Electroceutical Targeting of the Autonomic Nervous System

Abstract

Autonomic nerves are attractive targets for medical therapies using electroceutical devices because of the potential for selective control and few side effects. These devices use novel materials, electrode configurations, stimulation patterns, and closed-loop control to treat heart failure, hypertension, gastrointestinal and bladder diseases, obesity/diabetes, and inflammatory disorders. Critical to progress is a mechanistic understanding of multi-level controls of target organs, disease adaptation, and impact of neuromodulation to restore organ function.

Introduction

The autonomic nervous system (ANS) regulates a vast landscape of organ and tissue function using multi-level reflex control involving both central and peripheral neural networks. In addition to the well-recognized long-loop reflexes with sympathetic and parasympathetic efferent arms are complex peripheral neural networks in intimate association with their target tissues (9, 46). These peripheral neural networks are located in both the thorax (intrinsic cardiac nervous system) and viscera (enteric nervous system). Functions controlled by these interdependent neural networks include regional cardiac activity (chronotropy, inotropy, dromotropy, and lusitropy), arteriole (and venule) diameters, exocrine and endocrine pancreatic cell secretions, gastrointestinal (GI) motility and secretions, urinary bladder contraction and urination, tracheal and bronchial sensations, immune system function, and mobilization of energy stores from fat and liver (9, 33, 46, 61, 72, 111). Cranial and spinal nerves of the ANS are mixed populations of afferent and efferent fibers with cell bodies in peripheral ganglia and CNS. These interdependent neural networks are not relay stations but adaptable processing networks that allow for flexibility to respond to a myriad of competing endogenous and exogenous stressors.

ANS peripheral nerves and ganglia are attractive targets for electroceutical therapeutic approaches; these tissues typically have a favorable location for surgical interventions compared with invasive surgery into the brain. Indeed, peripheral nerves and ganglia lack the comparative complexity of targeting regions of the CNS. CNS areas contain diverse functions, with anatomically overlapping cell populations and fibers of passage; lowering an electrode into a deep brain structure can produce non-selective effects and trauma to surrounding areas. Electroceuticals use engineered electronic devices, with unique bio-compatible materials, form-fit nerve-cell interfaces, hardware components, and computer software to modulate sensory and motor neuronal signaling. In contrast to systemic pharmacotherapy, electroceutical devices have potential for highly selective organ-specific treatment, fewer side effects, and precise spatio-temporal therapeutic control. This is an old idea; for example, almost 50 years ago, Braunwald tested electrical stimulation of the carotid sinus nerve to treat cardiac disease (22). Then, as now, it is contingent on mechanistic-based preclinical studies to pave the way for translation to the clinic.

Research more than a decade ago led to the development of FDA-approved vagus nerve stimulation (VNS) devices to treat epileptic seizures (3) and depression (2, 83). Cervical vagus nerve bipolar circumferential electrodes were developed in the 1990s—the Cyberonics helical cuff electrode (FIGURE 1). Stimulation through this device, typically applied as 20–30 Hz of stimulation, has demonstrated effectiveness for controlling epileptic seizures; however, it produces adverse effects in some patients, including hoarseness, cough, dyspnea, pain, paresthesia, nausea, and headache (84). Many of these side effects result from non-selective stimulation of the cervical vagus trunk, which contains >100,000 afferent and efferent nerve fibers from almost all thoracic and abdominal organs (52). Unfortunately, this VNS approach is only effective in reducing epileptic seizures in ~1/3 of patients because stimulation levels must be reduced to avoid adverse effects (17). More recently, this approach was reported to control treatment-resistant depression, but again with similar off-target effects (14, 77). It remains unclear which mechanisms are affected by cervical VNS to produce therapeutic control of epileptic seizures or depression, but functional imaging studies indicate large-scale changes in CNS activity (29, 30). These early VNS approaches highlight the challenges of developing electroceutical devices, including how to achieve functional selectivity, reduce side effects, and determine physiological readouts to assess mechanisms of action and therapeutic efficacy in individual patients; importantly, these issues not only apply to the time of implant but to weeks and years into the future, which include the possibility of changes in device functioning and organ physiology.

FIGURE 1.

Electroceutical therapies approved by the FDA

The Cyberonics vagus nerve stimulation system was approved in 1997 to control epileptic seizures and later approved for use to treat depression (2005); this device is attached to the left cervical vagus trunk. The Maestro system was approved in 2015 to treat obesity and has attachments to the ventral and dorsal gastric branches of the vagus nerve. The Medtronic InterStim system was approved in 2011 to treat urinary incontinence; this system is placed adjacent to the third sacral nerve root.

