Multimodal evaluation of gait alterations in persons with multiple sclerosis

Gait impairment is a frequent manifestation of multiple sclerosis and is of the utmost functional importance for those who live with this chronic inflammatory neurological condition. It is also a useful clinical outcome measure, usually evaluated on the basis of walking speed measured on a short distance.In this work, our first hypothesis is that walking speed is a construct significantly influenced by several confounders. Through the use of conventional methods to test gait, we successively address the importance of the distance (and hence locomotor fatigability, first on 100 and next on 500 metres), acceleration capacity and type of walk instructed to the subject. We show that the Timed 25 foot walk test suffer from several shortcomings related to each of these factors. We demonstrate that these are differentially affected in persons with multiple sclerosis as compared to healthy subjects, representing potential individual outcome measures themselves. Next, our second hypothesis is that walking speed is not the only feature characterizing the gait of persons with multiple sclerosis. We review the different available gait analysis technologies, their application in multiple sclerosis and create a new gait analysis system adapted to our needs. After technical validation, we design 26 gait features in order to capture other dimensions of walk than its speed, such as ataxia. We define those using factorial analysis. Finally, we use this system to explore the variance of gait in a population of healthy subjects and persons with multiple sclerosis. Using a mixed model analysis, we show that while walking speed is the main contributing factor to gait variance in such populations, other dimensions significantly come into play and should be considered in order to fully characterize ambulation in multiple sclerosis.