The ability to reach, grasp, transport, and release objects is central to activities of daily living, such as feeding and grooming. However, children with cerebral palsy (CP) often have difficulty with these tasks, limiting their independence. While it is challenging to characterize and quantify specific upper limb movement disorders in CP, it is essential for identifying the underlying neural correlates and etiology, assessing movement disorder subtypes, i.e. spasticity, dystonia, and ataxia, which affect treatment selection, and measuring treatment outcomes. Current methods for measuring upper limb motion deficits are based predominantly on subjective, observational assessments. Thus, we have proposed three-dimensional motion analysis of the upper limbs during a Reach & Grasp Cycle to address the need for a standardized protocol for analysis of upper limb motion. The Reach & Grasp Cycle is a sequence of tasks that incorporates all major joints of the upper limb and simulates a functional task that is feasible yet challenging for individuals with CP.

Using a biomechanical model of the trunk and upper limbs, we calculated three-dimensional joint kinematics and temporal-spatial parameters for 30 typically developing (TD) children and 25 children with CP and upper limb involvement, ages 5-18 years, using an optoelectric motion analysis system. Consistent normative data and clinically significant differences in joint motions and temporal-spatial parameters between the CP and TD children suggest the Reach & Grasp Cycle is a repeatable protocol for objective and quantitative clinical evaluation of functional upper limb motor performance.

Next, we derived a single score of upper limb pathology from upper limb kinematics called the Pediatric Upper Limb Motion Index (PULMI). The root-meansquare difference was calculated between the data of each child with CP and the average from the TD population for eight kinematic variables over the Reach & Grasp Cycle. The raw value was then scaled such that a PULMI score ≥100 indicated the absence of upper limb pathology, and every 10 points below 100 corresponded to one standard deviation away from the TD PULMI mean. The PULMI was significantly different between the TD children and children with CP (Wilcoxon Z=-5.06, p<.0001), and between children with spastic CP and dyskinetic CP (Z=-2.47, p<.0135). There was a strong negative correlation between the PULMI and the standard Manual Ability Classification System for all children with CP (Spearman’s rho=-.78, p<.0001), indicating good validity of the PULMI. In addition, four key temporalspatial parameters (movement time, index of curvature during reach, ratio of the peak velocity of the transport and reach phases of the Reach & Grasp Cycle, and total number of movement units) revealed differences in movement patterns between CP and TD children. Furthermore, a multi-variable logistic regression of these temporalspatial parameters was derived which correctly predicted 19 of 22, or 86%, of movement disorder sub-types (spastic versus dyskinetic CP).

This research describes a pediatric upper limb motion index (PULMI) for children with cerebral palsy (CP) that provides information regarding the quality of upper limb motion during a functional sequence of tasks. The PULMI, calculated from upper limb kinematics, and key temporal-spatial parameters of the Reach & Grasp Cycle offer a quantitative approach to analyzing the quality of upper limb function in children with CP and identifying specific types of movement deficits. It is suggested for use in both research and clinical applications.