Algorithm for real-time gait detection and selectivity analysis for transcutaneous spinal cord stimulation

Walking is one of the most impactful activities of life and it can be limited by orthopaedic or neurological disorders. Neurological diseases like Stroke, Multiple sclerosis, Spinal cord injury or Parkinson affect millions of people all over the world, compromising their lives and hindering everyday actions. The aim of this thesis is to show how Transcutaneous spinal cord stimulation (tSCS) can be used to recruit, in a selective way, the lower limb muscles, in particular flexors and extensors, during gait. tSCS is a non-invasive approach that consists of electrical stimulation via surface electrodes on the skin to activate the spinal network and re-establish the connection with the muscles, helping people with gait impairments, to restore or improve their motor faculty. To achieve this goal, the work consisted in a preliminary investigation on 2 healthy subjects in which was studied the muscle activity of the leg muscles and kinematic features as Hip and Knee ROM (range of motion) behaviour, with respect to the frequency of stimulation, focusing the attention on two phases of the gait: Stance and Swing phase. This first analysis has brought to light some important trends that needed to be analysed more in details. Official experiments were carried out on 14 subjects, without any gait disorders, who were asked to walk on a treadmill with a fixed speed of 3.5km/h and with different frequencies of stimulation of the spinal cord (from 5 to 100Hz, with a fixed current, chosen according to the subject threshold). The muscle activation and the Hip range of motion were considered as parameters to watch out to see the effectiveness of stimulation on muscles. The experiments show a trend for the muscle activation that increases for the Flexors muscles in swing phase with respect to the stimulation frequency and decreases for Extensors muscles in stance phase. For the Hip ROM the test also showed an increase in swing phase with higher stimulation frequency and in stance phase with lower stimulation frequency. All the results were analysed with a statistical test (ANOVA unidirectional) to assess the significance of the results obtained. The work of this thesis is a very new investigation in the field of gait rehabilitation using tSCS but shows encouraging patterns for the recruitment of lower limb muscles during walking, these results are very promising because regarded only healthy subjects and can be even better in subjects with pathologies. In parallel was also implemented an algorithm for the real-time detection of gait’s phases and subphases. The test was done on the same 14 subjects and shows an accuracy of more than 95% for the phases and more than 88% for the stance subphases. This analysis is crucial for the implementation, in the future, of the modulated multisite tSCS which purpose is to determine real-time what stage of walking the subject is in, and to stimulate the spine with the frequency of stimulation required for the muscles activated in that stage.