A Modular Approach to Retraining Muscles after Stroke

Richmond FJR ¹, Baker LL ^2,3, Winstein C ², Waters RL ³, Loeb GE ¹

¹ AE Mann Institute for Biomedical Engineering, University of Southern California, Los Angeles,USA

² Dept. Biokinesiology and Physical Therapy, University of Southern California, Los Angeles, USA

³ Rancho Los Amigos National Rehabilitation Center, Downey, CA, USA

Email: fjr@usc.edu ; Website: http://ami.usc.edu

Abstract

A foot drop stimulator using implanted microstimulators (BIONs) was developed by modifying a WalkAide2 stimulator. BIONs were implanted in a person with incomplete spinal cord injury and severe foot drop in one leg. Compared to surface stimulation, BIONic stimulation of the deep peroneal nerve produces a more balanced ankle flexion without everting the foot. For effective stimulation, the BION must be within 10-15 mm of the nerve. The BIONic WalkAide elevates the foot so that the toe clears the ground by 3 cm, which is equivalent to the toe clearance in the less affected leg. The physiological cost index (PCI), a measure of effort during walking, is high without stimulation (2.29 ± 0.37; meanThe pathophysiology, clinical course and residual disability of patients presenting with a hemiparetic stroke are highly variable. ± S.D.) and greatly reduced with surface (1.29 ± 0.10) and BIONic stimulation (1.46 ± 0.24). Also, walking speed is increased from 9.4 ± 0.4 m/min. without stimulation to 19.6 ± 2.0 m/min with surface and 17.8 ± 0.7 m/min. with BIONic stimulation. We conclude that FES with BIONs is a practical alternative to surface stimulation and can provide a more balanced dorsiflexion.This makes it difficult to apply a highly specific treatment modality to a large population of patients and even more difficult to identify its safety and efficacy vis-à-vis the incremental and ad hoc application of myriad conventional treatments. We are conducting a set of clinical trials of a new class of modular injectable microstimulators that can be used in a wide variety of sites and exercise paradigms. Initial results from electrically induced exercise of the shoulder and wrist and finger muscles are encouraging. Actual clinical experience suggests, however, that the real power of such treatment resides in its ability to be adapted to the limitations, needs and progress of each patient and to complement other treatments rather than adhering slavishly to research protocols.

1 1. Introduction

Until recently, FES has been limited to three approaches: surface stimulation, percutaneous and fully implanted wire electrodes. Loeb et al. [1, 2] developed a novel, injectable microstimulator (BION^TM) that does not require surgical implantation and receives power and data through a wireless link. BIONs have already been used in clinical trials for therapeutic electrical stimulation (TES) [3]. TES applications do not require the BION systems to be portable, because stimulation is applied when the subject is sitting in a chair and the external coil is connected to a controller powered from an AC receptacle.

BIONs were originally designed to stimulate muscles near their end plate. With this method, increasing the stimulation level will increase the response gradually as more and more distant nerve branches are activated. This approach has a disadvantage, since many human muscles have multiple end plates that may be separated by several cm. Also, a functional movement may require activation of several muscles and so may involve implantation and control of a number of BIONs. An alternative approach is to place a BION near a peripheral nerve that innervates a number of muscles. In this way one or a small number of BIONs may suffice. The disadvantage is that the recruitment of nerve fibers is much steeper [4-6] and graded control becomes more problematic. Also, to the extent that the nerve innervates a number of muscles with different functions, separation of these functions may not be feasible.

We report here experiments in animals (two anesthetized cats) and a human subject. The animal experiments were designed to answer three questions: 1) How far can BIONs be from a peripheral nerve (sciatic) and still produce effective stimulation? 2) How steep are the recruitment curves with nerve stimulation? 3) Is the orientation of the BION with respect to the nerve critical?

