Innovations in electrical stimulation harness neural plasticity to restore motor function

Abstract

Novel technology and innovative stimulation paradigms allow for unprecedented spatiotemporal precision and closed-loop implementation of neurostimulation systems. In turn, precise, closed-loop neurostimulation appears to preferentially drive neural plasticity in motor networks, promoting neural repair. Recent clinical studies demonstrate that electrical stimulation can drive neural plasticity in damaged motor circuits, leading to meaningful improvement in users. Future advances in these areas hold promise for the treatment of a wide range of motor systems disorders.

Executive Summary

Advances in invasive and non-invasive neural interface technology have increased the spatiotemporal precision of electrical stimulation and allowed for closed-loop systems that respond to changes in physiological state. These innovative devices now allow electrical stimulation drive plasticity within motor circuits, leading to neural repair and restored function.

• The future research and innovation should focus on the directions listed below to improve specificity in stimulation protocols for user-individualized therapy.

  1. Closed-loop systems that drive stimulation based on patients’ real-time neurophysiological information.
  2. Increased spatial and temporal precision of stimulation delivery that can harness endogenous neural plasticity mechanisms.
  3. Improved neural interfaces designed to maximize recording and stimulation effectiveness.

Introduction

Restoring motor function after damage to the nervous system has been an intractable challenge for patients, clinicians and researchers. Motor systems disrupted by traumatic injury, pathological insult or neurodegeneration often do not regain full functionality. Currently, diverse types of electrical stimulation devices are designed to restore or substitute for lost motor functions. These include deep brain stimulation for essential tremor, dystonia and Parkinson’s disease and electrical stimulation of muscles/nerves to facilitate rehabilitation after spinal cord injury or stroke. However, current devices are primarily open-loop systems - controlled by external and arbitrary commands. As such, they lack precision in the delivery of stimulation, limiting the degree of motor control and circuit plasticity that these technologies can bring to users.

Recent advances in neurostimulation research suggests its potential to harness endogenous neural plasticity, broadening the therapeutic possibilities to a diverse range of motor disorders. Early clinical successes hint at the power of this approach. For instance, deep brain stimulation (DBS) applied earlier in disease progression may alter disease trajectory; spinal cord stimulation (SCS) has restored voluntary motor control in a small number of patients; and precisely-timed vagus nerve stimulation (VNS) speeds rehabilitation following stroke. The evolving landscape of innovative neurostimulation technology has been critical to opening up new therapeutic applications.

The ability for neuromodulation technology to drive neural plasticity relies on innovations in several key areas: precise timing of stimulation delivery, closed-loop device technology with integrated sensing capabilities and novel invasive and non-invasive neural interfaces to broaden the range of possible targets within the nervous system (Figure 1). Pairing precise electrical stimulation with rehabilitation appears to activate hyperpolarized neural circuits and drive neural plasticity. Closed-loop stimulation devices, which incorporate electrophysiological sensing and feedback stimulation, provide instantaneous targeted delivery - an ultimate form of precision medicine. Novel interface technology, such as smaller implantable electrodes and new non-invasive stimulation devices can reach a wider range of neural structures in the central and peripheral nervous system. Together, the expanding repertoire of medical device technology provides new means by which electrical stimulation can induce neuroplasticity and endogenous repair mechanisms, and is bringing a renewed sense of enthusiasm and hope to researchers, clinicians and patients.

Figure 1.

Innovations in electrical stimulation drive neural plasticity to restore motor function. Advances in invasive and non-invasive neural interface technology have increased the spatiotemporal precision of electrical stimulation and allowed for closed-loop systems that respond to changes in physiological state. These innovative devices now allow electrical stimulation drive plasticity within motor circuits, leading to neural repair and restored function.

Electrical stimulation for acute modulation of motor circuits

The most widely-used neurostimulation therapy for movement disorders is deep brain stimulation (DBS) for the treatment of Parkinson’s disease, essential tremor (ET) and dystonia. As a class, movement disorders effect more than 10 million individuals in the United States [1–4], cost billions of dollars each year in care, and the at-risk population is projected to double over the next three decades [4,5]. For this reason, there has been a focus in the academic and medical community on finding long lasting and effective treatments for these disorders. DBS was first approved by the FDA in 1997 for essential tremor [6–8] and in 2002 for PD. Since being approved for market, DBS implants have been performed at an increasing rate, over 150,000 users have received a DBS system, with over 60% implanted in the last five years [9,10]. These implants have also displayed robust life spans, providing users with significant improvements even after more than five years of use [11–13].

