BACKGROUND: Walking impairment is one of the most debilitating symptoms of multiple sclerosis (MS). A better understanding of brain mechanisms underlying successful gait training could help to improve development of targeted therapy. We therefore investigated changes in brain activation associated with improvements in walking function after rhythmic-cued gait training. METHODS: Thirty-one people with MS (pwMS
median EDSS = 2.5, range:2.0-5.0) and 17 age- and sex-matched healthy controls (HC) completed behavioural and MRI assessments at baseline and post-intervention (four weeks after baseline). All included pwMS received a four-week actual and/or imagined gait training with rhythmic-auditory cueing, while HC received no intervention. All participants performed a bipedal ankle plantar- and dorsiflexion and a corresponding motor-imagery task during fMRI. PwMS displaying a >
5 % walking distance increase in the 2-Minute Walk Test (2MWT) from baseline to post-intervention were defined as responders. RESULTS: Responders did not differ from non-responders in terms of demographics, clinical variables, and walking function at baseline. Responders, non-responders, and HC showed similar movement-related brain activation at baseline. At post-intervention, responders showed decreased brain activation within the premotor cortex, precuneus, and middle frontal gyrus during the movement task. Stronger decreases within these areas were associated with higher walking function improvements in all pwMS after controlling for potential confounders. No association was observed between walking function and motor imagery-related brain activation changes. CONCLUSION: Improved walking function after rhythmic-cued gait training was associated with reduced brain activation in motor planning and attention areas. This suggests a more efficient recruitment of areas subserving motor function after successful training.