Statistical Gait Analysis Based on Surface Electromyography.

To help neurologists, physicians, and physical therapists in the management of patients with altered locomotion patterns, it is of the uttermost importance relying on accurate measurements of gait. Gait analysis becomes even more informative if the electrical activity of muscles is recorded, non-invasively, during the dynamic task of walking, through surface electromyography (sEMG) probes. However, sEMG signals must be processed through advanced techniques to obtain reliable results, easily interpretable by healthcare practitioners. Indeed, the study of how muscles are activated during natural walking (in unconstrained environments) is complex for several reasons, including a high stride-to-stride variability, even more pronounced in pathological subjects. On the other hand, it is crucial to provide clinicians with aggregated information relying on validated parameters and easily usable representations that can be effectively included in clinical reports. This chapter is aimed at introducing: (1) Statistical Gait Analysis (SGA) to automatically analyze hundreds of gait cycles collected during a physiological or pathological walk lasting several minutes, (2) the extraction of principal and secondary muscle activations to obtain consistent clinical indexes, (3) the extraction of “muscle synergies” to quantitatively study motor control strategies. Each of these techniques are based on state-of-the-art processing algorithms of the sEMG signal. A brief review of the recent literature published in this field will be presented and discussed.