Robot-Assisted Locomotion Training after Spinal Cord Injury: Comparison of Rodent Stepping in Virtual and Physical Treadmill Environments

We are developing robotic devices for locomotion training after spinal cord injury. In this paper we compare two approaches to controlling and quantifying bipedal stepping of spinal rats with robots. In the first approach, the rats stepped on a physical treadmill with robot arms attached to their lower shanks. In the second, the rats stepped on a virtual treadmill generated by the robots. The rats could step on the virtual treadmill, but stepping was more consistent, step height greater and interlimb coordination improved on the physical treadmill. Implications for the role of sensory input in the control of locomotion and the design robotic of step trainers are discussed. 1 Introduction In the U.S. alone, over 10,000 people experience a traumatic spinal cord injury each year, and over 200,000 people with spinal cord injury are alive 1. Paralysis of the legs is a common consequence of spinal cord injury, resulting in loss of walking ability. Recently a new approach to locomotion rehabilitation called " body weight supported treadmill training " has shown promise 2-13. The technique involves suspending a spinal cord injured subject in a harness above a treadmill and manually assisting movement of the legs in a walking pattern. The goal of this technique is to enhance residual locomotor control circuitry that resides in the spinal cord. It is hypothesized that by providing appropriate sensory input (i.e. that associated with the force, position, and touch sensors that remain in the legs) in a repetitive manner, the spinal cord will learn to generate motor output appropriate for stepping.