A human subject with an incomplete spinal cord injury (C6/C7) was implanted both with BIONs near a peripheral nerve (deep peroneal) and in muscles innervated by that nerve (tibialis anterior, peroneus longus).. He had received treadmill training with partial body weight support [7], but had a serious, residual foot drop in his left leg. He had used a surface stimulator (WalkAide2) to correct this problem since 2001 and volunteered to be implanted with BIONs. A prototype of a portable (wearable) BIONic foot drop stimulator was developed for this subject. Stimulation is turned on and off by a WalkAide2 device, based on the tilt angle of the shin during walking [8]. When the leg tilts backward behind the body at the end of stance, the stimulator is turned on to help liftThe clinical disorder known as stroke covers a wide range of pathophysiology and clinical disability. The branching pattern of the vasculature varies from patient to patient and the premorbid capabilities and representation of a given sensory or motor function depends on the history of each individual. The amount of brain tissue that is permanently lost as opposed to temporarily nonfunctional depends on the extent of the obstruction as well as the availability and development of collateral circulation. The resulting heterogeneity of symptoms makes it difficult to apply and assess therapeutic interventions.

When physical therapists devise treatments to rehabilitate stroke patients, they proceed incrementally based on the history, current condition and needs of the individual patient. Spontaneous changes in the condition of the patient (for better or worse) may make a given treatment inappropriate, or may require different or supplemental modalities, particularly in the early stages of recovery. Postponing treatment until the patient is “stable” is likely to make recovery more difficult (see below). These circumstances are anathema to the usual design of a clinical trial intended to demonstrate efficacy of a specific treatment.

One physiological consequence of a paretic limb is actually quite predictable and amenable to treatment by neuromuscular electrical stimulation (NMES). This is the disuse atrophy and consequent weakness and fatigability of the paretic muscles. Disuse atrophy develops rapidly. Thus, it tends to interfere with any spontaneous recovery of function as motor function of the cortex improves. Disuse atrophy leads to a variety of sequelae (e.g. shoulder subluxation [1] and spasticity (e.g. flexion contractures of the hand). These secondary problems further compromise the ability of the patient to participate in and benefit from physical therapy such as constraint induced therapy, robotically assisted therapy, and other techniques that might encourage relearning to use the paretic limb.

Even modest amounts of daily NMES can substantially prevent or reverse disuse atrophy [2], but it is used relatively infrequently in most clinics and prescribed rarely for home use. The available technology of transcutaneous electrodes requires careful placement and adjustment, which many patients and their caregivers find difficult to master. Unpleasant sensations and skin irritation may result when relatively large currents are applied to the skin. Surgically implanted electrical stimulators can overcome these problems in principle, but they are typically expensive, bulky and difficult to adapt to changing rehabilitative needs.

the foot. When the leg is tilted in front of the body toward the end of swing, the stimulator is turned off. This is the first application of BIONs for FES to our knowledge, although implanted stimulators with and without wires have been surgically implanted previously for foot drop [9-11].

2 2. Animal StudiesMethods

Two adult cats were anesthetized with Somnotol and the sciatic nerve was exposed by an incision along the lateral surface of the thigh. Glass capsules the size of BIONs with wires at each end were sutured in nearby muscles at various distances from the nerve. In one experiment the capsule was oriented perpendicularly with the cathode closest to the nerve. In the other the capsule was oriented parallel to the nerve. A Grass stimulator (SD9) applied monophasic stimuli with 200s duration. The current was monitored with a current probe and displayed on a digital oscilloscope. EMG electrodes were implanted percutaneously into the lateral gastrocnemius (LG) muscle about 1 cm apart. The signals were amplified and also displayed on the oscilloscope. Peak-to-peak EMG and average BIONcurrent values were measured™ wireless microstimulators provide useful features for the stimulation of paretic muscles. One of more BIONs can be injected into various muscles or near muscle nerves to provide precise and selective control of the intensity and temporal patterning of individual muscle activation. Each device receives power and its own unique command signals from an externally worn RF transmission coil [3]. Additional BIONs can be injected at any time in a simple outpatient procedure. They are inert and biocompatible, so may be left in place indefinitely whether or not they are being used by the patient. After implantation, the clinician determines the threshold to produce a visible twitch of each muscle and designs one or more exercise programs consisting of trains of electrical pulses at a desired rate (1-50pps) and intensity (specified as a multiple of threshold stimulus charge, by controlling the pulse current over the range 0-30mA and pulse duration over the range 2-512μs). The onset, duration and interleaving of pulse trains from various implants are easily set using a graphical interface on a portable personal computer (ClinFit™). The exercise programs are downloaded into a portable, microprocessor-based controller (Personal Trainer™) that provides the patient with start/stop controls and records their usage. The externally worn coils come in several shapes to match various implantation sites.

