Abstract
Linear summation of torque, produced by selective activation of two motor fascicles in the cat sciatic nerve, was achieved using a four contact self-sizing spiral cuff electrode and a 900msec delay between pulses. Linear summation of two different torque outputs was successfully achieved in 125 out of 129 trials across five cats. These trials were performed using cuff electrodes that were in place for periods ranging up to 428 days. The results of these studies support the hypothesis that a single self-sizing spiral cuff with multiple contacts and a single lead may be used in place of several muscle-based electrodes each with its own separate lead.
Introduction
The goal of this project
was to demonstrate that separate motor nerves, activated selectively and
independently using a self-sizing spiral cuff electrode, can be activated in
combination to elicit a torque that is the linear sum of the individual torque
outputs. The ultimate goal of this work
is to support the hypothesis that a single multi-contact peripheral nerve based
cuff electrode can be used in motor prostheses to replace multiple muscle-based
electrodes. Multi-contact
cuff electrodes on peripheral nerves are an attractive alternative to muscle
base electrodes because they offer the possibility of using reduced currents,
simplified surgery to implant and more compact hardware. In a multiple contact
cuff electrode, however, the individual contacts are sufficiently close
together that attempting to simultaneously apply multiple stimulation patterns
will result in the excitatory fields summing together and result in a nonlinear
addition of the evoked torque. Although
a “look-up” table that relates every stimulation combination to joint torque
can be created, the number of combinations can become excessively large. The problem of field summation may be avoided,
if the applied stimuli are separated by a time that is long enough to allow
partially depolarized fibers to recover, yet close enough in time such that
fibers activated by the first stimuli are not activated by the second stimuli. Even though the two stimuli are separated in
time, the delay will be small enough to effect “simultaneous” activation at the
level of the muscle. Previous studies [1-4] suggest a 900 µs separation should be effective.
Methods
The objective of this study was to test if the linear addition of two individually produced outputs could be achieve by using a 900 µs delay between the stimuli. The ability to achieve linear addition was tested over the full range of torque outputs from partial activation to complete activation of individual fascicles.
These experiments used self-sizing nerve cuff electrodes, each containing 4 monopolar contacts, that were implanted, for periods ranging up to 428 days, on the sciatic nerves of five adult cats (masses ranging from 3.2 - 4.3 kg). The sciatic nerve of a cat, approximately 3mm in diameter, contains both smaller fascicles that serve only one or two muscles and larger fascicles that each serve multiple muscles and have sensory components. Together, these four motor nerves innervate all of the muscles controlling the ankle joint. The net resulting isometric torque about the ankle was measured as an indicator of the electrical excitation that was produced in the peripheral nerve [5].
Electrical excitation of the nerve was achieved using a 10µs cathodic pulse applied using either a monopolar configuration or using two contacts. In the case of monopolar stimulation, one contact was pulsed cathodically and the return electrode was a hypodermic needle placed in the nape of the animal’s neck. In the cases where two contacts were used, all of the current was passed between the two contacts (one contact was the cathode and the other contact was the anode).
The net torque output is graphically presented as a plot of the Plantar Flexion torque versus External Rotational torque. In this graphical presentation, isometric contraction of a single muscle produces a straight line [6]. The directionality of the torque vector is representative of which muscle was activated. The magnitude of the torque vector is representative of the level of muscle activation.
The tests performed in this study consisted of recording the output of each stimulus individually both immediately before and immediately after applying stimulation to the combination of the two stimuli (separated by 900µs). The difference between the average torque output produced by the two stimuli together and the sum of the average torque outputs produced by each stimulus individually was compared against the 98% confidence interval in each direction of torque output. The 98% confidence intervals, based on the variability of the torque resulting from supra-maximal stimulation, were 5.0, 2.5, and 1.7 N-cm for the plantar flexion, medial rotation and inversion axes respectively.
