Supported by NIH 2R01HD040289-15A1
Cerebellar damage impairs movement coordination and adaptive motor learning abilities, making actions like reaching inaccurate and very difficult to control. There are currently no medications that systematically improve cerebellar movement incoordination or `ataxia', making rehabilitation therapy the main treatment option. Yet, people with ataxia are notoriously difficult to treat with physical therapy, likely due to the fact that they are limited in their ability to learn new movement patterns. The studies proposed here address this challenge through the development of a staged, individualized approach to understanding if distinct interventions will be helpful to different people. We will use mathematical modeling of patient-specific movement deficits, alternative learning mechanisms, and robotic control of reaching to test whether we can systematically reduce arm ataxia. In Aim 1, we will determine if reinforcement based motor learning can be used as an alternative strategy in people with cerebellar damage. Our preliminary data show that reinforcement based motor learning is much more effective than adaptive motor learning for reaching movements in people with ataxia. We will test whether this can be used to change complex elements of 3D reaching that are more clinically meaningful for people with ataxia. In Aim 2, we will test whether long-term reinforcement-based training can reduce ataxia and improve 3D reaching performance. Here we will test whether people who show learning via reinforcement on a single day in Aim 1. We will determine if they benefit more from a 2-week course of virtual reality-based reinforcement reach training compared to 2 weeks of standard reaching practice. We will study how training transfers to natural reaching movements (i.e. those outside of the virtual reality environment) and clinical rating scales of arm ataxia and function. In Aim 3, we will develop compensatory robotic assistance for people with ataxia based on individualized models of cerebellar function. This is essential for individuals with the most severe ataxia and learning problems who do not learn at all in Aim 1. In sum, this proposal provides a scientific framework for determining an individual's motor learning potential for rehabilitation versus the need for intelligent compensatory robotic assistance. Our overarching goal is to provide a foundation for devising and choosing new rehabilitation strategies.
In this project, we will determine whether people with cerebellar damage can 1) learn to improve reaching on a short time scale using a motor learning mechanism that does not rely on the cerebellum, 2) improve long term reaching performance from weeks of training and 3) benefit from individualized robotic assistance for reaching. We hypothesize that many people with mild to moderate cerebellar ataxia can learn on short and long time scales using a reinforcement motor learning mechanism, and this may benefit them over weeks of training. We think that individuals with the most severe ataxia might benefit most from robotic assistance that is tailored to help their specific pattern of incoordination. We will use a staged testing strategy to determine which of these might be most useful for individuals with ataxia.