Neural correlates of gait variability in people with multiple sclerosis with fall history

Background and purpose: The aim was to investigate the association between step time variability and related brain structures in accordance with fall status in people with multiple sclerosis (PwMS). Methods: The study included 225 PwMS. Whole-brain magnetic resonance imaging was performed with a high-resolution 3.0 T magnetic resonance scanner in addition to volumetric analysis based on 3D T1-weighted images using the FreeSurfer image analysis suite. Step time variability was measured with an electronic walkway. Participants were defined as ‘fallers’ (at least two falls during the previous year) and ‘non-fallers’. Results: In all, 105 PwMS were defined as fallers and had a greater step time variability compared to non-fallers [5.6% (SD = 3.4) vs. 3.4% (SD = 1.5); P = 0.001]. MS fallers exhibited a reduced volume in the left caudate and both cerebellum hemispheres compared to non-fallers. On using a linear regression analysis no association was found between gait variability and related brain structures in the total cohort and the non-fallers group. However, the analysis found an association between the left hippocampus and left putamen volumes with step time variability in the faller group: P = 0.031, 0.048, respectively, controlling for total cranial volume, walking speed, disability, age and gender. Nevertheless, according to the hierarchical regression model, the contribution of these brain measures to predict gait variability was relatively small compared to walking speed. Conclusions: An association between low left hippocampal, putamen volumes and step time variability was found in PwMS with a history of falls, suggesting that brain structural characteristics may be related to falls and increased gait variability in PwMS.