Results
A
total of 129 data sets were collected from 5 to 22 combinations of amplitudes
on 10 different pairs of electrode configurations across 5 different
animals. For each data set, the torque
output produced by two mutually exclusive stimulation pulses applied 900µsec
apart was compared to the linear summation of the corresponding individual
torque outputs. The 900µsec delay was
chosen based on the results of the delayed stimulation section such that the
second pulse would be after the facilitative period but during the refractory
period of the first pulse. Examples of
seven of the 129 data sets are shown in Figure 1. These seven data sets used the same pair of
contacts in one animal. The two solid
lines represent the torque output due to applying stimuli to each contact
alone. Four torque outputs were
identified along both of those lines as indicated by the corresponding squares
in Figure 1. Originating at each of
these four torque outputs are dashed lines that illustrate the linear summation
of that particular torque output with the torque outputs produced by the other
contact. The intersection points of the
dashed lines are the linear summation of each of the achieved torque
outputs. Seven of these potential torque
summations were attempted, as indicated by the seven intersection points that
have been circled. In all seven of these
cases the actual and predicted torque output were found to be within the 98%
confidence interval of each other.
For each data set, the differences between the actual and predicted torque outputs were compared with respect to plantar flexion, lateral rotation and eversion torque. In Figure 2 is shown a histogram of the differences for each of the three axes. The dotted vertical lines are at the location of the 98% confidence interval for each of the three axes. Of the 387 data points (129 data points in each of the 3 dimensions), only 4 cases were found to fall outside of the 98% confidence region. Of the 4 cases found in Figure 2, 1 occurrence was found in plantar flexion, 2 occurrences were found in lateral rotation and 1 occurrence was found in eversion. Using a paired t-test in each of the three dimensions, the difference between the predicted torque output and the actual torque output was found to be less than one half of a N-cm at an alpha value better than 0.0001.
Discussion
The results of these experiments show that the torque output produced by two configurations activated separately can be linearly added. To be able to effect linear addition of the two torque values, the stimuli to the second configuration was delivered 900 µs after the first fascicle was activated. The resulting torque output of two stimuli separated by a delay of 900msec, was found to be a linear summation of the torque outputs produced by each stimulus alone. This technique can be used to effect linear addition of torque and thus produce complex patterns of limb movement.
A value of 900 msec was chosen based on previous results that characterized the duration of the facilitative and refractory periods for the cat sciatic nerve to be 700 and 900msec, respectively. For this study, stimulation was applied to two different contact sets to determine if a linear summation could be achieved. Since each pair of stimuli used in this experiment activated different sets of axons, the same axons were not activated twice. The regions of sub-threshold excitation resulting from each stimulus, however, may overlap and were therefore susceptible to activation due to facilitation. The longest delay value, within the range found in the delayed stimulation section (900msec), was therefore chosen.
Conclusions
Linear summation of torque,
produced by selective activation of two motor fascicles in the cat sciatic
nerve, was achieved using a four contact self-sizing spiral cuff
electrode. The results of these studies
support the hypothesis that a single self-sizing spiral cuff with multiple
contacts and a single lead may be used in place of several muscle-based
electrodes each with its own separate lead.
Bibliography
1. Peckham, P.H., Electrical Excitation of Skeletal Muscle: Alterations in Force, Fatigue, and Metabolic Properties, in Biomedical Engineering. 1972, Case Western Reserve University: Cleveland, OH.
2. Liang, D.H., et al., A Method for Evaluating the Selectivity of Electrodes Implanted for Nerve Simulation. IEEE Trans. Biomed., 1991. 38(5): p. 443-449.
3. Rutten, W.L.C., H.J.v. Wier, and J.H.M. Put, Sensitivity and Selectivity of Intraneural Stimulation Using a Silicon Electrode Array. IEEE Trans. BME, 1991. 38: p. 192-198.
4. Yoshida, K. and K. Horch, Selective Stimulation of Peripheral Nerve Fibers Using Dual Intrafascicular Electrodes. IEEE Trans. Biomed. Eng., 1993. 40: p. 492-4.
5. Grill, W.M. and J.T. Mortimer, Non-Invasive Measurement of the Input-Output Properties of Peripheral Nerve Stimulation Electrodes. J. Neurosci. Methods, 1996. 65: p. 43-50.
6. Lawrence, J.H., T.R. Nichols, and A.W. English, Cat Hindlimb Muscles Exert Substantial Torques Outside the Sagittal Plane. J Neurophys, 1993. 69: p. 282-285.
Acknowledgements
This work was funded by the National Institute of Health Contract # NO1-NS-32300. We would also like to thank Dustin Tyler and Harish Aiyar for their help and assistance in the experimental set-up and surgical procedures.