Endpoint Force Field Types

Dorn Spinal Therapy

Spine Healing Therapy

Get Instant Access

FIGURE 4.5 Endpoint force-field types produced by intraspinal myostimulation. All active responses were of one of the types shown: flexion withdrawal, rostral flexion, caudal extension, and rostral extension. EMG activity indicated that the fields were produced by coactivation of muscles (knee flexor, knee extensor, ankle flexor, and ankle extensor), rather than by activation of a single muscle.

FIGURE 4.5 Endpoint force-field types produced by intraspinal myostimulation. All active responses were of one of the types shown: flexion withdrawal, rostral flexion, caudal extension, and rostral extension. EMG activity indicated that the fields were produced by coactivation of muscles (knee flexor, knee extensor, ankle flexor, and ankle extensor), rather than by activation of a single muscle.

It is well established that the spinal cord is organized with many crossed reflexes and connections. Therefore, we also mapped the spinal cord contralateral to the instrumented limb. The responses evoked by contralateral stimulation were primarily extension responses: caudal (5 of 16 fields) or rostral (7 of 16 fields), and were more

CICI n TVDC niCTDIDI ITIOM DV RCDTLI

CICI n TVDC niCTDIDI ITIOM DV RCDTLI

Types Electrical Stimulation

FIGURE 4.6 Field type distribution by depth for intraspinal microstimulation applied to the side of the cord ipsilateral to the limb (top graph), and contralateral to the limb (bottom graph). The height of the bar represents the number of a particular field type found in the depth interval indicated on the horizontal axis. Flexion withdrawal was the most frequent response observed for ipsilateral stimulation, while extensor responses were more frequent for contralateral stimulation.

3000 4000 depth (|nm)

FIGURE 4.6 Field type distribution by depth for intraspinal microstimulation applied to the side of the cord ipsilateral to the limb (top graph), and contralateral to the limb (bottom graph). The height of the bar represents the number of a particular field type found in the depth interval indicated on the horizontal axis. Flexion withdrawal was the most frequent response observed for ipsilateral stimulation, while extensor responses were more frequent for contralateral stimulation.

readily evoked by stimulation in the dorsal aspects of the cord as shown in Figure 4.6, which illustrates the response distribution by depth of penetration. No new types of responses were identified, nor were any contralateral responses classified as rostral flexion (1 of 54 on the ipsilateral side). Interestingly, the contralateral extensor responses exhibited a point of zero active force within the measured workspace. A number of contralateral rostral extension responses showed a point of convergence in the workspace, while none of the flexion responses did. Since the extension responses were evoked when stimulating in the more dorsal aspects of the cord, it is likely that they are related to the crossed-extension reflex and may be due to afferent activation. However, responses evoked in the dorsal aspects of the cord (both ipsi- and contra-lateral) may also be due to activation of interneurons, as Tresch and Bizzi53 demonstrated that responses evoked by dorsal intraspinal microstimulation are preserved in deafferented rats. The stimulation may be activating the afferent terminals, interneurons, or most likely a combination of both. Based on the results of Tresch and Bizzi,53 it seems likely that interneurons are part of the circuitry responsible for the force patterns.

4.3.4 Perspective and Conclusions

The results of the high-resolution mapping studies identified regions of the lumbar spinal cord that generate hindlimb motor responses when electrically stimulated. The maps demonstrated spatial segregation of responses in the ventral horn that correlated well with previous anatomical tracing studies.54,55 The endpoint force measurements revealed that intraspinal microstimulation can generate organized, multiple-joint, multiple-muscle motor responses, which are not obtained by direct stimulation of muscles or by intraspinal stimulation of motor neurons. Stimulation in the dorsal aspect of the cord generated ipsilateral flexion and contralateral extension. These results suggest that electrical microstimulation can generate two classical spinal reflexes: flexion withdrawal and crossed extension.

Organized motor behaviors evoked by microstimulation support the hypothesis that intraspinal microstimulation and activation of spinal interneuronal circuitry may simplify neural prosthetic control of complex motor behaviors.22 Control of endpoint behavior via activation of individual muscles is possible, but the complexity of the hardware and software necessary to perform the task increases significantly, and systems must be devised to coordinate the action of the controllers around each joint. The spinal circuitry coordinates the activation of multiple limb muscles, and this coactivation gives the end-point an inherent stability that is not affected by either controller delay loops or muscle activation time.

Since this work was conducted in spinal intact animals, further research is necessary to establish the effects of cellular and network plasticity on the evoked motor responses. The results of Tresch et al. in rats53 indicate that responses are evoked at lower stimulation thresholds in chronically spinalized animals, but the patterns of responses are remarkably similar to those we measured in the cat. Possible differences in the structure of the fields between intact and spinalized animals have not been investigated, but it seems likely that the release of the supraspinal influences would affect the threshold, time course, and relative distribution of the fields, rather than their structures. The development of neural prostheses based on the technique of intraspinal microstimulation awaits answers to these questions, as well as further research into the dynamic and control properties of the evoked motor responses.

