood to Evaluate Weakness, Abnormal Motor Synergies, and Spasticity in Stroke Survivors," aims to create a scalable, objective assessment method that quantifies upper limb weakness, abnormal movement synergies, and spasticity from two minutes of free exploration movement. Current clinical evaluation methods—manual muscle testing and comprehensive assessments—are time-consuming and subjective. The research leverages kinematic probability distributions, termed action likelihoods, derived from markerless upper limb movement data to identify individualized signatures of motor impairment by comparing stroke survivor movement patterns to neurotypical participants and neuromusculoskeletal model predictions. The goal is to enable therapists to devote more time to therapy rather than assessment.

Work is performed in Chicago, Illinois. The period of performance runs through July 14, 2030. The award is classified as a Project Grant and is assistance funding, not a procurement contract.

Abstract

Upper limb motor impairments such as weakness, abnormal synergies and spasticity affect over two-thirds of stroke survivors, with one stroke occurring every 40 seconds in the united states. These deficits often persist long after the acute event, severely limiting individuals' ability to perform daily activities. To maximize recovery with personalized rehabilitation, clinicians rely on manual muscle testing and comprehensive clinical assessments. However, these evaluations are time-consuming and subjective. Attempts to streamline evaluations using electromyography or marker-based motion tracking are impractical for routine clinical use. There is a critical need for a scalable, objective tool that leverages markerless upper limb movement data to rapidly identify deficits and predict clinical scores so therapists can devote more time in therapy. Our long-term goal is to quantify weakness, abnormal synergies, and spasticity from just two minutes of free exploration movement. Previously, we have shown that the probability distributions of wrist position in a free exploration task where participants move based on their own preferences can uncover flexion synergy patterns in over half of stroke survivors and identify regions of weakness in all stroke survivors. Based on this encouraging preliminary evidence, our central hypothesis is that comparing a stroke survivor's portfolio of kinematic probability distributions, termed action likelihoods, to those of neurotypical participants will reveal individualized signatures of motor impairment. Our rationale is that each impairment alters the preferred movement patterns within a person's available range of motion in distinct, quantifiable ways. In aim 1, we compare action likelihoods of stroke survivors with weakness to those of neurotypical participants and to predictions from a neuromusculoskeletal model of weakness. In aim 2, we compare action likelihoods of individuals with abnormal synergies to those of neurotypical participants and to predictions from model of joint coupling. In aim 3, we examine spasticity during volitional movement by comparing high-frequency kinematic features in stroke survivors to those of neurotypical participants and to those generated by a computational model of spasticity. The expected outcome is a data-driven tool based on action likelihoods that can quantify upper limb impairments during free exploration and offer insight into underlying mechanisms of motor deficits. This work moves beyond predefined task performances and captures real-world variability and movement limitations. IT also establishes a novel framework for precision rehabilitation while generating a curated upper extremity movement data repository to support future research. Alongside the research project, the individualized training plan developed by the sponsor (james l. Patton, PHD, university of illinois chicago) and the CO-sponsor (r. James cotton, MD, PHD, northwestern university) will support the principal investigator in gaining skills necessary to become an impactful physician-scientist.