Estimation of parameters to evaluate asymmetries in movement in patients affected by multiple sclerosis during activities of daily life

Multiple sclerosis (MS) is one of the most prevalent causes of neurological disability in young and middle-aged adults, presenting through diverse symptoms, including primarily motor disruptions such as fatigue, coordination issues and even paralysis. The latter impairments lead to a deterioration of quality of life, safety, autonomy. In everyday life upper limbs motor abilities play a crucial role in performing typical activities of domestic environments and the ability to complete specific tasks is synonymous with individual functional independence. In several studies, the assessment of upper limb motor functions relies on the use of clinical scales or conventional methods such as optoelectronic system. However, these methods present some limits as they are based on standard clinical tasks conducted in a laboratory setting, and above all, not fully reflecting daily-life activities performances. To overcome these limitations, wearable devices as magneto-inertial sensors (MIMUs) has gained increasing prominence for assessing motor performances of the upper limbs. The practicality and ease of use of MIMUs allows to move the analysis of movement from standard clinical settings toward conditions that closely resemble real life. However, more efforts should be made in refining protocols which, mostly, involve the use of MIMUs for activity recognition or, alternatively, assessing range of motion during highly specific and well-defined tasks. The aim of the thesis is to investigate the feasibility of an original protocol based on MIMUs for detecting mobility changes trough quantitative evaluations of one-time or periodic performances in various conditions during real-life activities in individual with MS. The analysis enrolled 33 subjects (7 males; 46 ± 12 years; EDSS: 1 - 6), equipped with three MIMUs positioned on the lower back, right and left wrist. The motor assessment involved the execution of six activities of daily living in a predetermined sequence (wash floor, transfer light, laundry, meal management, clean surface and climb stairs) in two sessions, before and after undergoing an individualized rehabilitation training. Rehabilitation training consist of four types, one more comprehensive which includes both muscle strengthening and functional exercises, while the others primarily involve either muscular strength training, cardio training, or a combination of both. For each activity in every session, activity segmentation was automatically performed from wrist-acceleration signals. Then, for each activity several parameters were extracted: Standard Deviation of acceleration peaks, Peaks per Second, Peak Ratio, Mean Peak Acceleration, Acceleration Per Second, Log Dimensionless Jerk, Bilateral Magnitude and Magnitude Ratio. Various comparison was conducted, such as, for same subject and parameter, comparison between right and left limb in different tasks before and after rehabilitation exercise, then, after setting an arbitrary threshold of improvement on inter-session variations, comparison was done between different subjects for the same task and parameter. Preliminary findings suggested that intervention based on functional exercises led to improvements in a greater number of activities, as expected. The proposed protocol has the potential to enable an ecological assessment of upper limb mobility during everyday activities using IMUs, providing a quantitative and clinician independent evaluation. Results generalization should be confirmed by increasing the number of participants.