Identification of Motor Impairments Using Movement Distributions

The identification of motor impairments in human neuropathologies is daunting and often restricted to using isometric tests, identifying a single impairment, or investigating a single joint. Here, we present and evaluate our preliminary work in employing a novel, model-based framework for identification and quantification of multiple simultaneous impairments from multijoint upper extremity movement data in the form of a kinematic distribution. We use multiple synthetic models of idealized motor impairments to understand if and how it might be possible to differentiate between them based on the changes they induce in these movement distributions. We simulate these models in a variety of combinations and severities and use standard regression techniques to attempt to recover the “pathoparameters” defining them. We show that, given a well-chosen set of example distributions in a lookup table, one has the ability to identify pathoparameters and differentiate between impairment types without confusion.