Fabiola Spolaor

SVM classification of locomotion modes using surface electromyography for applications in rehabilitation robotics

The next generation of tools for rehabilitation robotics requires advanced human-robot interfaces able to activate the device as soon as patients motion intention is raised. This paper investigated the suitability of Support Vector Machine (SVM) classifiers for identification of locomotion intentions from surface electromyography (sEMG) data. A phase-dependent approach, based on foot contact and foot push off events, was employed in order to contextualize muscle activation signals. Good accuracy is demonstrated on experimental data from three healthy subjects. Classification has also been tested for different subsets of EMG features and muscles, aiming to identify a minimal setup required for the control of an EMG-based exoskeleton for rehabilitation purposes.

Gait and posture analysis in patients with maxillary transverse discrepancy, before and after RPE

AIM The purpose of this study was to evaluate the effects of the rapid palatal expansion (RPE) on posture and gait analysis in subjects with maxillary transverse discrepancies. MATERIAL AND METHODS Forty-one patients between 6 and 12 years were divided into 3 groups: 10 control subjects (Cs), 16 patients with unilateral posterior crossbite (CbMono), 15 patients with maxillary transverse discrepancy and no crossbite (Nocb). Every subject underwent gait analysis and posturographic examination in order to evaluate the presence of balance alterations before (T0) and after (T4) RPE application. The examinations were performed through a six-cameras stereophotogrammetric system (60-120Hz, BTS S.p.A.) synchronized with two force plates (FP4060, Bertec Corp.). Romberg test was performed on a force plate, and the statokinesiogram and joint kinematics were evaluated. One-way Anova was performed among the variables after evidence of normal distribution (Levenes test for equality of variances) and Kruskal-Wallis test (P<0.05), in order to compare the three groups of subjects. While paired t-test was performed, or Kruskal-Wallis test, instead when comparing pre- and post-RPE application within the same group of subjects (P<0.05). Tamane T2 or Bonferroni correction was applied where needed. RESULTS The posturographic analysis reveal significant differences across the 3 population: 95% power frequency in medio-lateral and antero-posterior direction in T0, median frequency in medio-lateral direction in T0, mean power frequency in medio-lateral direction in T0. Significant differences were also registered in the three-dimensional joints kinematics variables, mainly between Cs and Cbmono in T0 and T4 and between Cbmono and Nocb in T4. CONCLUSIONS A detectable correlation between dental occlusion and body posture is shown in this study that confirms another benefit of the RPE. This was mainly revealed in the dynamic posture where modifications at the mandibular level affect the whole body.

GMM-Based Single-Joint Angle Estimation Using EMG Signals

This paper aims to explore the possibility to use Electromyography (EMG) to train a Gaussian Mixture Model (GMM) in order to estimate the bending angle of a single human joint. In particular, EMG signals from eight leg muscles and the knee joint angle are acquired during a kick task from three different subjects. GMM is validated on new unseen data and the classification performances are compared with respect to the number of EMG channels and the number of collected trials used during the training phase. Achieved results show that our framework is able to obtain high performances even using few EMG channels and with a small training dataset (Normalized Mean Square Error: 0.96, 0.98, 0.98 for the three subjects, respectively), opening new and interesting perspectives for the hybrid control of humanoid robots and exoskeletons.

A methodological framework for detecting ulcers’ risk in diabetic foot subjects by combining gait analysis, a new musculoskeletal foot model and a foot finite element model

Diabetic foot is one of the most debilitating complications of diabetes and may lead to plantar ulcers. In the last decade, gait analysis, musculoskeletal modelling (MSM) and finite element modelling (FEM) have shown their ability to contribute to diabetic foot prevention and suggested that the origin of the plantar ulcers is in deeper tissue layers rather than on the plantar surface. Hence the aim of the current work is to develop a methodology that improves FEM-derived foot internal stresses prediction, for diabetic foot prevention applications. A 3D foot FEM was combined with MSM derived force to predict the sites of excessive internal stresses on the foot. In vivo gait analysis data, and an MRI scan of a foot from a healthy subject were acquired and used to develop a six degrees of freedom (6 DOF) foot MSM and a 3D subject-specific foot FEM. Ankle kinematics were applied as boundary conditions to the FEM together with: 1. only Ground Reaction Forces (GRFs); 2. OpenSim derived extrinsic muscles forces estimated with a standard OpenSim MSM; 3. extrinsic muscle forces derived through the (6 DOF) foot MSM; 4. intrinsic and extrinsic muscles forces derived through the 6 DOF foot MSM. For model validation purposes, simulated peak pressures were extracted and compared with those measured experimentally. The importance of foot muscles in controlling plantar pressure distribution and internal stresses is confirmed by the improved accuracy in the estimation of the peak pressures obtained with the inclusion of intrinsic and extrinsic muscle forces.

