Design of multiple degree-of-freedom sliding mode FES controller for concurrent stimulation of multiple mono and biarticulate muscles.

This article presents a theoretical design of an FES controller to be used for stimulation of multiple mono and biarticulate muscles to restore multiple degree-of-freedom (DOF) motion in paralyzed individuals. The overall control strategy is based on multiple DOF musculo-skeletal model, nonlinear sliding mode control design, constrained optimization techniques to determine the needed muscle activations, and an additional inversion of the neuro-muscular stimulation relationship in order to obtain the needed controller output (electrical current amplitude/pulse-width). The combination of these methods leads to a controller that guarantees asymptotically stable tracking of reference position trajectories while assuring minimal (optimal) muscle activation and fatigue.

Adaptive robotic rehabilitation of locomotion: a clinical study in spinally injured individuals.

STUDY DESIGN: Clinical study on six spinal cord-injured subjects. The performance of two automatic gait-pattern adaptation algorithms for automated treadmill training rehabilitation of locomotion (called DJATA1 and DJATA2) was tested and compared in this study. OBJECTIVES: To test the performance of the two algorithms and to evaluate the corresponding patient satisfaction. We also wanted to evaluate the motivation of the patients to train with a fixed gait pattern versus training where they can influence and change the gait pattern (gait-pattern adaptation). SETTING: Spinal Cord Injury Center Paracare, Balgrist, Zürich, Switzerland. METHODS: The experimental data were collected during six blinded and randomized training trials (comprising three different conditions per algorithm) split into two training sessions per patient. During the experiments, we have recorded the time courses of the six parameters describing the adaptation. Additionally, a special patient questionnaire was developed that allowed us to collect data regarding the quality, perception, speed, and required effort of the adaptation, as well as patients' opinion that addressed their motivation. The achieved adaptation was evaluated based on the time course of adaptation parameters and based on the patient questionnaire. A statistical analysis was made in order to quantify the data and to compare the two algorithms. RESULTS: Significant adaptation of the gait pattern took place. The patients were in most cases able to change the gait pattern to a desired one and have always perceived the adaptation. No statistically significant differences were found between the performances of the two algorithms based on the evaluated data. However, DJATA2 achieved better adaptation scores. All patients preferred treadmill training with gait-pattern adaptation. CONCLUSION: In the future, the patients would like to train with gait-pattern adaptation. Besides the subjective opinion indicating the choice of this training modality, gait-pattern adaptation also might lead to additional improvement of the rehabilitation of locomotion as it increases and promotes active training. SPONSORSHIP: The work was supported by The Swiss Commission for Technology and Innovation (Project No. 4005.1).

Neural network classification of nerve activity recorded in a mixed nerve.

Whole-nerve cuff electrodes can be used to record electrical nerve activity in peripheral nerves and are suitable for chronic implantation in animals or humans. If the whole nerve innervates multiple target organs or muscles then the recorded activity will be the superposition of the activity of different nerve fibers innervating these organs. In certain cases it is desirable to monitor mixed nerve activity and to determine the origin (modality) of the recorded activity. A method using the autocorrelation function of recorded nerve activity and an artificial neural network was developed to classify the modality of nerve signals. The method works in cases where different end organs are innervated by nerve fibers having different diameter distributions. The electrical activity in the cat S1 sacral spinal root was recorded using a cuff electrode during the activation of cutaneous, bladder, and rectal mechanoreceptors. Using the classification method, 87.5% of nerve signals were correctly classified. This result demonstrates the effectiveness of the neural network classification method to determine the modality of the nerve activity arising from activation of different receptors.

Optimal filtering of whole nerve signals.

Electroneurographic recordings suffer from low signal to noise (S/N) ratios. The S/N ratio can be improved by different signal processing methods including optimal filtering. A method to design two types of optimal filters (Wiener and Matched filters) was developed for use with neurographic signals, and the calculated filters were applied to nerve cuff recordings from the cat S1 spinal root that were recorded during the activation of cutaneous, bladder, and rectal mechanoreceptors. The S1 spinal root recordings were also filtered using various band-pass (BP) filters with different cut-off frequencies, since the frequency responses of the Wiener and Matched filters had a band-pass character. The mean increase in the S/N ratio across all recordings was 54, 89, and 85% for the selected best Wiener, Matched, and band-pass filters, respectively. There were no statistically significant differences between the performance of the selected filters when all three methods were compared. However, Matched filters yielded a greater increase in S/N ratio than Wiener filters when only two filtering techniques were compared. All three filtering methods have in most cases also improved the selectivity of the recordings for different sensory modalities. This might be important when recording nerve activity from a mixed nerve innervating multiple end-organs to increase the modality selectivity for the nerve fibers of interest. The mean Modality Selectivity Indices (MSI) over different receptor types and for the same selected filters as above were 1.12, 1.27, and 1.29, respectively, and indicate increases in modality selectivity (MSI>1). Improving the S/N ratio and modality selectivity of neurographic recordings is an important development to increase the utility of neural signals for understanding neural function and for use as feedback or control signals in neural prosthetic devices.

