Muscle recruitment with intrafascicular electrodes

We have studied muscle recruitment with Tefloninsulated, 25 pm diameter, Pt-Ir intrafasicular electrodes implanted in nerves innervating the gastrocnemius and soleus muscles of cats. The purpose of this study was to measure the performance of these bipolar electrodes, which had been designed to optimize their ability to record unit activity from peripheral nerves, as stimulating electrodes. Recruitment curves identified the optimal stimulus configuration as a biphasic rectangular pulse, with an interphase separation of about 500 ps and a duration of about 50 ps. The current required for a half-maximal twitch contraction was on the order of 50 PA. Current and charge densities needed for stimulation were well below levels believed to be safe for the tissue and electrode materials involved. When the spinal reflex pathway was interrupted by crushing the nerve, the force produced by a given stimulus changed in some cases, but not in others, implying that the spinal reflex contribution was not the same in all the implants. We conclude that intrafascicular recording electrodes are also a potentially valuable technology for functional neuromuscular stimulation, and warrant further development.

IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 38. NO. 8. AUGUST 1991 Muscle Recruitment with Intrafascicular Electrodes Nicola Nannini and Kenneth Horch, Member, IEEE Abstract-We have studied muscle recruitment with Teflon- insulated, 25 /im diameter, Pt-Ir intrafasicular electrodes im­planted in nerves innervating the gastrocnemius and soleus muscles of cats. The purpose of this study was to measure the performance of these bipolar electrodes, which had been de­signed to optimize their ability to record unit activity from pe­ripheral nerves, as stimulating electrodes. Recruitment curves identified the optimal stimulus configu­ration as a biphasic rectangular pulse, with an interphase sep­aration of about 500 fis and a duration of about 50 /xs. The current required for a half-maximal twitch contraction was on the order of 50 /j.A. Current and charge densities needed for stimulation were well below levels believed to be safe for the tissue and electrode materials involved. When the spinal reflex pathway was interrupted by crushing the nerve, the force produced by a given stimulus changed in some cases, but not in others, implying that the spinal reflex contribution was not the same in all the implants. We conclude that intrafascicular recording electrodes are also a potentially valuable technology for functional neuromus­cular stimulation, and warrant further development. Introduction PARALYSIS caused by spinal cord injury, head trauma, or stroke leaves the peripheral nerves and muscles in­tact, but interrupts communication between brain and pe­ripheral nervous system. In theory, mobility can be re­stored to paraplegic and quadriplegic patients by artificially stimulating motor nerves; the technique for doing so is called functional neuromuscular stimulation (FNS). The goal of the present project was to ascertain whether our new type of intrafascicular electrode [19] is elfective for FNS. The two types of electrodes commonly used for FNS are intramuscular electrodes and nerve culf electrodes. An intramuscular electrode is a metal wire inserted into the muscle which is used to stimulate the motor end plate [6], [8], [11], [26], A nerve cuff electrode is wrapped around the nerve and consists of an isolating cuff inside of which are two or more exposed metal areas which deliver current to trigger action potentials [13], [23], [25], [28]. Both types of electrodes elicit motoneuron activity with con­sequent muscle contraction and have been successfully used in clinical applications, but both have deficiencies. Manuscript received October 6, 1989; revised September 13, 1990. This work was supported by an NIH Biomedical Research Grant to the Univer­sity of Utah. N. Nannini was with the Department of Bioengineering, University of Utah, Salt Lake City, UT 84112. He is now a consultant for Italian hos­pitals, via Palazzine 7, Fiesole 50014, Italy. K. Horch is with the Department of Bioengineering, University of Utah, Salt Lake City, UT 84112. IEEE Log Number 9101195. Intramuscular electrodes are subject to mechanical fail­ure when implanted in contracting muscles, require high current pulses, and the recruitment curve is strongly de­pendent on the placement of the electrode relative to the motor point of the muscle [11]. Cuff electrodes are placed around the motor nerve, require custom fitting to