Michel A. Lemay

Tapping into spinal circuits to restore motor function

Motivated by the challenge of improving neuroprosthetic devices, the authors review current knowledge relating to harnessing the potential of spinal neural circuits, such as reflexes and pattern generators. If such spinal interneuronal circuits could be activated, they could provide the coordinated control of many muscles that is so complex to implement with a device that aims to address each participating muscle individually. The authors goal is to identify candidate spinal circuits and areas of research that might open opportunities to effect control of human limbs through electrical activation of such circuits. David McCreas discussion of the ways in which hindlimb reflexes in the cat modify motor activity may help in developing optimal strategies for functional neuromuscular stimulation (FNS), by using knowledge of how reflex actions can adapt to different conditions. Michael ODonovans discussion of the development of rhythmogenic networks in the chick embryo may provide clues to methods of generating rhythmic activity in the adult spinal cord. Serge Rossignol examines the spinal pattern generator for locomotion in cats, its trigger mechanisms, modulation and adaptation, and suggests how this knowledge can help guide therapeutic approaches in humans. Hugues Barbeau applies the work of Rossignol and others to locomotor training in human subjects who have suffered spinal cord injury (SCI) with incomplete motor function loss (IMFL). Michel Lemay and Warren Grill discuss some of the technical challenges that must be addressed by engineers to implement a neuroprosthesis using electrical stimulation of the spinal cord, particularly the control issues that would have to be resolved.

A DYNAMIC MODEL FOR SIMULATING MOVEMENTS OF THE ELBOW, FOREARM, AND WRIST

We developed a dynamic model of the upper extremity to simulate forearm and wrist movements. The model is based on the skeletal structure of the arm and is capable of elbow flexion/extension, forearm pronosupination, and wrist flexion/extension and radial/ulnar deviation movements. Movements are produced by activation of a Hill-type model of muscle, and limits on joint motion are imposed by passive moments modeled after experimental results. We investigated the muscle output force sensitivity, as well as wrist flexion/extension motion sensitivity to parameter variations. The tendon slack length and muscle fiber length were found to have the greatest influence on muscle output and flexion/extension wrist motion. The model captured the direction of the moment vectors at the wrist well, but predicted much higher moments than were measured by stimulating the paralyzed muscles of one tetraplegic subject.

Tendon transfers and functional electrical stimulation for restoration of hand function in spinal cord injury

Spinal cord injury at the C5 and C6 level results in loss of hand function. Electrical stimulation of paralyzed muscles is one approach that has demonstrated significant capacity for restoring grasp and release function. One potential limitation of this approach is that key muscles for stimulation may have lower motor neuron damage, rendering the muscles unexcitable. We have used surgical modification of the biomechanics of the hand to overcome this limitation. Tendon transfer of paralyzed but lower motor neuron intact muscles can compensate for potential function lost owing to muscles with lower motor neuron damage. Such procedures have been performed to provide finger extension, thumb extension, finger flexion, and wrist extension. Additional surgical procedures have been performed to enhance the function provided with electrical stimulation. These are side-to-side synchronization of the finger flexor and extensor tendons, the flexor digitorium superficialis Zancolli-lasso procedure, and thumb interphalangeal joint arthrodesis. These procedures have been performed in 11 patients with C5 and C6 level spinal injuries and functional electrical stimulation neuroprostheses. In these patients, 41 different functional electrical stimulation-related procedures were performed and 38 gave the desired result after surgery. One procedure resulted in no increase or decrease in function or muscle output, and two procedures resulted in a decrease in muscle force or joint range of motion. The issues that must be considered in performing functional electrical stimulation-related tendon transfers are discussed.

RESTORATION OF PRONOSUPINATION CONTROL BY FNS IN TETRAPLEGIA-EXPERIMENTAL AND BIOMECHANICAL EVALUATION OF FEASIBILITY

Individuals with C5/C6 tetraplegia lack voluntary control of the forearm pronators. We evaluated the feasibility of restoring forearm pronation/supination control using an electrically activated pronator opposed by voluntary supination. To this end, we measured the electrically produced pronation moments of subjects with tetraplegia. The maximal pronation moment achieved by stimulating the pronator quadratus ranged from 30 to 100 N cm in three forearms of two subjects. These moments were sufficient to produce forearm pronation in all three forearms. Voluntary control of pronosupination during constant pronator stimulation was achieved by having the subject voluntarily supinate or relax to change the balance of rotational torques acting on the forearm. In all cases, the subjects were able to supinate voluntarily against the continuously stimulated pronator, producing intermediate angles between full pronation and full supination. We also observed under some conditions that subjects could voluntarily pronate and supinate even without pronator stimulation. Using a biomechanical model, we show how pronation can be initiated from a supinated position using the brachioradialis, with gravity completing the pronation. This method of pronation without stimulation is extremely sensitive to the orientation of the forearm in the gravitational field, and thus is not a widely applicable technique. We conclude that forearm pronosupination via Functional Neuromuscular Stimulation is feasible, and would provide subjects the ability to pronate without the assistance of gravity.

