Skip secondary menu

Functional and Dysfunctional Spinal Circuitry: Role for Rehabilitation and Neural Prostheses

Chairman: Allan Tobin, University of California, Los Angeles
Organizer: Arlene Chiu, National Institute of Neurological Disorders and Stroke

June 14-15, 2000

Table of Contents


I. Introduction

A wealth of new tools is paving the way for the development of more effective rehabilitation practices and neural prostheses for the treatment of spinal cord injury (SCI). The development of microelectrode arrays capable of selectively stimulating small functional sets of cells, for example, provides hope that we may, in the future, be able to replace the activity lost with damage to neurons. On the other hand, the use of rehabilitative protocols that provide weight-loading and sensory feedback highlight the spinal cord's adaptability and its capacity for re-training. Both avenues could lead to restoration of functional capabilities, such as locomotion and bladder control, while minimizing invasive surgery and risk.

Much remains to be done however, before the full potential of these and other advances are translated into patient benefits. Gaps in our knowledge regarding the function of healthy spinal circuitry prevent us from optimizing strategies for harnessing natural mechanisms of plasticity and repair. In addition, a scarcity of objective and quantitative outcome measurements in both animal models and patient trials have hindered careful assessment of candidate therapies and comparisons between studies. Finally, critical information, such as the priorities of patients and the relative tractability of their symptoms, has been overlooked at times, due to a lack of communication between patients, clinicians, and basic scientists.

In the second workshop in the series "New Strategies in Spinal Cord Injury," held in Los Angeles, CA in June 2000, the National Institute of Neurological Disorders and Stroke (NINDS) took a first step towards addressing these issues. Workshop participants discussed and assessed strategies to replace the function of damaged neurons, facilitate their regeneration, and retrain the injured spinal cord. They also discussed opportunities for cross-fertilization between the study of SCI and space physiology. The National Aeronautics and Space Administration (NASA) co-sponsored the workshop, recognizing that astronauts exposed to microgravity experience some of the same physiological alterations that affect SCI patients.

 II. Background

Understanding the spinal cord's structure and function lies at the heart of developing technologies that aim to replace, and/or retrain neural function following SCI. As the mediator of communications between the body and the brain, the spinal cord shuttles sensory information from neurons distributed throughout the body to the brain, and motor instructions from the brain to the limbs and organs. Far from being a mere conduit of information, however, the endogenous circuitry within the spinal cord plays crucial roles in coordinating complex behaviors. For example, the spinal cord harbors a central pattern generator (CPG) -a group of neurons that generates oscillatory patterns of activity and coordinates movements required for locomotion. The CPG is not simply a hard-wired clock or pacemaker that sets a fixed pace for locomotion; it is modulated by sensory feedback which plays a key role in triggering successive movements. In addition, the spinal cord appears to contain collections or modules of neurons that activate specific groups of muscles, and allow the CPG, as well as supraspinal and reflex pathways, to translate their signals into actions. In amphibians and some mammals, different combinations of these modules, also known as "primitives", appear to be capable of generating a large range of complex movements.

One of the most exciting attributes of the spinal cord is its ability to adapt to change, and, when appropriately harnessed, it has the potential to overcome functional disruptions caused by SCI. Even when completely disconnected from the brain, the spinal cord is capable of learning and storing memories. Although the cellular and biochemical mechanisms underlying this "plasticity" or ability to change are not well understood, recent studies have shown that spinal cord neurons can undergo long-term potentiation, and long-term depression, two mechanisms believed to underlie memory and learning in the brain. Growth factors are some of the cues that may mediate adaptive changes in the spinal cord. Several growth factors, including Brain Derived Growth factor (BDNF), Neurotrophin-3 (NT-3) and Neurotrophin-4/5 (NT-4/5) are synthesized by neurons in the adult spinal cord. When stimulated by the glutamate agonist, kainic acid, subsets of these neurons show specific increases in their expression of BDNF and NT-4/5, suggesting that signaling between neurons controls the availability of these factors, that, in turn influence the properties of local circuits within the cord.