the nerve diameter, and may damage the nerve if the fit is not op­timal. To stimulate only one muscle, or muscle group, the electrode must be placed in the vicinity of the target mus­cle, which can cause damaging stress on the electrode due to movement of the muscle. We have developed a new type of bipolar, intrafasci­cular electrode for recording activity from peripheral nerves [19]. An intrafascicular electrode is a form of in- traneural electrode which penetrates the perineurium and resides within an individual fascicle of a nerve. The pres­ent study was undertaken to test our expectation that the threshold currents required to elicit muscle contractions with our intrafascicular electrodes would be low. In FNS applications one needs to be able to predict the force produced by a given stimulus amplitude, which re­quires characterizing the stimulus-force relationship of the electrode. But even with this information, we would not be able to accurately predict the output force, for the re­cruitment curve varies substantially with time and muscle fatigue [12], In addition, reflexes originating in the spinal cord are present in patients with high spinal injuries or stroke. These reflexes are important because with electri­cal stimulation we activate nerve fibers independently of the direction in which they carry information: electrical stimulation of motor nerves will activate both motoneu­rons and sensory afferent fibers, particularly muscle spin­dle and Golgi tendon organ fibers, which elicit spinal re­flexes. The level of activation of the afferent fibers will influence the force produced by the muscle, Golgi tendon organs producing a negative force feedback, and muscle spindle aflferents providing a negative length feedback. It is difficult to accurately predict the muscle response to electrical stimulation because of the complexity of these interdependent effects and because it has been shown that the length of individual muscle spindle fibers depends on muscle length, load, level of activation, and location of the spindle within the muscle itself [17]. For these reasons, a clinical orthosis could benefit from closed-loop control. In this respect our electrode design is of interest since it can be used both for stimulation and for recording of feedback signals from muscle, joint, and cutaneous sensory receptors. 0018-9294/91/0800-0769S01.00 © 1991 IEEE L/cparuiiciii, id e e ruuiiMiuig ocrvitcs, *+*+j nucs i-anu. r.v ; dux i j j i , nscaiaway, inj 0 Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY and at adc I, i jyrigni tyyi dv me institute 01 tiectricai ana itional mailing offices. Postmaster: Send address IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING. VOL. 38. NO. 8, AUGUST 1991 NANNINI AND HORCH: MUSCLE RECRUITMENT WITH INTRAFASCICULAR ELECTRODES cutting the nerve is that crushing leaves the current path­ways as unaltered as possible while still effecting a com­plete block of conduction [18]. Finally, this set of pro­cedures was repeated with the soleus muscle. The data presented here are from experiments in 15 cats, although not all measurements were made in all cats. Instrumentation The suture attached to the tendon was connected to a strain gauge force transducer, the output of which was amplified, filtered, displayed on an oscilloscope, and fed to an 8-bit A/D converter on a microcomputer system. The stimulator consisted of a controlled current source driven with waveforms generated by the computer through an 8-bit D/A converter. Amplitude, duration, and inter­phase separation for individual biphasic pulses could be changed during the experiment, as could the frequency and duration of trains of pulses. Stimulus Parameters Several different types of waveforms were used: they all had an initial rectangular stimulating (cathodal) pulse while the charge balancing wave had different shapes. The charge balancing waveform tends to repolarize the cell membrane, counteracting the effect of the cathodal stim­ulus [7], [14], [20]. To lessen this effect we tried two solutions: make the anodal pulse smaller in amplitude and/ or have it occur after a delay. We studied whether differ­ent anodic shapes and phase separations influenced the stimulus-force relationship. In all the waveforms studied, the stimulating and balancing phases injected equal amounts of charge. The anodal waveforms included simple rectangular, in which the balancing pulse had the same amplitude and duration as the stimulating pulse; balancing rectangle of duration four or eight times the stimulus duration, in which case the amplitude was adjusted