Adaptação do usuário de próteses mioelétricas : implicações na aprendizagem dos movimentos da mão
Nenhuma Miniatura disponível
Data
2017-03-30
Autores
Costa, Regina Mamede
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal do Espírito Santo
Resumo
From the physiological point of view, the knowledge of the anatomical and functional structure of the hand is essential for the understanding of the ostheomyoarticular mechanisms responsible for the movements of the fingers and their relation to the grasping functions. When injury occurs in one of these structures, the hand can be impaired, losing all functions as in the cases of amputation of the upper limb. The use of surface electromyography to control upper limb prostheses is an important clinical option, which offers the amputated an autonomy of control through the contraction of residual muscles. The functional biomechanical complexity of the hand involves a large area of representation in the cerebral cortex. In general, motor learning aims to maintain existing skills, the re-acquisition of lost skills and the learning of new skills. The goal of this Ph.D thesis was to propose and evaluate the application of a new experimental protocol for adaptation to the myoelectric prosthesis based on the distinction of hand movement patterns captured by sEMG of the remaining limb of amputees using myoelectric signals (SMEs).Ten upper limb amputees of both sexes, mean age 38.4 years ± 14.58, were evaluated. The inclusion criteria were: (1) transradial amputation or disarticulation of the wrist, unilateral or bilateral; (2) show no neurological or musculoskeletal disorder; (3) present no restriction of joint mobility. All of them were previously assessed including aspects of identification, anamnesis and physical examination. For MES acquisition, four Ag/AgCl active bipolar electrodes were used (TouchBionic ®). All the electrodes were placed according to the SENIAM (Surface Electromyography for the Non-Invasive Assessment of Muscle) recommendation. To the MES digitalization, an National Instrument NI USB-9001 acquisition system was used, and to the visualization of the MES, a digital processing software was developed, with interface to Matlab. The experimental protocol established a total of thirteen movements, which were grouped into two categories: GA (individual finger movements and hands opening and closure) and GB (grasping movements). The participants executed the tasks in three consecutive days. Two schemes were defined for the data capture: training phase and validation phase. The data concerning the first session (S1) were used to obtain a model of mechanical learning to the patterns classification, and the second (S2) and third (S3) sessions were used for the system validation. Although all tasks were performed in the same experiment, each category was studied and analyzed independently. Effectiveness (Acc), Kappa Coefficient (k) and Specificity (Sp) were calculated to evaluate the performance of each classifier of the executed movement. Positive-Negative Measurement (PNM) indicator was used to measure the performance of the thirteen proposed movements, and Goal Attainment Scale (GAS) was used to assess the extent to which individual objectives of each user were reached during the intervention. During the sessions, there were differences in the performance of the subjects during the proposed movements, which means that some participants could easily maintain repeated patterns, even with few training sessions, while others may need a longer training time to ensure good performance. Regarding the results of effectiveness, specificity, Kappa coefficient and PNM, the fact that the tasks of group A are simpler may explain the better performance of the volunteers in this group of tasks in relation to the performance in the tasks of group B (GB). On the other hand, the values obtained by GAS showed a satisfactory amount of correctness for the objectives outlined. Thus, this study showed that the subjects were able to perform muscular contractions, that is, perform the same movement with distinguishable MES patterns in the three experimental sessions, therefore, the proposed experimental design was validated in all the amputees of this study.
Descrição
Palavras-chave
Amputees , Myoelectric control , Pattern recognition , Hand movements , Myoelectric prosthesis , Rehabilitation , Amputados , Controle mioelétrico , Padrões de reconhecimento , Movimentos da mão , Próteses mioelétricas , Reabilitação