Changes or plasticity within the spinal cord does not always result in improved function. Working with larval lampreys, Avis Cohen identified dysfunctional changes that can result from SCI. These animals demonstrated functional recovery when allowed to recuperate at room temperature from experimentally-induced crushing of their spinal cords,. Yet when recovery took place at 40C - which is closer to the ambient temperature at which lampreys live in the wild - Cohen discovered that the coupling of oscillatory activities between segments of the cord was disrupted. These studies show that altered coupling during recovery can wreak havoc in a simple system such as lamprey swimming. The fact that a relatively minor difference in recovery conditions had such a dramatic effect on the outcome underscores the difficulty in predicting the response to injury in higher vertebrates with more complex networks of coupled oscillators.

There are a multitude of injury-induced changes that have consequences upstream and downstream of the spinal cord. Imaging studies of the brain, for example, reveal changes in the cortical maps of patients with SCI indicating a functional reorganization of their primary sensorimotor cortices. In addition, as muscles become deprived of their normal synaptic inputs and weight-bearing activity, they begin to atrophy and undergo changes in the complement and levels of many muscle-specific proteins, altering their functional properties. These alterations, in turn, have retrograde effects on the metabolic state and health of the motor neuron innervating the muscles. The lack of weight-bearing activity also affects bones, causing them to de-mineralize. In summary, SCI affects normal physiology well beyond the central nervous system. 

III. Workshop Discussion

IIIA. Tools for Extending Basic Understanding
Animal models

Although rats, cats, and lampreys provide powerful models for examining particular aspects of SCI, additional animal models are needed. Studies in larger animals, such as primates, are needed to mimic the human disorder more closely. Also, some researchers are turning to mouse models to take advantage of the recent revolution in genetics and the availability of mutant animals. The need for developing chronic models of injury was also stressed. The majority of current studies focus on acute SCI which do not reflect the status of the tens of thousands of patients who have lived with SCI for years.

Spinal cord circuitry

Circuitry maps describing functional connectivity within the normal spinal cord are greatly needed if we are to understand how the spinal cord is altered by injury. To address this need, one approach is the use of neurotropic viruses, such as the pseudorabies and herpes simplex viruses, which invade permissive cells, replicate, and move to infect other neurons tran-synaptically, that is, from one neuron to the next connected by active synapses. The viruses act as self-amplifying tracers that light up specific pathways or circuits. Viral strains vary in their direction of motion - retrograde or anterograde - and in their ability to infect different classes of neurons. By choosing the appropriate viral strains, it is possible to map a wide variety of neuronal circuits.

Reginald Edgerton applied this technique to compare the connections of soleus motor neurons between healthy controls and injured animals. By back-label motor neurons from identified muscles, he found that the virus spread to more neurons in injured animals than in healthy controls. The increased distribution may be due to the establishment of new synapses through sprouting as a result of the injury, or to the appearance of gap junctions between motor neurons. Other groups have observed the formation of gap junctions between lumbar spinal motor neurons in the adult cat following injury to the peripheral nerve. This electrical coupling may be a compensatory or protective mechanism that helps keep motor neurons alive until they reestablish synaptic connections.

Lea Ziskind-Conhaim is employing a different method - optical imaging - to examine spinal cord circuitry. To understand how the CPG is set up during development, she uses voltage-sensitive dyes and large arrays of photo detectors to monitor the simultaneous activities of large groups of neurons in real-time. Both sides of the cord fire in synchrony early in development, but begins to fire in alternating bursts at the time when descending fibers arrive from the brain and when afferent, sensory inputs arrive from the periphery. This correlation suggests that these inputs, from the brain and/or from the periphery, influence the endogenous pacemaker circuits in the cord. She plans to use semi-intact preparations, including the brainstem, spinal cord, and hindlimbs, to dissect the system's functional modules in a systematic fashion, and identify the regions of the spinal cord where oscillations originate. Finally, she will extend her imaging findings with electrical recordings to pinpoint the individual neurons involved in the process.

IIIB. Therapeutic Strategies
a. Replacing cell activity

Cell transplantation is one of the strategies that could help replace lost function in SCI. Animal studies have shown that multipotent neural progenitors and stem cells can integrate into the central nervous system and restore function. For example, stem cells have been used to replace damaged cerebellar Purkinje neurons in mice. Ongoing research suggests that multipotent cells can also be used to replace genes and provide neuroprotective molecules, such as growth factors. Using genetic engineering, it is possible to create stem cells that package and produce vectors carrying specific genes, or create cells that synthesize and secrete the products of inserted genes.