proportionately; and triangular balancing phase of duration three or seven times the stimulus duration. The results showed that there were no statistically significant differences between the different anodal waveforms in the force of evoked muscle contractions. On the other hand, separation between cathodal and an­odal waveforms did have an effect. There was a small (on the order of 10-15%), but statistically not significant, in­crease in force with a phase separation of 500 fis com­pared to unseparated waveforms. Separations longer than 1000 /xs showed no further increase. One might argue that the increased force generated by separated waveforms is due not to the effects described above, but to activation of a second set of action potentials by the balancing pulse. However, we, like others [14], found that the biphasic stimuli were not more effective than simple, monophasic stimuli, ruling out this possibility. For the results presented below, we used simple rect­angular waveforms with an interphase separation of 1 ms. In order to minimize muscle fatigue and the time re­quired by the experiments, recruitment curves were pri­marily studied with single pulses. The results obtained with the single pulses were periodically tested with 50 Hz, 300 ms duration pulse trains. In the former case, peak twitch force was the measure used in determining recruit­ment. For pulse trains, plateau force was the measure used. To control for muscle fatigue, we periodically tested the muscle with a standard stimulus. The standard stim­ulus was selected early in each experiment as that stimu­lus which produced approximately 50% of peak contrac­tion force. If the contraction force remained the same, we knew that the muscle was not undergoing fatigue. If the force changed, we suspended further testing until the muscle had recovered. Recruitment curves were obtained by varying the stim­ulus intensity between the threshold value and the value that gave maximal force production. This was done either by increasing the current amplitude while the stimulus du­ration was maintained constant or by changing the dura­tion while the current remained constant. Advantages of one method with respect to the other were examined. The main problem we faced was variability in contrac­tion responses not only between animals, but within a sin­gle experiment as well. This is a common factor in this type of experiment [10], [12] and is influenced by phys­iological changes which are not completely understood or identified. In spite of this, the data are consistent when normalized and averaged. Results Preload Effects Recruitment curve measurements showed that as long as the preload was high enough, the shape of the recruit­ment curve did not change over a wide range of preloads. This proved to be true for both muscles and allowed us to use only one preload for each experiment. Preload was applied by changing the rest length of the muscle. The chosen length was the one at which the force produced above the preload was maximal, but was always within the normal physiologic range of the muscle. Voltage-Current Relationship A series of tests was performed to study the electrical characteristics of the electrodes both in vivo and in saline solution after the electrode had been explanted. Both tests gave similar results, and the equivalent circuit for the range of current densities and pulse durations used in the animal experiments is presented in Fig. 1. The electrode could be modelled by a series resistor (Re) and capacitor (Ce), and the tissue by a simple resistance. Re was small, so the behavior of the system was dominated by the elec­trode capacitance and tissue resistance. The relationship between stimulus duration and ampli­tude for activation of single nerve fibers was shown by Hill in 1936 [15], [16] to be I, = Ir/(1 - e-'h L/cparuiiciii, ic o n , ruuiiMimg ocrvuxs, nucs i-anu. r .v ; dux i j j i , riscaiaway, inj 0 Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY and at adc li i jyrigni iy y i dv m e institute 01 tiectricai ana itional mailing offices. Postmaster: Send address 10 ' NANNINI AND HORCH: MUSCLE RECRUITMENT WITH INTRAFASCICULAR ELECTRODES c Q) D O Duration (t/7") Fig. 3. Strength-duration curve (soleus muscle). The current needed to ob­tain a midrange force value is plotted versus duration for three different experiments. The current was normalized with respect to the rheobase and the duration to the time constant for each experiment. The theoretical curve from Hill's equation is superimposed. pulse duration modulation. In each plot, force has been