Neuro-fuzzy extraction of angular information from muscle afferents for ankle control during standing in paraplegic subjects: an animal model

Abstract

This paper is part of a project whose aim is the implementation of closed-loop control of ankle angular position during functional electrical stimulation (FES) assisted standing in paraplegic subjects using natural sensory information. In this paper, a neural fuzzy (NF) model is implemented to extract angular position information from the electroneurographic signals recorded from muscle afferents using cuff electrodes in an animal model. The NF model, named dynamic nonsingleton fuzzy logic system is a Mamdani-like fuzzy system, implemented in the framework of recurrent neural networks. The fuzzification procedure implemented was the nonsingleton technique which has been shown in previous works to be able to take into account the uncertainty in the data. The proposed algorithm was tested in different situations and was able to predict reasonably well the ankle angular trajectories especially for small excursions (as during standing) and when the stimulation sites are far from the registration sites. This suggests it may be possible to use activity from muscle afferents recorded with cuff electrodes for FES closed-loop control of ankle position during quite standing.

On the use of wavelet denoising and spike sorting techniques to process electroneurographic signals recorded using intraneural electrodes

“…Therefore, in most cases, the neural activity recorded in this way has been used for onset detection, for example for the control of a 1-DoF hand prosthesis (Stein et al, 1980) or of FES-systems in hemiplegics Hoffer et al (2005). Even if these limits can be partly overcome by using multi-site cuff electrodes (Yoo and Durand, 2005) and advanced processing techniques (Micera et al, 2001;Cavallaro et al, 2003;Lin et al, 2007;Tesfayesus and Durand, 2007), more selective PNS interfaces may be necessary to access more specific information. In fact, higher selectivity interfaces make possible the identification of single spikes from single axons (or a small group of axons) and to access the natural frequency coded information (Micera et al, 2006) in ENG signals.…”

Review of EMG-based neuromuscular interfaces for rehabilitation: elbow joint as an example

“…The input for the model is the data for 16 muscles through normalised EMG values, and the output is the angle and torque of the hip, knee and ankle (Sepulveda et al, 1997). The system was later improved for neuro-fuzzy control (Micera et al, 2001) and emotion-based interaction (Leon et al, 2004). Another representative neural network model is Prentice's model.…”

“…Neural network-based models were used to build up a relationship between EMG signals and the correspondent kinematic data of human movement (Prentice et al, 2001;Micera et al, 2001;Sepulveda et al, 1997;Jamwal and Xie, 2012). Compared with the original method of solving the inverse model, neural network-based models are much simpler, because they do not have complicated mathematical formula or time delays which are generated and associated by EMG.…”

Neuroprosthetic technology for individuals with spinal cord injury

“…119 Penetrating electrode arrays may offer additional advantages including the possibility of recording multiple independent commands from a single array of implanted electrodes. 116,120,121 Because of paralysis and weakness after SCI, EMG and ENG-based control may require the user to learn a new and often unnatural strategy to command FES stimulation. The muscles or movements chosen as actuators are often functionally and anatomically unrelated to the muscle(s) they control in order to avoid complications with coupling during use.…”