DBS directly modulates motor circuit function DBS is achieved by surgically implanting a stimulating electrode unilaterally or bilaterally into the thalamus or striatum and applying continuous electrical stimulation to those brain structures. DBS produces an acute alleviation of symptoms, and implanted patients revert back to baseline level function minutes after stimulation ends [12–14]. DBS was initially thought to inhibit activity in the basal ganglia circuit as an ‘informational lesion’ [15,16]. However, later research suggests that DBS may serve as a low-pass filter for signal transmission [17,18], disrupting the synchronization of the basal ganglia and downstream structures [19–21] and inducing short-term inhibitory plasticity [22].

Future directions to improve DBS therapy

Novel treatment paradigms, stimulation protocols and device technology are driving forward innovations in the treatment of movement disorders with DBS and other forms of electrical stimulation. The application of DBS earlier in disease progression is currently under investigation as a way to alter disease trajectory in Parkinson’s patients by enhancing adaptive plasticity [23–26]. Animal studies suggest that DBS drives markers of plasticity, such as BDNF expression, within motor circuits[27–29]. In addition, a pilot clinical study examining DBS in early stage Parkinson’s disease patients shows promise that DBS may slow the progression of the rest tremor [30,31]

To improve the efficacy and efficiency of DBS, there is a growing focus in the field on adaptive DBS (aDBS) [32]. aDBS utilizes feedback from a user’s unique neurophysiology, typically through recorded local field potentials (LFPs), for real-time information of the user’s state to modify stimulation parameters. During implantation, aDBS strategies use LFPs to help target brain regions more accurately [33–35]. Importantly, LFPs can be reliably detected from DBS users over several years [36], providing critical longitudinal insight into the disease [37–42] and fundamentals of the human motor system [43,44]. Moreover, aDBS may also be used to reduce side effects and extend the longevity of DBS devices [45–47].

Although DBS is the most common electrical stimulation device to treat ET, the FDA recently approved a non-invasive median and ulnar nerve stimulation device that reduces tremor in ET patients [48]. This device is worn on the wrist, and applies transcutaneous electrical stimulation to the nerves, minimizing tremor in some users. Further development of non-invasive techniques could open the door for electrical stimulation devices to be used on a far broader range of patient populations due to their drastically decreased risk profile.

Electrical stimulation to accelerate recovery from injury

While DBS produces instantaneous relief from the symptoms of Parkinson’s disease, electrical stimulation for treating neural injury relies on driving neural plasticity to enhance functional recovery. In this treatment paradigm, chronic electrical stimulation is paired with physical rehabilitation over months to years, and leads to outcomes that surpass physical rehabilitation alone. These stimulation modalities include spinal cord stimulation (SCS), vagus nerve stimulation (VNS), and direct brain stimulation. These paradigms have been applied to treat spinal cord injury and stroke, and may have broader implications for other types of injuries in central and peripheral nervous system.

Epidural spinal cord stimulation to drive neuroplasticity

Spinal cord injury (SCI) is a devastating condition that affects roughly 2.5 million people worldwide[49]. Depending on the location of the injury, patients fully or partially lose motor or sensory functions. After initial diagnosis, patients begin physical rehabilitation programs that aim to maintain muscle tone, reduce muscle atrophy and pain, and may improve prognosis[50]. Physical rehabilitation may induce plasticity in spared spinal cord circuits[51], but has not been able to restore volitional movement to patients with motor complete spinal cord injury.

Historically, electrical stimulation in the spinal cord has been demonstrated to activate motor pools, allowing patients with SCI to initiate some movements, but did not restore function of higher-order movements such as walking[52,53]. A major advance in SCS involved pairing stimulation with ongoing physical rehabilitation. In patients with complete motor injuries and either ‘sensory incomplete’[54,55] or ‘sensory complete’ injuries[56], epidural stimulation paired with rehabilitation led to partial recovery of intentional movement in lower and upper limbs. In addition, EMG activity was detected in the lower limbs during stimulation, suggesting that previously non-functional motor circuits in motor complete SCI patients were re-activated[56]. Several groups have since shown amazing recovery of stepping[57] and over-ground walking[58,59] in patients with SCI, an unprecedented leap forward in treatment for these patients. A perhaps unexpected additional benefit of SCS is the effect on autonomic function, including elevating blood pressure from hypotensive states[60,61] and improving bladder and bowel function[62]. Excitingly, similar results may also be produced by non-invasive transcutaneous electrical stimulation, opening up the possibility of treating a broader range of patients without the need for surgical procedures [63,64].

