Improving signal reliability for on-line joint angle estimation from nerve cuff recordings of muscle afferents

Abstract

Closed-loop functional electrical stimulation (FES) applications depend on sensory feedback, thus, it is important to continuously investigate new methods to obtain reliable feedback signals. The objective of the present paper was to examine the feasibility of using an artificial neural network (ANN) to predict joint angle from whole nerve cuff recordings of muscle afferent activity within a physiological range of motion. Furthermore, we estimated how small changes in joint angle that can be detected from the nerve cuff recordings. Neural networks were tested with data obtained from ten acute rabbit experiments in simulated, on-line experiments. The electroneurograms (ENG) of the tibial and peroneal nerves were recorded during passive ankle joint rotation. To decrease the joint angle prediction error with new rabbit data, we attempted to pretune the nerve signals and re-trained the ANNs with the pretuned data. With these procedures we were able to compensate for interrabbit variability. On average the mean prediction errors were less than 2.0 degrees (a total excursion of 20 degrees) and we were able to predict joint angles from muscle afferent activity with accuracy close to the best-estimated angular resolution. The angular resolution was found to depend on the initial joint angle and the actual step size taken and we found that there was a low probability of detecting joint angle changes less than 1.5 degrees. We thus suggest that muscle afferent activity is applicable as feedback in real-time closed-loop control, when the motion speed is restricted and when the movement is limited to a portion of the joint's physiological range.

Model-based ankle joint angle tracing by cuff electrode recordings of peroneal and tibial nerves

“…Therefore, the generalization ability and model-free structure characteristics of soft-computing algorithms [2,18] have been exploited to extract the relevant kinematic information. Jensen et al [14] proposed an artificial neural network to predict the joint angle. This network improved the reliability of the on-line joint angle estimation and provided a compensation for inter-subject variability.…”

Using nerve signals from muscle afferent electrodes to control FES-based ankle motion in a rabbit

“…Control was first performed with a neural network trained off-line only with data from rabbits used in previous experiments (see [1]). The performance of the closed-loop system was found to be very poor with this network, termed NET A.…”

“…Thus, our research group has been using implanted cuff electrodes to extract relevant information by analyzing signals obtained directly from the nerve bundles that carry it [1,2,3]. So far, a simple approach has been taken: rectified and bin integrated (RBIN) Electroneurographic (ENG) signals have been monitored and have been found to allow for a reasonable mapping onto angular and torque data by means of neural and fuzzy models [1,2]. The present paper shows our first findings from using the extracted angular information in a closed-loop controller.…”

“…A number of features in both time and frequency domains can somehow convey the necessary information albeit in a non-linear and time-variant fashion, and with signal-to-noise ratios much less than unity for cuff electrodes. Thus, our research group has been using implanted cuff electrodes to extract relevant information by analyzing signals obtained directly from the nerve bundles that carry it [1,2,3]. So far, a simple approach has been taken: rectified and bin integrated (RBIN) Electroneurographic (ENG) signals have been monitored and have been found to allow for a reasonable mapping onto angular and torque data by means of neural and fuzzy models [1,2].…”

Wavelet packet analysis for angular data extraction from muscle afferent cuff electrode signals

“…Thus, our research group has been using implanted cuff electrodes to extract angular information by analyzing electroneurographic (ENG) signals obtained directly from the nerve bundles that carry it. So far [1,2,3] rectified and bin integrated (RBIN) ENG signals have been monitored and have been found to allow for a reasonable mapping onto angular data by means of neural and fuzzy techniques. However, in [4] it was shown that the RBIN method is unsuited for signals with very low signal-tonoise ratios, which is often our case.…”