normalized to the maximum value obtained with ampli­tude modulation, while charge has been normalized to the amount needed to produce 50% of the maximal force in each case. Fig. 5 shows two recruitment curves obtained with dif­ferent pulse durations: 50 ns (circles) and 20 /xs (dia­monds). In Fig. 5(a) the force is normalized to the max­imum value obtained with the 50 /us duration and is plotted versus the current. With the shorter duration stimulus, the current required to obtain the same force is higher, but the slope of the curve is less steep. If the force is plotted versus charge [Fig. 5(b)], the two curves are substantially the same and the charge required is slightly lower for the curve with shorter duration stimuli. Pulse Trains Recruitment curves obtained with 50 Hz pulse trains showed no significant differences, other than expected in­creases in force amplitude, from the curves obtained with single pulses. Effect of Nerve Crush Data were also collected when the spinal reflex had been suppressed by crushing the nerve proximal to the stimulus site. With both single pulses and trains we had two classes of results: in one there was no significant difference be­tween the curves obtained in normal and crushed situa­tions [Fig. 6(a) and (b)], in the other the curves were shifted to the right after nerve crush [Fig. 6(c) and (d)]. As both types of behavior were seen in about the same number of cases, we could not conclude that one of the two was a result of measurement errors. The same effect was also seen in the soleus muscle. Discussion Control of Force Although different types of force modulation have been proposed [1], [24], most neuroprostheses increase muscle force by recruiting an increasing number of motor units. This is accomplished by increasing the charge injected, modulating either the current amplitude or the pulse du­ration. In our experiments, duration and amplitude mod­ulation did not show any significant differences in terms of charge requirements (Fig. 4). The choice of one over the other is therefore dictated by other factors. There are at least two important reasons for suggesting that the parameter to minimize in optimizing stimulation control may be charge. First, electrochemical studies of the behavior of materials used for stimulation show that damage is a function of the charge injected in each pulse and safe charge values for several materials have been determined [2]-[5]. Second, our long term goal is to use this electrode as a chronic implant in patients. As this will probably be associated with a battery power supply, it is important to limit the charge required in each pulse to maximize battery life. From these considerations, we should choose short pulse durations that minimize charge. Moreover, short durations have the advantage of producing a less steep recruitment curve when current is used as the control pa­rameter. Fig. 5(b) shows that with the shorter duration stimuli the charge requirement is slightly lower, while the two curves have almost identical shapes when force is plotted as a function of charge. But, the slope of the re­cruitment curve obtained with shorter duration stimuli is flatter when the force is plotted versus current [Fig. 5(a)]. This reflects the possibility of a finer force control with shorter duration stimuli, as has been noted before [10]. L/cparuiiciii, id e e ruuiiMiuig ocrvitcs, *+*+j nucs i-anu. r.v; dux i jj i , riscaiaway, inj 0 Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY and at adc I, i jyrignt iy y i dv m e institute 01 tiectricai ana itional mailing offices. Postmaster: Send address NANNINI AND HORCH: MUSCLE RECRUITMENT WITH 1NTR A FASCICULAR ELECTRODES S 150 I £ a __ 300 200 100 125 r ioo . __" S i E 75 . / " n 3 1 / 4> 50 ■ P 0 L. T d 0 ^ U. 25 / O d Current (uA) (a) 100 Current (uA) (C) 25 50 75 Current (uA) (b) 50 100 150 Current (uA) (d) 200 Fig. 6. Effect of nerve crush (gastrocnemius muscle). Plotted are recruit­ment curves from four different preparations obtained before (filled sym­bols, solid line) and after (open symbols, broken line) crushing the nerve proximal to the stimulation site. Parts (a) and (b) show no significant dif­ference between the two curves, suggesting that in these experiments the spinal reflex was not an important recruitment factor. In (c) and (d), the recruitment curve after nerve crush was shifted to the right, suggesting that the spinal reflex was an important recruitment tool. The data in (a) and (c) were obtained with single pulse stimuli, while the data in (b) and (d) were collected using 50 Hz trains of pulses. The