Although less versatile, electrodes offer a potentially safer alternative for restoring lost function because they can be controlled more easily. The time, place, magnitude, and polarity of stimulating currents can be regulated with great precision. Recording electrodes can also be used as sensors to provide feedback for modulating the activities of stimulating electrodes. If spinal cord primitives are identified in humans, a few strategically positioned electrodes could, in principle, control and produce a variety of complex motions, including those involved in locomotion.

Graham Creasey explained that electrical prostheses to directly stimulate bladder and sphincter muscles, or their motor neurons, already provide some control for micturition. Although the prostheses can't coordinate the timing of contractions as precisely as uninjured circuits, they can generate out-of-phase contractions that allow voiding or emptying of the bladder. Because the muscle wall of the bladder contracts at a slower but more sustained rate than the sphincter that prevents release, a somatic nerve reflex can be used to achieve voluntary control of voiding. The somatic efferent is surgically connected to the bladder to trigger micturition by a voluntary act, such as scratching a limb. Identification of a primitive that drives micturition allow the design of potentially even better solutions. William Agnew showed that micturition can be induced with a single, unilateral electrode implanted near the central canal in sacral levels of the cord in normal, anesthetized cats.

Despite these promising results, electrodes have drawbacks. One challenge is to keep them anchored for extended periods of time. In addition, the cables that connect the electrodes to the stimulating and recording devices are cumbersome. Finally, the process of inserting electrodes causes tissue damage and the connective tissue that ultimately encapsulates them can block their function.

Many of these problems have been reduced and new advantages have been gained by the development of microelectrode arrays. Using photo-lithographic technologies adopted from the semiconductor industry, McCreery and others are creating microelectrodes with stimulation radii as small as ten microns that are capable of stimulating single neurons. Since each probe can hold multiple electrodes, a single array can stimulate several sites simultaneously, and the number of wires can be drastically reduced. Because of their increased stability, the arrays are also better than conventional electrodes for long-term applications. Once an array has been designed and optimized, its mass production is easy and cheap. McCreery is now trying to improve the arrays using new materials to make electrodes that are stiffer, longer and more slender; these properties should increase stability and reduce the trauma induced by penetration.

b. Regeneration

Growth factors can be powerful tools for repairing the nervous system because they protect neurons against death and induce them to sprout or regenerate. By administering NT-3, Mark Tuszynski obtained partial recovery of function after a selective lesion of the corticospinal tract. To avoid inducing sprouting in unaffected areas and provide some guidance in regeneration, this approach could be combined with genetic engineering to control local expression of growth factors, or conversely, to down-regulate growth factor receptors in non-target areas. The development of increasingly small, hollow microelectrodes for the targeted delivery of minute volumes may also help localize the effects of specific growth factors. Pico- and nano-liter pumps, with no moving parts, could allow further control by providing both localized and time-regulated delivery.

Protecting neurons and encouraging them to sprout are not the only challenges in promoting regeneration. Cut nerve fibers in the central nervous system often sprout spontaneously, but fail to elongate along their original pathways. Inhibitory factors on the surface of glial cells and in the extracellular matrix contribute to an environment that is inhospitable for regeneration. In contrast, neurons in the peripheral nervous system are capable of retracing their paths and restoring proper connections. Several researchers are attempting to repair injured spinal cords by transplanting segments of peripheral nerve or Schwann cells cultured from peripheral nerves as scaffolding or paths for axons to use. Although these efforts have allowed cut nerve fibers from the spinal cord to regenerate into the transplants, the fibers fail to leave the transplants and re-enter the host nervous system.

An alternate approach was described by Almudena Ramon-Cueto. Noting that neurons in the mammalian olfactory bulb are able to elongate and connect with their targets in adulthood, she transplanted olfactory ensheathing glial cells to promote regeneration in the cord. In contrast to peripheral grafts, the olfactory glia allowed axons of injured central neurons in rats to elongate for long distances well into the segments of the cord caudal to the lesion, accompanied by a striking recovery of function. Paraplegic rats regained locomotor and sensimotor reflexes and were able to move their hindlimbs voluntarily, and respond to touch and propioceptive stimuli applied to their hindlimbs.

c. Retraining

For the past 25 years, rehabilitation practices have focused on compensation rather than recovery. Assistive devices have been used to supplement visible functions - mostly intact voluntary movements - but little effort has been directed at understanding and exploiting the nervous system's capacity for retraining. Recent work has greatly enhanced rehabilitation strategies by demonstrating how weight-loading and propioceptive feedback can lead to recovery of locomotor function.