Detection and inhibition of hyperreflexia-like bladder contractions in the cat by sacral nerve root recording and electrical stimulation.

Detection of bladder volume and hyperreflexive bladder contractions would be useful in individuals with overactive bladders. We sought to determine whether bladder filling and/or reflex bladder contractions could be detected by electrical recording from the sacral nerve roots, and whether bladder contractions could be inhibited by stimulation of sacral afferents. Six male cats were anesthetized with alpha-chloralose and bipolar cuff electrodes were used to measure sacral nerve root electroneurograms (ENG) during slow bladder filling, during rapid injections of fluid into the bladder, and during hyperreflexia-like bladder contractions. The rectified and time-averaged activity of the S1 extradural root increased by 0-5 % above the baseline during bladder filling. Rapid injections caused a sudden increase in bladder pressure, and a 3-36 % increase over baseline in the S1 nerve activity. Withdrawal of the same volume caused a reduction in pressure and a decrease in recorded activity (4-14 %). At the onset of a bladder contraction, there was a 7-38 % increase over baseline in the S1 nerve activity. This activity increase was sustained for the duration of the contraction and decreased during bladder relaxation. The onset and duration of bladder contractions could be detected consistently from these nerve activity changes. Recording only afferent activity showed that the increased nerve activity was due to S1 sensory rather than motor fibers. In two cats, it was demonstrated that an ongoing bladder contraction could be inhibited by rectal distension. In one cat, the contractions could be terminated by electrical stimulation of the S1 dorsal root. The results demonstrate that afferent sacral root nerve activity can be used to detect hyperreflexive bladder contractions at low bladder pressures. Such a signal might be used to trigger bladder inhibition via electrical stimulation of specific sacral afferents.

Analysis of bladder related nerve cuff electrode recordings from preganglionic pelvic nerve and sacral roots in pigs.

PURPOSE: Electrical stimulation of appropriate lower urinary tract (LUT) nerves may be used in bladder dysfunction to achieve continence and abolish hyper-reflexic detrusor contractions. It can also be used for consequent emptying of the bladder. To control the time course of the described functional phases, knowledge of bladder sensory information is needed. We investigated if the latter could be extracted from the LUT nerve activity. MATERIALS AND METHODS: In acute experiments using 10 pigs, tripolar cuff electrodes were placed unilaterally around the pelvic nerve and the S3 and S2 roots. The cuff electrode signals, filling rate and the bladder and rectal pressures were recorded during slow and fast bladder fillings/emptyings. RESULTS: Two pigs were excluded from the analysis because of no observed changes in the nerve signals in one animal, and because of electrical noise problems in the other animal. Fast bladder pressure increases resulted in a sudden pelvic nerve signal rise in 6 out of 7 pigs (3 out of 6 for the S3 nerve signal). Slow bladder pressure increase was reflected in the recorded nerve activity only in 3 out of 8 and in 3 out of 7 pigs for the pelvic and S3 cuff signals respectively. In 2 animals small spontaneous bladder contractions were clearly reflected in the pelvic nerve signal (contractions were observed only in 3 pigs). Except in one pig, there were no slow/fast bladder filling responses recorded in the S2 roots. It is shown that the recorded responses were afferent. CONCLUSIONS: Cuff electrodes can be used to record bladder afferent information from the pelvic nerve and the sacral root S3 in pig. Pelvic nerve recordings were more selective than the sacral root recordings. Nerve activity increases were more distinct and repeatable during rapid bladder pressure changes and small spontaneous bladder contractions than during slow bladder fillings.

On statistical properties of whole nerve cuff recordings.

Whole nerve cuff electrodes can record an electric signal generated by the superposition of single fiber action potentials (AP's). Using a simple stochastic model for the superposition of AP's, the statistical properties of nerve cuff signals are mathematically derived in this study. Consequences of common signal processing methods like rectification and time-averaging are also explained. The nerve cuff signals are found to be approximately identically, independently distributed Gaussian signals with zero mean and varying variance. The spectral properties of the cuff signals generated by single AP shape or different AP shapes are also addressed and investigated by examining the properties of the autocorrelation functions of the nerve cuff signals. The theoretical results were found to be in accordance with computer simulations and processing of actual recorded data.