Sensor-based Quantitative Assessment of Sensorimotor Deficits in Multiple Sclerosis

Multiple Sclerosis (MS) is a chronic neurological disease characterised by heterogeneous and progressive impairments that can affect gait, balance, motor control, and neuromuscular function. Conventional clinical assessments provide essential information about overall disability, but may have limited sensitivity to subtle changes in specific domains of motor function. Quantitative sensing technologies may complement clinical assessment by providing objective measures of movement and motor performance. However, the clinical relevance of sensor-derived outcomes depends on the specific motor construct assessed and on the measurement properties established for each outcome. The aim of this thesis was to investigate the use of sensing technologies for the quantitative assessment of selected dimensions of motor and neuromuscular function in people with MS, with a particular focus on approaches that may complement conventional clinical evaluation. Three studies investigated complementary aspects of motor performance using different sensing modalities and experimental tasks. The first study investigated a video-based, markerless approach for the quantitative assessment of gait characteristics in people with MS. Two-dimensional pose estimation was used to derive lower-limb kinematic measures from RGB video recordings. The study demonstrated the feasibility of extracting quantitative gait-related measures from video data and explored their associations with clinical disability, providing evidence for the potential of accessible, non-contact approaches to characterise motor impairment. The second study investigated an instrumented eyes-closed forward-and-backward walking task using a visual-inertial tracking sensor. A trajectory-derived measure of normalised lateral trajectory excursion (NLE) was introduced as a geometric descriptor of movement, together with measures of whole-trial average speed and directional deviation. NLE was associated with disability severity and showed within-session repeatability, suggesting that challenging walking tasks performed without visual feedback may reveal clinically relevant aspects of motor performance that are not fully captured by conventional assessments. Importantly, NLE is interpreted as a descriptor of trajectory excursion rather than as a direct measure of dynamic stability or vestibular function. The third study investigated force control during dynamic force tracking in older people with MS. Force variability and mean absolute yank were used to characterise complementary aspects of force regulation during a dynamic isometric tracking task. At baseline, older people with MS exhibited greater force variability than age-matched healthy controls, whereas mean absolute yank did not differ significantly between groups. Following a 24-week power training intervention, maximal force improved in the power training group, but no differential changes in force variability or mean absolute yank were observed compared with usual care. These findings suggest that maximal force production and dynamic force control represent distinct dimensions of neuromuscular function. Overall, the findings of this thesis demonstrate the potential of sensing technologies to provide objective and quantitative information about specific dimensions of motor function in people with MS. The results support their use as complementary tools to established clinical assessments rather than as replacements for conventional measures. At the same time, the studies highlight that feasibility, validity, reliability, clinical association, group discrimination, responsiveness, and clinical utility represent distinct measurement properties that must be evaluated separately for each sensor-derived outcome. Further longitudinal and validation studies are required to determine whether these measures can reliably detect clinically meaningful changes over time and ultimately support the monitoring of motor function and disease progression in MS.