Altered EMG patterns in diabetic neuropathic and not neuropathic patients during step ascending and descending.

Diabetic peripheral neuropathy (DPN) causes motor control alterations during daily life activities. Tripping during walking or stair climbing is the predominant cause of falls in the elderly subjects with DPN and without (NoDPN). Surface Electromyography (sEMG) has been shown to be a valid tool for detecting alterations of motor functions in subjects with DPN. This study aims at investigating the presence of functional alterations in diabetic subjects during stair climbing and at exploring the relationship between altered muscle activation and temporal parameter. Lower limb muscle activities, temporal parameters and speed were evaluated in 50 subjects (10 controls, 20 with DPN, 20 without DPN), while climbing up and down a stair, using sEMG, three-dimentional motion capture and force plates. Magnitude and timing of sEMG linear envelopes peaks were extracted. Level walking was used as reference condition for the comparison with step negotiation. sEMG, speed and temporal parameters revealed significant differences among all groups of patients. Results showed an association between earlier activation of lower limb muscles and reduced speed in subjects with DPN. Speed and temporal parameters significantly correlated with sEMG (p<0.05). The findings of this study are encouraging and could be used to improve rehabilitation programs aiming at reducing falls risk in diabetic subjects.

Processing of sEMG signals for online motion of a single robot joint through GMM modelization

This paper evaluates the use of Gaussian Mixture Model (GMM) trained through Electromyography (EMG) signals to online estimate the bending angle of a single human joint. The parameters involved in the evaluation are the number of Gaussian components, the channel used for model, the feature extraction method, and the size of the training set. The feature extraction is performed through Wavelet Transform by investigating several kind of configuration. Two set of experimental data are collected to validate the proposed framework from 6 different healthy subjects. Trained GMMs are validated by comparing the joint angle estimated through Gaussian Mixture Regression (GMR) with the one measured on new unseen data. The goodness of the estimated date are evaluated by means of Normalized Mean Square Error (NMSE), while the time performances of the retrieval system are measured at each phase in order to analyze possible critical situations. Achieved results show that our framework is able to obtain high performances in both accuracy and computation time. The whole procedure is tested on a real humanoid robot by remapping the human motion to the robotic platform in order to verify the proper execution of the original movement.

Comparison of lower limb muscle strength between diabetic neuropathic and healthy subjects using OpenSim

Diabetes neuropathy and vasculopathy are the two major complications of diabetes mellitus, leading to diabetic foot disease, of which the worst consequences are plantar ulcers and amputations. Motor impairments like joint stiffness and loss of balance are distinctive effects of diabetes and they have been extensively explored. However, while altered muscle function has been also assessed through experimentally measured surface electromyography, little is known about muscle forces. The objective of this study was to estimate muscle forces in subjects with diabetes and to use these data to identify differences with respect to a population of healthy subjects matched for age and BMI. This was obtained by generating musculoskeletal models of 10 diabetic and 10 control subjects in OpenSim starting from experimentally recorded data. Dynamic simulations of motion were run and hence muscle forces calculated. Student T test (p<0.05) was used to compare joints kinematics, kinetics and muscle forces between the two populations. Significant changes were observed between lower limb muscle forces and activation of diabetic and healthy subjects, as well as between joints kinematics and kinetics. In particular muscles related to foot movements proved to be stronger in the healthy population. The typical ankle rigidity of the diabetic population was confirmed by a lower range of motion registered at the ankle plantar/flexion angle associated with weaker dorsal-plantar flexor muscles. The information provided by this methodology can help planning specific training programs aiming at augmenting muscle strength and joints mobility, and they can also improve the evaluation of the potential benefits.

Biomechanical evaluation of diabetic foot through hierarchical cluster analysis

Type 2 diabetes is predicted to become the 7th leading cause of death in the world by the year 2030 [1]. Diabetic foot is the most common long-term diabetic complication, and it is a major risk factor for plantar ulceration (PU), it is determined by peripheral neuropathy (PN), vascular disease, increased foot pressures, foot trauma, deformity and callus [1]. The aim of this study is to develop a methodology for automatic detection of patients at risk for PU based on 3 dimensional (3D) multisegment foot biomechanics through cluster analysis.