Automated tuning of a closed-loop hand grasp neuroprosthesis

An automated tuning algorithm was developed to reduce the time and skill required to tune a closed-loop hand grasp neuroprosthesis. The time reduction results from simultaneous tuning of four gain parameters controlling the dynamic response of the system, and from automation of the calculation and decision processes. The method is therefore an automated parallel tuning method, replacing a manual sequential method in which only one parameter at a time was tuned. RMS error between the step input and the grasp output is minimized, with absence of oscillation as a constraint. The difference between the systems RMS ramp tracking errors for the two tuning methods was less than 1% of the ramp size regardless of the initial values of the parameters, implying that the tuning methods were equivalent. However, the parallel tuning method was faster and required fewer trials than the sequential method. The capability of the closed-loop system to regulate grasp output in the presence of disturbances was shown to be better than the capability without feedback.<<ETX>>

Endpoint forces obtained during intraspinal microstimulation of the cat lumbar spinal cord - experimental and biomechanical model results

We studied the structure of endpoint forces produced by microstimulation of the cat spinal cord, and from combinations of muscles using a biomechanical model of the cat hindlimb. The forces evoked by microstimulation were of four types. At some stimulation sites, the force patterns exhibited a point of convergence where the active endpoint force was zero. The endpoint forces produced by activating combinations of muscles in the biomechanical model were of types similar to the experimental ones, although they demonstrated points of convergence in locations not observed experimentally. These results suggest that the spinal circuitry uses a subset of the possible muscular combinations.

Endpoint forces evoked by microstimulation of the cat spinal cord

The authors studied the mapping and structure of endpoint forces produced by microstimulation of the cat spinal cord. The results demonstrate that stimulation in the dorsal and intermediate aspects of the spinal cord generated organized, convergent force patterns, while stimulation in the ventral aspect of the cord did not. These results suggest that electrical activation of higher-order neurons can be used to coordinate the activation of multiple muscles in multi-joint movements.

Endpoint force patterns evoked by intraspinal stimulation-ipsilateral and contralateral responses

We studied the mapping and structure of endpoint forces produced by microstimulation of the cat spinal cord. The forces evoked by microstimulation varied in magnitude and direction as a function of limb configuration. At some stimulation sites, the force patterns exhibited a point of convergence where the net endpoint force was zero. Ipsilateral stimulation in the dorsal aspect of the cord evoked flexion responses that exhibited a convergent point, while extension responses evoked by ipsilateral stimulation were not convergent. Conversely, contralateral stimulation in the dorsal aspect of the cord evoked extension responses that exhibited a convergent point, while flexion responses evoked by contralateral stimulation were not convergent. Stimulation in the ventral aspects of the cord, as well as intramuscular stimulation of single muscles did not produce convergent force patterns. The results demonstrate that stimulation in the dorsal and intermediate aspects of the spinal cord can generate organized, convergent force patterns, while stimulation in the ventral aspect of the cord can not. These results suggest that electrical activation of higher-order neurons may be used to coordinate the activation of the multiple muscles required in multi-joint movements.

Active regulation of grasp stiffness in neuroprosthesis for restoration of hand function in quadriplegics

An active, closed-loop stiffness regulation system is being incorporated into a neuroprosthesis for hand grasp for quadriplegic patients. The stiffness regulator controls the activation of the thumb in lateral grasp and the fingers in palmar grasp. The opposing digits are fixed in place by an open-loop system that operates in parallel. A single continuous input specifies the size of grasp opening and the grasp force under a wide range of loading conditions. The parameters of the system can be tuned to specify the static input-output properties of the grasp as well as a stable and rapid response under isometric, compliant and unloaded conditions.<<ETX>>

Experimental and biomechanical model force fields produced by intraspinal microstimulation of the cat lumbar spinal cord

Using a biomechanical model of the cat hindlimb, we studied patterns of endpoint forces created by all muscle combinations of fourteen selected muscles, and compared them to the force patterns produced by intraspinal microstimulation of the lumbar spinal gray matter. We ran the model with two different activation schemes for the muscles. The first run used combinations of the fourteen selected muscles stimulated at the same level of activation. The second run used combinations where muscle forces were normalized to produce the same maximum end-point force. These results were compared to force field patterns obtained experimentally during intraspinal microstimulation. Although there were slight variations in the force patterns produced, both methods converged to four dominant patterns. When muscles in the model were normalized, some force patterns were found that were not observed experimentally. These results show the significance of specific levels of muscle activation to the production of the experimental patterns.