A team led by Edgerton discovered that mammals with thoracic transections of the spinal cord could relearn to step on a treadmill when they experienced the sensory input normally associated with stepping. They then extended these studies to humans and examined the role of sensory information in modifying the motor patterns of patients suffering from SCI. Using various levels of "loading", they analyzed the movements and electrical activities of the patients' leg muscles during assisted-stepping on a treadmill with body weight support. As suggested by the animal studies, they concluded that weight-loading provides cues that enable the human spinal cord to correct its output in a way that helps stepping.

The results also indicated that traditional rehabilitation practices could be optimized. When a patient is asked to move a joint voluntarily and produces a very low electromyographic response, physicians typically assume that the function is lost, and try to compensate for it with assistive devices. Yet the treadmill studies show that several of these patients can generate robust electromyographic responses when they receive appropriate sensory feedback. Susan Harkema, Bruce Dobkin, Edgerton, and others have applied these findings to retrain the spinal cords of SCI patients, and have, in many cases, succeeded in improving their locomotion.

Edgerton is now exploring the use of robotic training devices to make training more consistent, and to tailor the procedure to the evolving conditions of individual patients. Robotic devices could allow physicians to control critical parameters, such as weight-loading and stepping-speed, accurately and reproducibly. In addition, these devices could help to track patients' progress quantitatively. Paritcipants also noted the possibility of using microelectrode arrays to directly retrain neural circuits.

d. Combination therapies

Most participants predicted that the path to successful treatments would involve a combination of approaches, as illustrated by the work of Hugues Barbeau. He trained spinalized cats on a treadmill after injecting them with the noradrenergic alpha-2 agonist, clonidine. Although clonidine enhances locomotion for only 4-6 hours after administration, it creates a permissive state during which training is much more effective. Consequently, they were able to reduce the three week process of teaching spinalized cats to walk to approximately six days.

Beyond creating permissive states, a combinatorial approach can bolster specificity. Combining electrical stimulation with growth factors, for example, could allow selective modulation of targeted cell populations. Combining cell transplantation with growth factors or anti-myelin antibodies could enhance local regeneration.


IV. Workshop Conclusions and Recommendations

Participants discussed how to strike a balance between accelerated paths that could lead to therapy and careful, robust pre-clinical research, and proposed the following recommendations for advancing SCI research:

IVA. Therapeutic Priorities
a. Promote therapies in advanced stages of development

Findings on weight-bearing training and proprioceptive feedback should be incorporated into rehabilitation regimens and practices. An intermediary phase in clinical trials in which the effects of new candidate therapies are tested in combination with more established approaches, such as retraining rehabilitation should be encouraged.

b. Increase focus on the more tractable problems of SCI

Most SCI research has focused on locomotion -one of the most complex functions affected by SCI. Re-directing some efforts towards more tractable problems, such as micturition and blood pressure control, could yield quicker and more tangible results.

c. Increase focus on patient priorities

Most research efforts have focused on locomotion as an outcome, even though patients often rank sexual dysfunction, bladder control and bowel control as their most immediate and pressing problems. To improve communication, patients could be invited on grant-awarding advisory boards. Setting up websites to foster discussions is another possibility. In addition, private foundations could conduct surveys of the patient population to help establish patient priorities, which could then help governmental funding agencies set priorities.

d. Test candidate therapies for dysfunctional side-effects

Participants criticized some ongoing efforts that apply experimental findings to humans before they have been tested thoroughly in animals. Given the unpredictability of the spinal cord's responses to injury and the potential for eliciting dysfunctional changes, it is crucial to require comprehensive pre-clinical studies before conducting clinical trials. It will also be important to determine the critical time periods for each candidate therapy.

e. Develop objective and quantitative outcome measurements.

Creating adequate outcome measurements for humans was considered a top priority. Three basic approaches were identified as potential candidates: 1) the use of imaging technologies - including functional magnetic resonance imaging (fMRI) and positron emission tomography (PET)-- to track changes in the activity of the brain and spinal cord, 2) electrophysiological monitoring, including the use of electromyograms and microelectrode arrays, and 3) the use of robotic devices to measure movement trajectories, speed of movements, and degree of weight bearing.