The New Landscape of Electroceuticals

Over the last 5 years, there has been substantial increase in U.S. government and private research focused on electroceuticals, including development of novel bio-compatible materials, miniaturized electronics, software, stimulation patterns, and electrode geometry. A central goal of these efforts is to produce electroceuticals with greater selectivity, enhanced efficacy, and fewer complications (18). These efforts also include a significant focus on determining the biological mechanisms of how electroceutical devices produce therapeutic effects as a lead up to targeting a patient population (108). In 2013, bioelectronic medical research, which includes electroceuticals, was catalyzed by the launch of Bioelectronics R&D at Glaxo-Smith Kline, followed closely in 2014 by the ElectRx program at DARPA (66) and the NIH SPARC (Stimulating Peripheral Nerves to Relieve Conditions) initiative (https://sparc.science; Refs. 4, 51). New bioelectronic device approaches, many funded by these initiatives, include novel applications of electrical and other energy transfer devices, with combined investments of more than $500 million to teams of neuroscientists, engineers, and clinicians. Below, we highlight recent advances in electroceuticals, including application to an array of diseases, device configurations, and patterns and parameters of stimulation.

Thoracic Targets

Within the thorax, targeting of cardiac disease is arguably the clearest example of electroceutical device development in recent years, with significant insight into physiological effects of stimulation and effective device configurations and parameters; as such, these efforts could provide a useful template for understanding effects on other organ systems, e.g., the GI tract.

Cardiac Disease

Cardiovascular disease is the leading cause of morbidity and mortality in the U.S. and the world (110). The ANS plays a central role in the pathogenesis of several cardiovascular diseases such as atrial fibrillation, hypertension, myocardial infarction, and ventricular tachyarrhythmias/fibrillation, and in the progression of heart failure (97, 106). Progression of cardiac disease reflects adverse remodeling in the heart and cardiac nervous system. In general, this is characterized by a hyperdynamic sympathetic response with withdrawal of central parasympathetic tone in an attempt to maintain cardiac output (9, 43, 45). Figure 2 summarizes the structural organization for autonomic control of the heart, GI tract, and spleen, with primary intervention points for neuromodulation based therapies. In this review, we will focus on device-based electroceuticals. The reader is referred to recent reports for a comprehensive review of neuromodulation-based therapies for cardiac disease (9, 97).

FIGURE 2.

Autonomic nerve connections and electroceutical targets for the cardiac system, gastrointestinal system, and spleen

BAT, baroreceptor activation therapy; AMT, axonal modulation therapy, applied to the T1-T2 paravertebral chain to control sympathetic projections to heart; CVS, cervical vagus nerve stimulation; GVS, gastric vagus nerve stimulation; SCS, spinal cord stimulation. Other systems (not shown) also contain autonomic nerve foci for targeting lungs, liver, pancreas, kidney, and bladder function.

Recent preclinical studies have pioneered the concept of the neural fulcrum as the basis for effective control of multi-level neural circuits with electroceutical therapy for treatment of specific disease processes (11). The neural fulcrum is based on the concept that, when bioelectric interventions push the autonomic neural networks in one direction, the endogenous reflex control pushes back. For example, with cervical VNS, central parasympathetic drive is reduced by afferent activation, starting at low stimulus intensities, which is counteracted by direct activation of parasympathetic efferent preganglionic axons at higher intensities (11, 12). Effective bioelectric control of the entire network is achieved when these two counteracting events are equal and opposite. The net effect is that, although there are minimal changes in basal function, the response of the entire network to stress is restrained, counteracting the hyperdynamic reflex responses commonly associated with disease progression. These studies have further shown that the neural networks can be conditioned to bioelectric interventions, allowing for titration to a broad spectrum of stimulation protocols with minimal side effects (11). Such titrations utilize lower frequencies and pulse widths than those used in FDA-approved epileptic seizure and depression VNS applications. Recent clinical trials, based on the neural fulcrum, are moving forward for treatment of heart failure (39, 90). VNS likewise has shown efficacy, with potential for clinical application, for management of arrhythmias (93, 101).

Spinal cord stimulation.

Spinal cord stimulation (SCS) is FDA-approved for pain management, including angina (44). With reference to cardiac disease, SCS impacts multiple levels of the cardiac nervous system and the heart itself (8). SCS stabilizes reflex processing in the spinal cord (40, 41), stellate ganglia (10), and intrinsic cardiac ganglia (13), and induces a cardioprotective state within cardiomyocytes (100), likely involving changes in substrate utilization. Two recent clinical studies have evaluated the efficacy of SCS in heart failure. A prospective, multisite clinical trial known as SCS Heart (104) enrolled patients with NYHA functional class III symptoms to undergo continuous SCS for 6 months. This study reported significant improvements in cardiac function and NYHA class. In contrast, in a randomized controlled clinical trial called DEFEAT-HF (113), SCS was applied to patients with a similar clinical profile of heart failure but with a duty cycle of 12 h on and 12 h off. This second study showed no significant difference in cardiac function with SCS. From a mechanistic perspective, SCS exhibits a neural memory function of ~45 min when transitioning from on-phase to off-phase (13). The difference in results between these two studies may distill down to differences in protocols. Such results also reinforce the theme of this review—that rational clinical applications require a fundamental understanding of what neural processes and neural target interfaces are being impacted and that these interactions need to be considered in both a temporal and a spatial context.

Baroreceptor activation therapy.

