Novel spatiotemporal analysis of gait changes in body weight supported treadmill trained rats following cervical spinal cord injury

BACKGROUND: Common gait measures such as stride length, cycle time, and step height are not independent variables, but different aspects of the same multidimensional step. This complicates comparisons between experimental groups. Here we present a novel multidimensional gait analysis method and use this method to assess the ability of body weight supported treadmill training (BWSTT) to improve rodent stepping after spinal cord injury (SCI). METHODS: In lieu of reducing a step to a collection of gait measures and comparing the means of several of these, we developed a multidimensional analysis technique that compares the step as a whole. While in a passive robotic gait training device, the pre-injury hindlimb stepping of 108 rats was recorded while they walked in a quadrupedal posture at 8 cm/s. Following a C4/5 over-hemisection spinal cord injury the weekly changes in stepping were tracked for 17 untrained and 10 BWSTT animals for 7 weeks. The performance of trained rats was recorded during training with BWS, as well as at the end of the training week without BWS. An additional six uninjured rats were trained for 5 weeks. RESULTS: Our novel multidimensional analysis shows that stepping is asymmetrically altered 1 week after SCI. The differences in stepping change over the following weeks, with the less impaired left hindlimb deviating further away from pre-injury than the more impaired right hindlimb. Uninjured rats do not significantly alter their stepping over 5 weeks. BWSTT improves the stepping of the right hindlimb, but only when the BWS is active. If the BWS is not present, the performance of trained animals is worse than untrained rats. The left hindlimb performance of BWSTT rats is worse than untrained rats, during both training sessions and weekly assessments. CONCLUSIONS: We feel that our novel multidimensional analysis is a more appropriate method to address the inter-dependencies of gait measures. Untrained rats exhibit both initial impairments as well as the development of compensatory techniques. BWSTT does not improve this spontaneous recovery, but exacerbates it, particularly in the less impaired left hindlimb.