force axis is plotted in terms of gram-force (1 g-wt = 9.81 * X0 1 N). The present experiments demonstrate that they are also efficient in evoking neural activity. One salient feature of our electrode design is its relatively large (1 mm) stimu­lation zone. This corresponds to an area of about 0.8 mm2 and means that small leakage paths due, for example, to breaks in the insulation will not cause significant shunting of current, which could affect the electrode performance. The gassing limit charge density for platinum elec­trodes has been shown to be 300 /.iC per cm2 geometric surface area [2]-[5], [21], [27], In our electrodes, this corresponds to a pulse charge of 2.4 /xC. The median of the charge required to produce half maximal force in the present experiments was 3.75 * 10"3 ^C, almost three orders of magnitude below this limit, and well below the safe limit for stimulation of neural tissue [2], An upper limit for current density has been suggested as 300 mA/cm2 [3]-[5], [21]; for our electrode this cor­responds to 2.4 mA. The maximum current needed in the present experiments was 200 /xA, over one order of mag­nitude below this level. Conclusions This paper shows that intrafascicular recording elec­trodes are also suitable for motor nerve stimulation. Ef­ficient recruitment of muscle force can be generated with these electrodes using biphasic, charge balanced, 50 /xs duration pulses separated by about 500 fxs. Current levels are significantly lower than those needed with either intramuscular or cuff electrodes. This reduces the amount of charge required, with a consequent possible increase in safety and decreased power supply require­ments for portable clinical stimulators. Charge density and current density requirements are well within the range considered safe for stimulation of neural tissue. Acknowledgment We thank T. Lefurge for assistance with preparation of the electrodes, and Dr. R. Normann and Dr. J. Wood for review of earlier versions of the manuscript. References [1] N. Aconero, G. Bini, G. L. Lenzi, and M. Manfredi, "Selective ac­tivation of peripheral nerve fibre groups of different diameter by tri­angular shaped stimulus pulses," J. Physiol., vol. 273, pp. 539-560, 1977. ' [2] W. F. Agnew, D. B. McCreery, T. G. H. Yuen, and L. A. Bullara, L/cparuiiciii, icon, ruuiiMimg ,3civiccs, hhj nucs i-anu. r.v ; dux i j j i , riscaiaway, inj 0 Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY and at adc I, i jyrigni iyyi dv me institute or tiectricai ana itional mailing offices. Postmaster: Send address IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING. VOL. 38, NO. 8. AUGUST 1991 "Histologic and physiologic evaluation of electrically stimulated pe­ripheral nerve: Considerations for the selection ot parameters," Ann. Biomed. Eng., vol. 17, pp. 39-60, 1989. , [3] S. B. Brummer, J. McHardy, and M. J. Turner. "Electrical stimu­lation with Pt electrodes: Trace analysis for dissolved platinum and other dissolved electrochemical products," Brain. Behav Evol vol 14, pp. 10-22, 1977. ' ' [4] S. B. Brummer and M. J. Turner, "Electrochemical considerations for safe electrical stimulation of nervous system with platinum elec­trodes," IEEE Trans. Biomed. Eng., vol. BME-24, pp 59-63 Jan 1977. ' ' [5] . ‘‘Electrical stimulation with Pt electrodes: II-estimation of maximum surface redox (theoretical non-gassing) limits." IEEE Trans. Biomed. Eng., vol. BME-24, pp. 440-443, Sept. 1977. [6] C. W. Caldwell and J. B. Reswick, "A percutaneous wire electrode for chronic research use," IEEE Trans. Biomed. Eng., vol. BME-22. pp. 429-432, 1975. [7] N. E. Cotter, "Modeling of auditory prostheses," Tech. rep. No G906-8, Stanford electronic laboratories, Stanford Univ., Stanford, CA, June 1986. [8] P. E. Crago, J. T. Mortimer, and P. H. Peckham. "Closed-loop con­trol of force during electrical stimulation of muscle," IEEE Trans. Biomed. Eng., vol. BME-27, pp. 306-311. June 1980. [9] P. E. Crago, J. T. Mortimer, P. H. Peckham. and H. J. Chizeck, "Authors' reply," IEEE Trans. Biomed. Eng., vol. BME-34 pp 261-263, Mar. 1987. ' ' [10] P. E. Crago, P. H. Peckham, J. T. Mortimer, and J. P. Van Der Meulen, "The choice of pulse duration for chronic electrical stimu­lation via surface, nerve, and intramuscular electrodes," Ann. Biomed. Eng., vol. 2, pp. 252-264, 1974. [11] P. E. Crago, P. H. Peckham, and G. B. Thorpe, "Modulation of muscle force by recruitment during intramuscular stimulation," IEEE Trans. Biomed. Eng., vol. BME-27, pp. 679-684. Dec. 1980. [12] W. K. Durfee and K. E. MacLean. "Methods for estimating isomet­ric recruitment curves of electrically stimulated muscle," IEEE Trans. Biomed. Eng., vol. 36, pp. 654-667, 1989. [13] J. Duysens and R. B. Stein, "Reflexes induced by nerve stimulation in walking cats with implanted cuff electrodes," Exp. Brain Res.. vol. 32, pp. 213-224. 