4.4 HINDLIMB MOTOR RESPONSES INDUCED IN AWAKE BEHAVING CATS*

4.4.1 Background

Electrical stimulation of muscles or nerves has long been used for the purpose of restoring motor function following spinal cord injury, head trauma or stroke.56 Neuroprostheses based on this approach have been reasonably successful especially in the areas of diaphragm pacing57 and foot-drop correction.58 But restoring locomotion using this approach has proved to be rather difficult.18,59 We proposed the use of electrical microstimulation of the lumbosacral spinal cord for improving motor function following injury for several reasons.60,61 The spinal cord is distant from contracting muscles, so electrodes implanted therein would be subjected to significantly less stress and strain than epimysial or nerve cuff electrodes. The region of the spinal cord containing the cell bodies of motor neurons innervating the lower extremity muscles is relatively small (~5 cm in humans). Therefore, by implanting electrodes in a compact and protected region, most of the muscles in the lower extremities could be accessed. Finally, tapping into the movement control centers in the spinal cord would allow for the activation of intact subsets of locomotor networks. However, for various reasons, it is still somewhat debatable whether spinal cord microstimulation (SCpStim) would generate functional and controllable movements in people with spinal cord injury.

4.4.2 Acute Mapping of the Lumbosacral Spinal Cord

In 1992 the University of Utah Neuroprosthetics Laboratory initiated the investigation of the feasibility of SCpStim for restoring functional movements following spinal cord injury.60 The long-term objective of the study was to develop and evaluate an electrode array and a stimulator suitable for chronic implantation and stimulation of motor pathways in the mammalian lumbosacral spinal cord. Prior to this, spinal cord segmental and intersegmental networks involved in producing stereotyped and programmed reflexive movements had been extensively studied using methods ranging from intracellular recordings to epidural spinal cord stimulation and systemic administration of locomotor-inducing drugs. Yet, the ability to generate fractionated, controllable movements by electrically stimulating the spinal cord had not been demonstrated. We posited that the generation of fractionated movements is essential for restoring functional mobility in paraplegia since such movements allow for the performance of simple daily acts like rising from a wheelchair, getting in and out of bed and immediately correcting for unanticipated obstacles. Generation of fractionated movements requires that individual muscles, or synergistic muscle groups, are activated in isolation and that the produced forces are graded through a large

* By Vivian K. Mushahwar and Arthur Prochazka.

range of stimulus strengths. Furthermore, functional movements require the maintenance of muscle force for relatively long periods without excessive muscle fatigue.

To determine whether selective activation of muscles can be obtained using SCpStim, the lumbosacral portion of the spinal cord was mapped in pentobarbital-anaesthetized adult cats, and locations generating activation within the main knee and ankle flexor and extensor muscles (i.e., hamstrings, quadriceps, tibialis anterior, triceps surae/plantaris) were noted. The maps demonstrated that there were regions within the spinal cord from which single muscles (i.e., tibialis anterior) or synergistic muscle groups (i.e., quadriceps) can be activated in isolation.60-63 We referred to these regions as "activation pools." Within these pools, muscles are selectively activated with low stimulus strengths (<40 pA) and the selectivity is maintained with further increases in stimulus amplitude (>2x threshold). The results therefore affirmed that the overlap between various motoneuronal pools within the lumbar cord is minimal and functional muscle groups could be selectively activated using SCpStim. A recent comprehensive horseradish peroxidase map of motoneuronal pools in the lumbosacral spinal cord further confirmed these findings.55

The characteristics of force recruitment using SCpStim were studied by stimulating numerous locations within the activation pool of each of the muscles of interest with varying stimulus strengths and measuring the resulting force. The constructed force recruitment curves demonstrated that smooth and graded control of muscle force could be obtained using SCpStim. Correlations between stimulus strength and duration of force twitches suggested that SCpStim could also produce a near-normal physiological order of motor unit recruitment.63 This comes in contrast to the mixed or reversed motor unit recruitment order observed in peripheral functional neuro-muscular stimulation systems using epimysial or motor point electrodes. Finally, when two electrodes implanted in a given activation pool were stimulated with interleaved pulse trains, muscle fatigue was significantly reduced.64 Interleaved stimulation mimics the natural asynchronous activation order of motor units.59 65 66 Muscle fatigue is significantly reduced because fused contractions of whole muscles are produced without having to sustain tetanic contractions in single motor units by stimulating them at high rates.

The findings of the acute studies affirmed that functional limb movements could be produced through SCpStim. They demonstrated that individual muscles or synergistic muscle groups can be selectively activated; that the generated muscle force can be smoothly graded by varying stimulus strength; and that the rate of muscle fatigue can be reduced using simple interleaved stimulation. Therefore, it was concluded that SCpStim may be feasible for restoring functional mobility in paraplegia, and specifications for a high-density electrode array to be implanted in the spinal cord were proposed based on the dimensions of activation pools and the amount of effective stimulus spread.64 Yet, whether the acute results could be reproduced in awake, chronic preparations remained unclear.