SCS drives both acute and chronic alterations in motor circuit function

Spinal cord stimulation is thought to aid recovery by: 1) depolarizing motor pools in the spinal cord, reducing the activation threshold of spared motor and sensory tracts and 2) evoking plasticity to remodel the spared motor tracts[65]. Spinal cord injuries typically do not result in a complete “separation” of tissue, rather contusions or lacerations leading to damage and inflammation. Remaining supraspinal connectivity is not sufficient to activate lower motor neuron pools, so SCS is required to acutely activate downstream neural populations. However, for complex voluntary movement, such as over-ground walking, acute SCS must be paired with physical rehabilitation over numerous training sessions, to drive chronic circuit plasticity. Recent clinical reports have suggested that patients with higher likelihood of recovery are those with partial lesions, specifically those with some somatosensory or proprioceptive input[66,67]. Yet, even in the most severe SCI injury, there is evidence that EMG activity can be generated by descending pathways, suggesting some white matter tracts are spared from damage and may allow patients to recover partially through plasticity of spared tracts during SCS paired with rehabilitation[68,69].

Future improvements in technology: Closed-loop SCS

Closed-loop stimulation synchronizes motor intent and SCS, thus inducing more salient plasticity preferentially in connections relevant to higher-order movements[70,71]. Two methods of closed-loop stimulation are the focus for future clinical implementation: cortico-spinal decoding and EMG-based closed loop stimulation. Cortico-spinal decoding records signals in motor cortex to decode movement intentions, and then deliver temporally precise stimulation to the spinal cord below the lesion, leading to greatly improved motor control in rodents[72] and non-human primates [73]. Enhanced plasticity in cortical control was driven by spared ascending sensory afferents in turn increasing the density of cortical fibers projecting to subcortical motor regions[68]. Alternatively, EMG and other physiological measures, can approximate motor intention, allowing for precisely timed delivery of epidural stimulation[74]. In a study of several patients with chronic cervical spinal cord injury, closed-loop stimulation accelerated performance of over–ground walking [59]. Tailoring closed-loop stimulation based on cortical decoding methods may result in different varieties of neuroplasticity in spinal cord circuits, suggesting opportunities for engineered neuroplasticity[75].

Paired VNS for the recovery of motor function after stroke

New advances in VNS therapy gives promise to the recovery of movement capabilities. Currently, VNS is marketed as a treatment for medically-intractable epilepsy and depression. For these therapies, VNS is applied continuously, similar to the stimulation protocol for marketed DBS devices. However, a new VNS stimulation paradigm, where brief bursts of VNS are paired with physical rehabilitation, is thought to drive long-lasting circuit plasticity and speed motor rehabilitation [76–83]. In theory, by inducing plasticity and speeding rehabilitation, paired VNS has potential to improve therapy for a broad range of disorders of motor systems.

Clinical trials are exploring the use of paired VNS to enhance motor rehabilitation in cases of stroke involving upper limb paresis. There are over 795,000 incidents of stroke every year, and approximately 40% of stroke survivors have functional impairment [84–86]. Recovery from stroke depends on structural and functional reorganization of lesioned tissue and intact contralateral structures. A pilot safety and feasibility study of VNS for motor rehabilitation following stroke demonstrated that the device is reasonable and safe [87]. Implantation of the stimulating cuff electrode has low surgical risk and is generally well-tolerated by users, with common adverse events including hoarseness of voice, coughing, and nausea ([88]. Patients receiving paired VNS showed a greater improvement in their Fugl-Meyer Assessment (FMA) compared to those receiving rehabilitation alone, providing justification for a future pivotal trial [89].

VNS drives functional recovery through plasticity in motor circuits

There is a strong basis of animal research to support the use of VNS for stroke rehabilitation. Rat models of ischemic and hemorrhagic stroke that received VNS paired with a forelimb reach showed significant recovery of strength in a forelimb lever pull task. When VNS paired with rehabilitation, the rate of full recovery doubles comparing to rehabilitation only control [80–83], and motor improvements can generalize across motor task [90]. VNS paired with a specific movement causes cortical plasticity, reorganization of motor cortex and increased representation of the corresponding area in healthy[91] and injured animals[92]. There is evidence to suggest that VNS-induced cortical plasticity is mediated cholinergic basal forebrain and noradrenergic locus coeruleus[93]. Neurons in the locus coeruleus are activated and release norepinephrine in response to VNS [94–96]. In SCI, VNS drives plasticity in spared spinal circuits[97], similar to the results in epidural spinal cord stimulation following spinal cord injury.