1978. [14] P. H. Gorman and J. T. Mortimer. "The effect of stimulus parameters on the recruitment characteristics of direct nerve stimulation." IEEE Trans. Biomed. Eng., vol. BME-30, pp. 407-414, 1983. [15] A. V. Hill, "Excitation and accommodation in nerve," in Proc. Royal Soc. Lond. Biol., vol. 119, pp. 305-335. 1936. ' 1161 -----• "The strength-duration relation for electric excitation of med- ullated nerves," in Proc. Royal Soc. Lond. Biol., vol 119 pp 440­453, 1936. ' ' [17] J. A. Hoffer, A. A. Caputi, I. E. Pose, and R. 1. Griffiths, "Roles of muscle activity and load on the relationship between muscle spindle length and whole muscle length in the freely walking cat." Prog. Brain Res., vol. 80, pp. 75-85, 1989. [18] K. Horch, "Central responses of cutaneous sensory neurons to pe­ripheral nerve crush in the cat," Brain Res., vol. 151. pp. 581-586 1978. [19] M. S. Malagodi, K. W. Horch, and A. A. Schoenberg, "An intra­fascicular electrode for chronic recording of action potentials in pe­ripheral nerves," Ann. Biomed. Eng., vol. 17, pp. 397-410, 1989. [20] J. T. Mortimer, "Electrical excitation of nerve." in Neural Prostheses, W. F. Agnew and D. B. McCreery. eds. Englewood Cliffs, NJ: Prentice Hall, 1990, ch. 3, pp. 67-83. [21] L. S. Robblee and T. L. Rose. "Electrochemical guidelines for se­lection of protocols and electrode materials for neural stimulation," in Neural Prostheses, W. F. Agnew and D. B. McCreery. eds. En­glewood Cliffs, NJ: Prentice Hall, 1990, ch. 2, pp. 25-66. 122J M. Solomonow, "Comments on ‘Design and evaluation of digital closed-loop controller for the regulation of muscle force by recruit­ment'," IEEE Trans. Biomed. Eng., vol. BME-34, pp. 260-261 Mar 1987. [23] M. Solomonow, R. Baratta. B. H. Zhou, H. Shoji, and R. B. D'Am- brosia, "The EMG-force model of electrically stimulated muscle: Dependence of control strategy and predominant fiber composition," IEEE Trans. Biomed. Eng., vol. BME-34, pp. 692-703, Sept. 1987. [24] M. Solomonov, E. Eldred, J. Lyman, and J. Foster, "Control of muscle contractile force through indirect high-frequency stimula­tion," Amer. J. Phys. Med., vol. 62. pp. 71-82. 1983. [25] R. B. Stein, D. Charles, L. Davis. J. Jhamandas, A. Mannard, and T. R. Nichols, "Principles underlying new methods for chronic neural recording," Can. J. Neurol. Sci.. no. 2, pp. 235-244, 1975. [26] G. F. Wilhere, P. Crago. and H. J. Chizeck, "Design and evaluation of a digital closed-loop controller for the regulation of muscle force by recruitment," IEEE Trans. Biomed. Eng., vol. BME-32, pp. 668­676, Sept. 1985. [27] T. G. H. Yuen, W. F. Agnew, L. A. Bullara, S. Jacques, and D. B. McCreery. "Histological evaluation of neural damage from electrical stimulation: Considerations for the selection of parameters for clini­cal applications." Neurosurg., vol. 9, pp. 292-299, 1981. [28] B. Zhou, R. Baratta, and M. Solomonow. "Manipulation of muscle force with various firing rate and recruitment control strategies." IEEE Trans. Biomed. Eng., vol. BME-34, pp. 128-139, Feb. 1987. Nicola Nannini was born in Florence, Italy, in 1960. He received the Laurea in electrical engi­neering from the University of Florence in 1987, and the Master of Science in Bioengineering from the University of Utah, Salt Lake City, in 1989. As a Research Assistant he was involved in the University of Utah project in neuroprostheses. He is currently working as a consultant for Italian hospitals in the field of biomedical technology management. His research interests include power spectrum analysis of biomedical images and neuroscience. Kenneth Horch (M'88) was born in Cleveland, OH, in 1942. He received the B.S. degree from Lehigh University, Bethlehem, PA, and the Ph.D. degree from Yale University, New Haven, CT. He is an Associate Professor of Physiology and Bioengineering at the University of Utah, and Di­rector of Neurological Testing at Topical Testing, Inc. His research activities are in the areas of neu­roprosthetics, peripheral nerve regeneration and repair, sensory evaluation, and somatosensory physiology. L/wpai miviiii i uuiuiiiiig uvi Tiwi}) nuvo ijau^, i ,vy. uua t j j i , i uvaiana^1, u o o j j * u j i , _ _ Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY and at adc 1, i jjiigm \s/ i yy\ uv me umiiuic ui mccincai anu itional mailing offices. Postmaster: Send address