4.4.3 Chronic Fixation Technique

Given the delicate nature of the spinal cord and its ability to move within the vertebral column, special attention was given to electrode design and fixation to the vertebrae and dura mater. Floating microwires were used and the techniques of their implantation and stabilization were based on those developed for chronic recordings from single neurons in spinal dorsal roots.67 Healthy adult cats were anaesthetized with sodium pentobarbital and the hindlimbs and back were shaved. A skin incision was made from the L4 to L6 vertebral spinous processes and the dorsal surface of vertebra L5 was removed to expose spinal cord segments L5 to S1. Six to twelve microwire electrodes contained within a silastic tube and attached to a connector on one end were passed subcutaneously from the back incision to the head. The wires were 30-pm diameter stainless steel or platinum-iridium (80%-20%) with 30-70 pm tip exposure and 10-30 kQ impedance. They were pre-cut and bent to an appropriate angle and length. The head connector was embedded in dental acrylic secured to the skull by screws and the silastic tube was anchored to the L4 spinous process.67 The microwires emerging from the tube were tethered to the dura mater with 8/0 ophthalmic sutures and further bonded with cyanoacrylate glue. A 5-cm length of bared multistrand stainless steel wire was used as the indifferent electrode and was placed in the back muscles close to the vertebral column. The microwires were inserted through the dura mater on both sides of the spinal cord, 2 mm from the midline. Their placement was based on maps established in the acute experi-ments,62,63 and stimulation through each electrode was used to guide final positioning. Once inserted in the spinal cord (to a depth of 3.5 to 4.5 mm), the epidural portions of the wires lay flat on the dura mater (Figure 4.7). A small piece of plastic thin film, serving as a barrier to the growth of connective tissue in and around the electrodes, was spot-glued over the dura mater and microwires with cyanoacrylate. The back wound was sutured closed in layers and the cats recovered in an intensive care unit for 1 to 2 days. Animals received doses at 12-hour intervals of a strong opioid analgesic (buprenorphine) during the recovery period. Animals were maintained from 2 to 24 weeks following surgery. Twice a week throughout the duration of implantation, the cats were stimulated through individual spinal cord microwires. Visual inspection, palpation and intramuscular EMG recordings were used to detect muscle activation and the generated movements were noted.

4.4.4 Types of Movements Evoked by Spinal Cord Microstimulation

In the chronic experiments, we were faced with two challenges. First, very few electrodes (six per side) were used to target activation pools spanning a 30-mm length of lumbosacral spinal cord. On average, individual motoneuronal pools are <1mm in diameter and ~10 mm in length.61 6264 The acute experiments indicated that several electrodes implanted in a single pool and stimulated in a patterned manner produce optimal recruitment of whole muscles while maintaining selectivity of activation, smoothness of force generation and minimal fatigue. Second, in awake, behaving animals, the spinal cord is more excitable than in pentobarbital-anaesthe-tized preparations. Given the large number of interneurons in the cord ventral horn, it was questionable whether SCpStim could generate coordinated and useful limb movements. Yet, functional movements were generated when spinal cord microwires were stimulated in the awake animals (nine implanted to date). Taken collectively, 60% of the implanted electrodes predominantly generated movements across a single

FIGURE 4.7 Microwire Implantation Technique. The microwires targeted motoneuron pools located in the ventral horn of the spinal cord. Their epidural portions lay flat on the spinal cord and were tethered to the dura mater with a fine ophthalmic suture. The silastic tube containing the microwires was anchored to the L4 spinous process.

FIGURE 4.7 Microwire Implantation Technique. The microwires targeted motoneuron pools located in the ventral horn of the spinal cord. Their epidural portions lay flat on the spinal cord and were tethered to the dura mater with a fine ophthalmic suture. The silastic tube containing the microwires was anchored to the L4 spinous process.

joint throughout the duration of the experiments. Stimulation through 30% of the electrodes generated whole limb synergies (involving hip, ankle and knee) with torques large enough to lift the animals' hindquarters. The final 10% of the implanted electrodes elicited cocontraction of multiple muscles resulting in stiffening of one or two joints without producing any net movement. Figure 4.8 gives examples of the single-joint and whole limb synergistic movements generated with SCjuStim in awake animals. In (A) and (B), knee extension and ankle dorsiflexion were generated primarily due to the activation of quadriceps and tibialis anterior, respectively. In some cats, quadriceps contractions generated knee torques up to 1 Nm (an estimated torque of about 0.6 Nm is required to support the hindquarters in stance). In (C), stimulation through a single electrode produced a powerful hip-knee-ankle extensor synergy capable of bearing the weight of the animal's hindquarters. These whole limb synergies may represent the force-field primitives that arise from focal stimulation of premotoneuronal areas in the spinal cord.10 Several electrodes implanted on both sides of the spinal cord generated synergies similar to the one shown in Figure 4.8C. Therefore, patterned stimulation through as few as two or three electrodes on each side of the spinal cord could, in principle, restore simple functional activities such as maintaining a standing posture in human spinal cord injured subjects.

Was this article helpful?

0 0

Post a comment