Reconstructing muscle activation during normal walking: a comparison of symbolic and connectionist machine learning techniques

Abstract

Abstract.One symbolic (rule-based inductive learning) and one connectionist (neural network) machine learning technique were used to reconstruct muscle activation patterns from kinematic data measured during normal human walking at several speeds. The activation patterns (or desired outputs) consisted of surface electromyographic (EMG) signals from the semitendinosus and vastus medialis muscles. The inputs consisted of flexion and extension angles measured at the hip and knee of the ipsilateral leg, their first and second derivatives, and bilateral foot contact information. The training set consisted of data from six trials, at two different speeds. The testing set consisted of data from two additional trials (one at each speed), which were not in the training set. It was possible to reconstruct the muscular activation at both speeds using both techniques. Timing of the reconstructed signals was accurate. The integrated value of the activation bursts was less accurate. The neural network gave a continuous output, whereas the rule-based inductive learning rule tree gave a quantised activation level. The advantage of rule-based inductive learning was that the rules used were both explicit and comprehensible, whilst the rules used by the neural network were implicit within its structure and not easily comprehended. The neural network was able to reconstruct the activation patterns of both muscles from one network, whereas two separate rule sets were needed for the rule-based technique. It is concluded that machine learning techniques, in comparison to explicit inverse muscular skeletal models, show good promise in modelling nearly cyclic movements such as locomotion at varying walking speeds. However, they do not provide insight into the

Fundamental patterns of bilateral muscle activity in human locomotion

“…It contains three maxima, and three minima which occur 50% later in the cycle than the maxima. The homologous pair of tibialis anterior and gluteus maximus load highly on this factor in opposite directions for the right and left sides ( Table 3), suggesting that its purpose is to maintain left/right symmetry cessfully modelled with artificial neural networks (Heller et al 1993;Sepulveda et al 1993), there has been no implementation to date of a model which incorporates the full complexity of the neural, muscular and skeletal systems in human gait (Vaughan et al 1995). Our findings, which strongly suggest that the central nervous Note the very high correlation for shifted factor 1 and factor 2, as well as for shifted factor 3 and factor 4.…”

Árvores De Decisão Como Ferramenta De Auxílio Na Análise Biomecânica

“…As séries temporais da VFRS força plantar foram coletadas em 8 Os sinais de VFRS foram obtidos a partir das sequências de mapas de pressão utilizando o software FScan e foram filtradas por um filtro passa-baixa Butterworth de fase zero, com ordem efetiva igual a 8 e frequência de corte em 6 Hz [1]. Rotinas no Matlab (2015a) foram desenvolvidas para a extração de 5 parâmetros temporais (dos lados direito e esquerdo), sendo eles o tempo de duplo apoio (TDA), tempo de apoio (TA), tempo de primeiro pico de força (TPP), tempo de vale (TV), tempo de segundo pico (TSP).…”

“…Derlatka e Ihnatouski analisaram e classificaram dados de pressão plantar de indivíduos que sofrem de pé valgo e paralisia cerebral [6], e Heller et al reconstruíram os padrões de ativação muscular a partir de dados cinemáticos [8]. O uso de árvores de decisão tem como vantagem a facilidade de extração de informação acerca dos dados devido ao formato simples das regras de decisão resultantes do seu processo de aprendizado, que, ao apontarem os principais discriminantes, reduzem a complexidade de interpretação dos dados, mitigando possíveis limitações impostas pela experiência a priori do investigador.…”

Inverse Dynamics and Artificial Neural Network Applications in Gait Analysis of the Disabled Subjects

“…Specifically, the authors discussed the modelling of muscle activity as a basis of modelling changes in kinematic gait parameters [50,51], modelling muscle work of lower extremities by using kinematic and kinetic parameters [52,53], or assessing muscle strength from EMG signals [13,54].…”

“…Muscle activity was determined by EMG and kinematic variables on a mechanical treadmill. Hel-ler, Veltink et al [52] designed a network fed by time functions of changes in hip and knee joint angles, angular velocities and accelerations, and durations of the contacts of feet with the ground, and the network generated EMG values for the semitendinosus muscle. The network could reproduce muscle activities and changes in kinematic variables in cyclical movements performed in six consecutive steps at various velocities.…”