IVB. Basic Science Priorities
a. Elucidate key components of spinal cord circuitry

The development of neural prostheses and rehabilitation regimens will require a better understanding of the normal circuitry of the spinal cord. For example, understanding how the CPG is linked to cortical and sensory inputs should help guide the development of prostheses with feedback systems, and help optimize the sensory stimuli used for rehabilitation. In addition, identifying movement primitives in human and non-human primates could aid the development of prostheses that trigger complex behaviors, such as locomotion. If the primitives exist, determining how they are regulated by sensory and supraspinal inputs will also be necessary to develop appropriate feedback devices.

A relatively new area of research is to understand the contribution of trophic- and activity-dependent influences between neurons within a circuit, and how these factors establish, maintain and modulate function. There has been few studies investigating structural re-modeling in the spinal cord; this is an important mechanism of plasticity in other areas of the nervous system, such as the hippocampus. Finally, understanding the development of spinal cord circuitry, including the establishment of the CPG, during embryogenesis and neonatal life should help reveal the developing cord's capacity for change, and how this becomes limited in adulthood.

b. Identify areas of cross-fertilization between SCI and space physiology

Astronauts exposed to microgravity suffer from physiological alterations that resemble those experienced by patients with SCI, including muscle atrophy, bone loss, disruption of locomotion and coordination, and impairment of functions regulated by the autonomic system. Although astronauts suffer from a mostly reversible and milder degree of the symptoms, the similarities are significant enough to suggest that both areas of research could benefit from each other's findings and therapeutic developments.

For example, one of Edgerton's recent space flight studies suggests that alterations that occur during spaceflight may also occur in SCI. He found that the control of motor pools changes in response to changes in weight-bearing activity. After exposure to microgravity for 14 days, rhesus monkeys showed adaptations in the tendon force and electromyographic amplitude ratios of different muscles, indicating that their patterns of muscle recruitment were reorganizing. Since SCI also involves changes in weight-bearing activity, it is possible that motor output in SCI is affected by similar reorganizations of muscle recruitment.

c. Support projects that avoid conflicts over intellectual property rights

Collaborations can fall apart when researchers perceive a risk of losing their intellectual rights to a new discovery or invention. Although it is impossible to predict the exact evolution of a project, the probability of establishing successful collaborations could be increased by setting up projects where the roles of each team were clearly defined from the outset, and competitive overlap was minimized.  

V. Participants

Jimmy Abbas
Center for Biomedical Engineering
University of Kentucky

Ken Baldwin
Department of Physiology and Biophysics
University of California, Irvine

Hugues Barbeau
School of Physical and Occupational Therapy
McGill University

Emilio Bizzi
Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology

Lisa A. Cash
Spinal Cord Connections

Marina Chicurel
Science Writer

Arlene Chiu

Avis Cohen
Department of Biology
University of Maryland

Graham Creasey
Case Western Reserve

Bruce Dobkin
Department of Neurology
University of California, Los Angeles

Reggie Edgerton
Department of Physiological Science
University of California, Los Angeles

Lisa Fugate
Physical Medicine & Rehabilitation
Ohio State University

Simon F. Giszter
Department of Neurobiology and Anatomy
MCP Hahnemann University

Susan Harkema
Department of Neurology
University of California, Los Angeles

Leif Havton
Department of Neurology
Reed Neurological Research Center
University of California, Los Angeles

Bill Heetderks

Susan Howley
Christopher Reeve Paralysis Foundation

Doug McCreery
Huntington Medical Research Institute

Almudena Ramon-Cueto
Neural Regeneration Group
Institute of Biomedicine, Spanish Council for Scientific Research (CSIC)

Allan Tobin
Brain Research Institute
University of California, Los Angeles

David Tomko
Biomedical Research and Countermeasures Program

Mark Tuszynski
Department of Neurosciences
University of California, San Diego

Kent Waldrep
Kent Waldrep National Paralysis Foundation

Michael Weinrich
National Center for Medical Rehabilitation
National Institute of Child Health and Human Development

Lea Ziskind-Conhaim
Department of Physiology
University of Wisconsin Medical School

Last updated April 15, 2011