Future directions for VNS in multiple neurological conditions

VNS holds potential therapeutic efficacy across a rapidly growing range of medical indications, including autoimmune, metabolic and neurological disorders[78,98]. The diverse therapeutic space is reflective of the anatomy of the vagus nerve, which has widespread peripheral afferent and efferent innervation of heart, lungs, gastrointestinal tract and brainstem. In addition to the studies involving recovery from stroke, temporally-precise VNS shows promise in animal models of other motor system injuries, including spinal cord injury, traumatic brain injury and peripheral nerve injury. As with epidural stimulation for spinal cord injury, closed-loop stimulation may enhance the effectiveness of paired VNS [97].

While the implantation surgery for vagus nerve stimulators is relatively low risk, a non-invasive stimulation solution could have a profound impact on the safety, widespread use, and potential therapeutic indications for VNS. Transcutaneous stimulation of vagus afferents can occur at either the cervical or auricular branches. Evidence that non-invasive stimulation can sufficiently activate vagal afferents is accumulating[99–101]. For instance, non-invasive auricular and cervical stimulation showed a vagal sensory evoked potential similar to the one produced with invasive VNS [102]. Moreover, pilot studies of auricular VNS applied for 60 minutes prior to physical rehabilitation [103] or paired with movements during rehabilitation [104] show a small increase in FMA scores compared to sham stimulation.

Cortical stimulation for the recovery of motor function after stroke or spinal cord injury

In addition to VNS, invasive cortical stimulation paired with rehabilitation may also speed motor function recovery [105]. Initial clinical trials for cortical stimulation in patients with hemiparetic stroke showed minor improvements in upper extremity FMA, yet a broader pivotal trial found no increase in recovery over controls [106–108]. However, these trials may not have accounted for variable lesion size [109], lesion location [110] and delivery timing[111,112]. Non-invasive electrical stimulation, such as transcranial direct current stimulation (tDCS), may enhance recovery from stroke[113] and spinal cord injury[114], but large randomized controlled trials are needed to determine effectiveness[115].

Electrical stimulation to directly restore motor and sensory function

In addition to restoring motor function by driving plasticity in damaged neural circuits, electrical stimulation can transmit information directly into, or out of, motor and sensory pathways. Referred generally as neuroprosthetic systems, these devices restore uni- or bi-directional communication between the brain and the user’s peripheral nervous system. By directly activating peripheral motor units in patients with paralysis, electrical stimulation devices can bypass injured motor systems to restore motor function[116]. Alternatively, such systems can restore sensory function through stimulation of remaining peripheral or central sensory pathways.

Direct restoration of motor function through electrical stimulation

Foundational work on neuroprosthetic motor devices for spinal cord injured patients lead to the first implantable motor prosthesis, the Freehand System [117–119]. This device relies on direct stimulation of forearm and hand muscles to produce two hand grasp positions: lateral grasp for small objects and palmar grasp for large objects. Patients controlled the first-generation device through mechanical actuators on the shoulder, and later generations through EMG signals from implanted electrodes in user’s neck and shoulder muscles. The second generation system provide better control of grasp-release, forearm pronation, and elbow extension[120,121]. While the Freehand System is no longer commercially available, advanced work on these functional electrical stimulation systems continues under the direction of the Institute for Functional Restoration at Case Western Reserve University. Newer functional electrical stimulation systems often incorporate muscle-based control commands, allowing for closed-loop control by physiological measurements[122]. These devices provide motor control of the limbs, allowing users to complete activities of daily living[123,124].

The early work on myoelectric neuroprostheses inspired a search for a richer, and perhaps more intuitive set of control signals, leading to the development of brain-controlled neuroprosthetic devices. In these devices, control signals are generated by higher dimension natural cortical activities recorded through microelectrodes in the brain - devices called brain computer interfaces (BCIs)[125–127]. BCIs have been used to control virtual or robotic prosthetic assist devices, providing new movement capabilities to patients with tetraplegia. Moreover, recent work in non-human primates [128] and human patients [129–131] demonstrated that BCI signals can control direct stimulation of paralyzed forelimb muscles to restore voluntary forelimb grasp and reach movements.

