Identification des synergies musculaires lors de l'activité de la marche chez la personne atteinte d'arthrose du genou

Abstract

Automatic identification and precise on/off muscle activity from surface electromyography signal (EMG) is very important in the analysis of neuromuscular alterations in patients with osteoarthritis of the knee during locomotion. So far, there is no gold standard for automatic on/off EMG activity detection, except the visual method whom still subjective. There are several methods for automatic EMG activity detection based on different approaches: an approach by thresholding, approach by energy TKE, approach by maximum likelihood AGLR, and finally approach by combined TKE / AGLR. These methods have limitations when the signal to noise ratio is low, which is the case in the analysis of the locomotion of knee osteoarthritis patients. The aim of our study is to develop a new method for automatic EMG activity during walking in knee OA patients, based on the results of the analysis of surface EMG signals of each gait cycle. These results allow analyzing the patterns of muscle activation by gait cycle of a group of patient with osteoarthritis, to estimate changes in neuromuscular muscle of the lower limbs. Our study will be conducted on 2 healthy and 4 knee OA patients grade KL = 3.4. 16 bilateral muscles of 2 lower limbs will be evaluated during the comfortable speed walking on a treadmill with two belts. Four automatic methods have been developed to determine the EMG activity: MS (threshold approach) method, TKE (energy approach) method, the probabilistic method (maximum likelihood approach AGLR) and finally the combined method TKE / AGLR. The patterns of activation of the quadriceps and hamstring muscles of healthy differ from one subject to another, while those of knee OA patients seem to activate during the same cycle phases, but with different numbers of activations and frequencies of occurrence. This suggests that the neuromuscular control of knee OA patients tend to use the same strategies, and therefore a "typical pattern of knee OA patients." The patterns of activations of TA, GL and GM muscles are different from one participant to another. The duration of activation of the quadriceps and BF muscles of knee OA patients is greater during the stance phase and shorter during the swing phase, compared to healthy subjects. In addition, the ST muscle of knee OA patients seems to be more active during the swing phase and the beginning of the stance phase. The TA muscle seems to be more used by knee OA patients to counter plantar flexion caused by the activation of G muscle, thus limiting knee flexion during the stance phase. The level of activation of VL, VM and BF muscles of knee OA patients is greater during the stance phase compared to those of healthy subjects. In addition, knee OA patients deploy greater muscle activation on the lateral side during the stance phase. The level of activation of ST, TA and GM muscles of knee OA patients is lower compared to healthy subjects. Analysis of EMG signals by gait cycle allowed us to observe that the knee OA patients use different neuromuscular control strategies from those of healthy subjects, in terms of pattern, level, and duration of activation. In addition, the new method for determining the muscle activation, according to the minimum RMS criterion enabled the TKE method to automatically detect muscle activations on different cycles of EMG signals.

Date	8 Oct 2013
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Rachid Aissaoui (Supervisor) & Jacques A. de Guise (Co-supervisor)