Relationship between gait variables and domains of neurologic dysfunction in multiple sclerosis using six-minute walk test.

Most multiple sclerosis (MS) patients eventually suffer from mobility impairment, and thus it is critical that walking disability in MS be accurately assessed. The six-minute walk test (6MWT), a reliable MS measure, is traditionally used to determine the distance covered in six minutes using a standard protocol. With the availability of body sensor networks (BSNs), researchers are interested in leveraging BSN data for finding new gait assessment anchors for improved separability performance. Further, current methods for gait assessments are insufficient since assessments are absolute, performed by comparing outcomes to the statistical norms established from diverse patient data despite natural inter-patient variability. For example, walking at a slower pace may be a personal trait rather than a fatigue-induced impact in MS. Thus instead of using actual values, changes in gait features during a 6MWT can be used to draw clinical conclusions and remove inter-subject variability. We use 6MWT gait time series BSN data of MS subjects to determine changes in their gait speed and gait cycle length variance over six minutes. We also determine the relationships between these variables and three subjective assessments, namely functional systems scores, modified fatigue impact scale, and multiple sclerosis walking scale. These relationships are important in order to validate features and make them clinically meaningful for a doctor. In MS, feelings of fatigue can occur more often, and have a greater impact on a subject's daily life. The results show that MS patients who reported high motor-related fatigue slowed down and had high gait cycle length variance in the last minute of the test, in contrast to those on whom fatigue had minimal impact, enhancing the separability between MS subjects of varied disabilities based on their scores. It is also observed that FSS components related to coordination, balance, and physical activities show strong correlation with gait parameters.