Restoration of sensory function with electrical stimulation

Electrical stimulation can also be used to reproduce sensory perceptions in patients who have lost sensory function. As sensory perception is tightly linked to motor systems, restored sensation can have additional benefits to improve motor control in users. In these technologies, sensory input, such as that from tactile sensors on a prosthetic hand, is transmitted to stimulation electrodes in the remaining afferent nerves, allowing users to receive tactile and proprioceptive feedback [132–134]. The addition of sensory feedback improves the amputees performance using myoelectric prosthetic devices, and improves the embodiment of the device[135,136]. Similar systems have been used in pilot investigations to restore somatosensory feedback in SCI patients. Since SCI patients can’t relay peripheral sensation back to brain through damaged spinal cord, the stimulation is delivered directly to somatosensory cortex [137,138], producing cutaneous sensations.

Future directions for neuroprosthetic devices

Information processing in motor systems is highly complex and the circuits can change over short and long timescales. Advances in neuroprosthetic technology will include more robust, reliable interface with high information content. These may include better recording and stimulating electrodes with novel materials, biocompatible coatings, increased channel count and high-throughput wireless connectors[139]. There is also a need for better analytical methods to interpret incoming signals, encode stimulation parameters and calibrate analytical systems. Since both neural circuits and computer algorithms can learn and adapt over time, performance of neuroprosthetic devices needs to be continually optimized based on current benchmarks[140–142]. New computational strategies that incorporate dynamical systems, nonlinear models, rapid control feedback and other strategies to acknowledge network plasticity and account for learning across distributed neural networks may optimize solutions[143–148]. Finally, closed-loop neuroprosthetics systems are complex, and may require individualized configurations to meet each user’s needs. Modular devices, with interconnecting components can allow for personalized solutions, but they require standardized interconnections and may have a more complex regulatory pathway for use in humans[149].

Conclusions and future directions

Electrical stimulation is changing the lives of thousands of patients. Deep brain stimulation brings immediate and often remarkable changes to each user’s motor control and quality of life. Implanted functional electrical stimulation and neuroprosthetic systems have restored motor functions to patients with spinal cord injury, stroke and neurodegenerative disorders. Vagus nerve stimulation, spinal cord stimulation, and direct cortical stimulation can accelerate the rehabilitation and induce plasticity in remaining circuits. Yet, each of these technologies faces challenges in implementation, and has exciting new avenues for improvement.

In contrast to pharmaceutical treatment options, electrical stimulation allows for precise, targeted therapy for movement disorders. By nature, stimulation electrodes are placed intentionally in specific targets to provide local treatment. Precisely timed VNS during rehabilitation may enhance cortical plasticity to speed recovery of injured motor circuits. Similarly, spinal cord stimulation for spinal cord injury paired with physical rehabilitation allows users to regain volitional movement. Neuroprosthetic systems deliver temporally and spatially precise stimulation to target muscles or neural tissue, both to re-animate musculature or provide sensation. Improved specificity of device placement, stimulation delivery and stimulation protocols can drive future improvements in user-specific therapy.

Building on the advances brought by spatiotemporal precision, closed-loop stimulation includes a control command that drives stimulation onset. Real-time neurophysiological information recorded by electrodes sensing neural firing or EMG activity are a powerful source of control commands. Adaptive DBS systems use the patient’s neurophysiology to guide stimulation for improvement of both the efficiency, and likely the effectiveness of therapy for movement disorders. Spinal cord stimulation may be improved by pairing stimulation with EMG activity, increasing the speed of recovery from injury. VNS driven by rehabilitation activities that reach a success threshold may lead to increased functional outcomes. And BCI systems that convey sensory information during movement may allow the users to have the most effective control over motor effectors.

The innovations in precise, closed-loop stimulation capitalize on novel small, high density microelectrodes. Increasingly, these electrodes can be effectively targeted to the optimal location within neural tissue, to maximize recording and stimulation effectiveness. In addition, advances in non-invasive stimulation has the benefits of precise, responsive electrical stimulation without the risk of invasive implanted devices. Pushing the boundaries of technology development to create robust yet precise neural interfaces allows for unique and localized electrical stimulation. It also enhances physiological sensing, opening the door for fully closed-loop therapeutic systems.

Advances in neural interface technology, stimulation delivery, closed-loop systems allow devices to respond to each individual’s neurophysiology, optimizing device performance and recovery of motor function. Critically, these new neurostimulation devices are customized to meet each user’s individual needs and respond to real-time physiological cues. The innovative device technology opens up the frontiers for precise and responsive closed-loop neurostimulation systems to harness endogenous neural plasticity mechanisms to repair or compensate for poorly-functioning motor circuits. Moreover, increased accessibility to neural plasticity has implications that transcend the beyond movement disorders: in essence, it can allow injured or poorly-functioning nervous systems to maximize their repair capacity. Electrical stimulation is expanding the boundary of what is considered possible in restoration of motor function.