Assessment of hindlimb locomotor strength in spinal cord transected rats through animal-robot contact force.

Robotic locomotor training devices have gained popularity in recent years, yet little has been reported regarding contact forces experienced by the subject performing automated locomotor training, particularly in animal models of neurological injury. The purpose of this study was to develop a means for acquiring contact forces between a robotic device and a rodent model of spinal cord injury through instrumentation of a robotic gait training device (the rat stepper) with miniature force/torque sensors. Sensors were placed at each interface between the robot arm and animal's hindlimb and underneath the stepping surface of both hindpaws (four sensors total). Twenty four female, Sprague-Dawley rats received mid-thoracic spinal cord transections as neonates and were included in the study. Of these 24 animals, training began for 18 animals at 21 days of age and continued for four weeks at five min/day, five days/week. The remaining six animals were untrained. Animal-robot contact forces were acquired for trained animals weekly and untrained animals every two weeks while stepping in the robotic device with both 60 and 90% of their body weight supported (BWS). Animals that received training significantly increased the number of weight supported steps over the four week training period. Analysis of raw contact forces revealed significant increases in forward swing and ground reaction forces during this time, and multiple aspects of animal-robot contact forces were significantly correlated with weight bearing stepping. However, when contact forces were normalized to animal body weight, these increasing trends were no longer present. Comparison of trained and untrained animals revealed significant differences in normalized ground reaction forces (both horizontal and vertical) and normalized forward swing force. Finally, both forward swing and ground reaction forces were significantly reduced at 90% BWS when compared to the 60% condition. These results suggest that measurement of animal-robot contact forces using the instrumented rat stepper can provide a sensitive and reliable measure of hindlimb locomotor strength and control of flexor and extensor muscle activity in neurologically impaired animals. Additionally, these measures may be useful as a means to quantify training intensity or dose-related functional outcomes of automated training.