3 3. Clinical Studies

We are conducting several clinical studies that target weakness of the arm and hand. Because of regulatory considerations, each study deals with a different clinical problem and tests the safety and efficacy of a specific NMES intervention against various control treatments, as summarized below:

3.1 3.1 Prevention of shoulder subluxation in subacute stroke patients

The methods and results of the first ten patients enrolled in this prospectively randomized control and cross-over study have been published [4]. The center continues to accrue patients at a low rate. Briefly, patients with a flaccid arm at 4-10 weeks post-stroke who are otherwise mentally intact are implanted with one BION in each of the supraspinatus and middle deltoid muscles. They self-administer exercise programs for 20-30 minutes, 2-3 times per day. The programs are designed to produce maximal recruitment of the implanted muscles at low stimulus rates (typically 5 pps) in interrupted trains (typically 3s on and 3s off) that produce relatively low forces and little arm movement. Most patients were already clinically subluxed at the start of treatment; subluxation was substantially reduced or eliminated after a few weeks of stimulation. Pain is not well correlated with subluxation in timing or intensity but is known to develop eventually in 60-80% of such patients with conservative treatment. None of our initial 10 patients developed shoulder pain (2-4 year follow-up available on most).

The anecdotal findings may be even more revealing than the formal hypothesis, which had already been proven using surface stimulation [1]. The implanted patients all liked the sensations associated with muscle stimulation and most elected to continue self-administering the exercise after the end of the formal study period, whether or not their subluxation tended to recur. Muscles were strikingly weak and fatigueable at the start of stimulation but improved greatly within the first week or two. Some patients regained substantial voluntary control over shoulder abduction and elevation (the function of the stimulated muscles) even while the rest of their arm remained paralyzed. In one patient, such movement was functionally useful because he could place his hand on a pole which could then be grasped by using flexion tone, in order to help support himself when standing.

3.2 3.2 Reversal of shoulder subluxation in chronic stroke patients

BION treatment similar to the subacute study described above is currently being applied to patients with stable, chronic shoulder subluxation (>6 mo post-stroke) and compared to surface stimulation in a cross-over study design. Implantation of the correct sites is more difficult in chronic patients because of the more severe atrophy. Two of the first three BION patients have done better with surface stimulation; one achieved complete reduction with BION stimulation, which relapsed when stimulation was stopped. Three of four patients started on surface stimulation have expressed a strong desire to receive BION implants. Initial use of surface stimulation may be an effective way to identify patients whose symptoms are likely to respond to muscle exercise and to condition their muscles to facilitate BION injection.

3.3 3.3 Prevention and/or reversal of flexion contractures of the hand and fingers in chronic stroke patients

Patients with paralyzed extensor muscles of the wrist and fingers usually develop permanent contractures because of the greater strength and spastic tone of the flexor muscles. Such contractures prevent even passive use of the hand. They are unsightly and, in the extreme, result in skin care problems when the fingernails press into the palm. The first three patients in this study constitute a pilot group intended to identify appropriate implantation sites for BIONs to produce balanced contraction of the various wrist and finger extensor muscles. The motor branching pattern of the radial nerve suggests that multiple synergists can be recruited from individual sites, but there is substantial individual variability and disuse atrophy complicates assessment during implantation.

All three chronic stroke patients implanted with BIONs obtained rapid and substantial reduction of their contractures. The stimulated muscles had rapid improvement in strength and fatigability, similar to that seen in shoulder patients. Unexpectedly, two patients progressed from minimal to substantial voluntary wrist extension, but did not achieve significant voluntary finger extension.

3.4 3.4 Retraining of voluntary extensor function in the hand and fingers of chronic stroke patients

One of our team (CW) leads a research program to assess the efficacy of constraint induced therapy (CIT), that combines intensive task-specific practice with encouraged use of the paretic limb by restraining the intact limb during activities of daily living