Innovative Techniques for Biomechanical Evaluation of Stroke Survivors: Combined fMRI-Gait Analysis Assessment and Fugl-Meyer Clinical Scores Estimation Through Wearable Sensors

Stroke is a leading cause among cerebrovascular injuries responsible for disabilities and social functioning difficulties. This pathology can really affect patients functionality, so that they might not be able to perform the daily simple activities like walking, eating and so on. So individual rehabilitation programs are designed to address subject-specific motor impairments and functional limitations, therefore it is of paramount importance to optimize the outcomes of rehabilitation on a subject-by-subject basis. What clinicians usually do to assess the effectiveness of rehabilitation programs is to use some clinical scales, which are based on the observation of subjects’ motor behavior, they are usually time consuming, expensive to perform and quite subjective, so therapists and clinicians often favor increasing the time devoted to therapy at the cost of not performing longitudinal assessments of motor abilities. Therefore the aim of this work is trying to provide an objective and fast methodology to evaluate stroke survivors performances: the first method is based on a combined functional magnetic resonance imaging (fMRI)-gait analysis study, where fMRI is a MRI procedure that measures brain activity by detecting associated changes in blood flow, and gait analysis is the motion analysis of the human gait; the second one relies on the use of the wearable technology. The first section of the thesis describes the first method which focuses on an electromyographic biofeedback rehabilitation program for the lower limbs, while the second one develops a method to estimate clinical scores through the use of wearable sensors, centering mostly on the upper limbs. As regards the first method, one of the primary goals for neurological rehabilitation after stroke is to walk independently with a velocity and endurance which allow the patient to take part in home and community activities of daily living. In particular looking at the gait patterns of hemiparetic stroke survivors they tend to have a severe reduction of ankle power in the push-off phase of the gait cycle, which can lead to an overall reduced gait quality and velocity. The ankle power produced at this stage is a parameter which can be measured only by a laboratory integrated system for gait analysis. Recent studies have shown that rehabilitation protocols based on electromyographic biofeedback (BFB), in a motor learning context, in post-stroke patients, might improve the ankle joint power production during the push off, increase the speed and enhance the quality of their gait. The functional magnetic resonance imaging provides an effective approach to analyze brain activity during cognitive and motor tasks in both healthy subjects and, in this context, patients with neurological damages, thanks to the high spatial resolution and safety. So far fMRI has been applied almost exclusively to the upper limbs' analysis, only a few studies have investigated the brain activation during tasks related to the ankle joint. The overall analysis of the current state of art highlights the need to investigate the ankle plantar-dorsiflexion in these patients, from a clinical, kinematic, dynamic point of view and to follow the ankle recovery in relation to possible changes in brain activity. The study involves the recruitment of 1 control subject and 4 patients with chronic hemiparesis due to ischemic or hemorrhagic stroke, the patients underwent clinical evaluation, instrumented gait analysis and fMRI analysis. First a preliminary analysis at T0 (i.e. pre-rehabilitation treatment) was performed on post-stroke patients in order to identify the brain activation areas during the motor task of active and passive plantar-dorsiflexion, both for the unaffected and affected side, comparing the obtained results with the healthy subject data. Afterwards a longitudinal study was carried out on post-stroke patients, before and after a BFB rehabilitation program, aimed at improving the global ankle functionality during gait. Both gait analysis data (kinetics, kinematics and electromyographic activity (EMG)) and fMRI data were collected at T1 (i.e. 2 months after T0) to assess the stabilization of the functional clinical picture and to test the repeatability/reliability of the data collected at T0. Then patients underwent a BFB rehabilitation intervention and they were asked to come back for further assessments within 1 week after the end of rehabilitation (T2) and after 3 months of follow-up (T3) after the initial assessment. The T2 and T3 assessments included the same tests performed at T1. The healthy subject underwent the fMRI analysis, in order to determine the normative brain activation areas. Gait analysis was performed by means of an integrated motion analysis system consisting of infrared cameras synchronized with force plates and an EMG device in order to evaluate the subjects' kinematics, kinetics and the patterns of muscle activation. The data analysis, the functional assessment, the correlation between the gait analysis and the brain activity parameters were used to identify the possible relationship between the motor ability improvement and the brain reorganization after the BFB treatment; this might become a crucial aspect for the longitudinal assessment of the motor recovery in neurological patients and for the analysis of the mutual interactions between brain activation maps and gait analysis parameters. In this context clinical assessment scales to evaluate motor abilities in stroke survivors could be used to individualize rehabilitation interventions thus maximizing motor gains. Unfortunately, these scales are not widely utilized in clinical practice because their administration is excessively time-consuming. Wearable sensors could be relied upon to address this issue, so the second method exploits this technology. Sensor data could be unobtrusively gathered during the performance of motor tasks. Features extracted from the sensor data could provide the input to models designed to estimate the severity of motor impairments and functional limitations. In previous work, it has been shown that wearable sensor data collected during the performance of items of the Wolf Motor Function Test (a clinical scale designed to assess functional capability) can be used to estimate scores derived using the Functional Ability Scale, a clinical scale focused on quality of movement. The purpose of the study herein presented was to investigate whether the same dataset could be used to estimate clinical scores derived using the Fugl-Meyer Assessment scale (a clinical scale designed to assess motor impairments). The results showed that Fugl-Meyer Assessment Test scores can be estimated by feeding a Random Forest with features derived from wearable sensor data recorded during the performance of as few as a single item of the Wolf Motor Function Test. Estimates achieved using the proposed method were marked by a root mean squared error as low as 4.7 points of the Fugl-Meyer Assessment Test scale.