Correctly identifying gait phases is a prerequisite to achieve a spatial/temporal characterization of muscular recruitment during walking. Recent approaches have addressed this issue by applying machine learning techniques to treadmill-walking data. We propose a deep learning approach for surface electromyographic (sEMG)-based classification of stance/swing phases and prediction of the foot–floor-contact signal in more natural walking conditions (similar to everyday walking ones), overcoming constraints of a controlled environment, such as treadmill walking. To this aim, sEMG signals were acquired from eight lower-limb muscles in about 10.000 strides from 23 healthy adults during level ground walking, following an eight-shaped path including natural deceleration, reversing, curve, and acceleration. By means of an extensive evaluation, we show that using a multi layer perceptron to learn hidden features provides state of the art performances while avoiding features engineering. Results, indeed, showed an average classification accuracy of 94.9 for learned subjects and 93.4 for unlearned ones, while mean absolute difference ( ± S D ) between phase transitions timing predictions and footswitch data was 21.6 ms and 38.1 ms for heel-strike and toe off, respectively. The suitable performance achieved by the proposed method suggests that it could be successfully used to automatically classify gait phases and predict foot–floor-contact signal from sEMG signals during level ground walking.

The aim of this work is to determine whether postural sway can be well described by nonlinear deterministic modelling. Since the results of nonlinear analysis depend on experimental data processing, emphasis was given to the assessment of a proper methodology to process posturographic data. Centre of Pressure (CoP) anterior-posterior (AP) displacements (stabilogram) were obtained by static posturography tests performed on control subjects. A nonlinear determinism test was applied to investigate the nature of data. A nonlinear filtering method allowed us to estimate properly the parameters of the nonlinear model without altering signal dynamics. The largest Lyapunov exponent (LLE) was estimated to quantify the chaotic behaviour of postural sway. LLE values were found to be positive although close to zero. This suggests that postural sway derives from a process exhibiting weakly chaotic dynamics.

In the present study, spontaneous postural behavior has been analyzed in freely standing multiple sclerosis (MS) patients, exhibiting no clinically assessable abnormalities of postural control. This population has been compared with two other groups, healthy people and hemiparetic patients. This latter group represents a situation where the central nervous system (CNS) lesion is precisely localized in one anatomical site and no signal-conduction disorders are present; i.e., it has an opposite anatomical character with respect to the MS at a preclinical stage. The hypothesis underlying the modeling study is the presence of a controller block working in a feedback posture control system. This controller block receives the body sway as input, and produces the corresponding ankle torque stabilizing the body, the latter being modeled as an inverted pendulum. The CNS damage, caused by MS, is supposed to be reflected in some detectable change in the structure of the controller of the posture control system. The identification of the controller has been performed by means of a parametric estimation procedure which employed as input sequences, data recorded by means of a movement-analysis (MA) system. Reported findings show a structural changes of the model of the controller block in the posture control system. This result may suggest the presence of an MS-specific reorganization of the posture control system. Some speculation is finally made on the black-box approach in comparison with traditional posturography, to arrive at hypothesizing a progression path for postural disorders.

No abstract

Differences among methods rely on different techniques used to extract fECG. If pure abdominal electrode configurations are used, fECG is extracted directly from the abdominal recording using independent component analysis or template subtraction. Eventually, if mixed electrode configurations are used, the fECG can be extracted using the adaptive filtering fed with the maternal ECG recorded by the electrodes located in the woman thorax or shoulder.