Electrical stimulation of primary sensory neuritis of the carotid sinus, known as baroreceptor activation therapy (BAT), has been evaluated with respect to hypertension and heart failure. This approach is based on the premise that such stimulation will reduce sympathetic outflow while augmenting parasympathetic tone. Although BAT has been effective for patients with resistant hypertension (37), its efficacy in heart failure remains unresolved (112).

Axonal modulation therapy.

Electroceutical therapies, and other forms of energy delivery, have the potential to increase or decrease neural activity, depending on how they are delivered. One promising line of investigation utilizes either kilohertz frequency AC [KHFAC (63)] or charge-balanced DC [CBDC (107)] stimulation to reduce or completely block neural activity in an on-demand and scalable fashion. Novel bioelectric interventions to control sympathetic outflows are of obvious clinical relevance. The T1-T2 region of the paravertebral chain was recently identified as a critical nexus point for sympathetic control of the heart (27). Application of either KHFAC (26) or CBDC (35) to the T1-T2 paravertebral chain produced rapid, scalable, and reproducible block of sympathetic outflow to the heart without compromising basal cardiac function. Moreover, this blocking technology markedly reduced the potential for sudden cardiac death even in the setting of chronic ischemic heart disease (35). As such, in some cases, this therapeutic approach has the potential to supplant cardio-defibrillators.

Abdominal Targets

Below the thorax, the vagus nerve contains critical afferent and efferent pathways for bi-directional communication between brain and abdominal organs, including the GI tract and pancreas (15, 31, 103). There are also small vagal branches innervating the hepatic portal vein, which could be important for sensing GI hormones and absorbed nutrients (15). The GI tract receives both parasympathetic (vagus) and sympathetic innervation, whereas the bladder contains only sympathetic innervation. Our initial discussion focuses on the vagus nerve because it is a complex therapeutic nerve target, including pathways linked to GI, metabolic, and inflammatory diseases. We end this section with a discussion of electroceutical research applied to treat bladder dysfunction.

Functional GI Disease

Functional GI diseases are disorders of gut-brain interaction, primarily motility disturbances. These diseases affect >20% of the population (102) and include esophageal disorders, gastresophageal reflux disease, functional dyspepsia, gastroparesis, chronic intestinal pseudo-obstruction, postoperative ileus, irritable bowel syndrome, diarrhea, and constipation. We will focus on electroceutical devices used to treat gastroparesis and its associated chronic nausea and vomiting (64). For a more complete review of electroceutical targeting of GI motility disorders, the reader is referred a recent review (34).

Regulation of gastric slow-wave responses originating from the gastric pacemaker region is a mechanistic target of treatments for GI motility disorders (81). The Medtronic Enterra device is FDA-approved for controlling the symptoms of gastroparesis, including nausea and vomiting (59). This device contains bipolar electrode leads attached to the gastric serosal surface in approximately the location of the gastric pacemaker. This gastric electrical stimulation (GES) device uses stimulation of 0.3-ms pulse duration, 0.1 s on and 5 s off up to 6 mA. It remains unclear whether the device stimulates extrinsic nerve fibers, such as the vagus, enteric neurons, or muscle, to produce therapeutic effects. The effectiveness of Enterra therapy, however, is controversial and includes few controlled studies, which show slight or no effects of stimulation on emesis (or nausea) when comparing “on” vs. “off” stimulation conditions (82). Also, in a preclinical model in musk shrew, electrical stimulation of the stomach surface triggered emesis (55).

There are also promising preclinical studies to control gastric motility. SCS increases gastric tone and accelerates gastric emptying in rats (99). Furthermore, a recent report showed that cervical VNS in rats can control gastric emptying, using 10 Hz at 20 s on and 40 s off, 0.6 mA, for 4 h (69). These techniques could also be applied to the control of food intake to produce satiation and limit meal size, which could control body weight gain.

Obesity and Diabetes

Obesity affects nearly 40% of U.S. adults and is associated with high levels of comorbidities, including cancer, cardiovascular disease, and diabetes (49, 60). Effective obesity treatments without significant complications are lacking, but there is compelling preclinical evidence that VNS, either at cervical or abdominal levels, can suppress body weight; the reader is referred to more comprehensive reviews for a full discussion (36, 89). Preclinical studies show that abdominal VNS at up to 30 Hz can reduce food intake and body weight (e.g., 105), but VNS can also stimulate emetic pathways (6, 54), which in preclinical studies could be confused with satiation of food intake. Importantly, rats and mice were used in VNS reports on feeding and body weight, but these species lack an emetic reflex, unlike humans, musk shrews, ferrets, and many other mammals (54). Although a change in gastric emptying and motility could be a possible mechanism for the effects of VNS on body weight (see Functional GI Disease above), there are also other VNS-induced responses that could change metabolism and food processing, including insulin, glucagon, and gastric acid secretion (16).

Abdominal VNS is a recent surgical option for obese patients, which is an attractive alternative because it can be tailored to each patient using different stimulation parameters and, in contrast to bariatric surgery, can be turned off or removed. Gastric VNS, known as vBloc therapy and approved by the FDA, results in weight loss of ~11% in moderately obese patients (76). Although vBloc therapy is presented as a nerve signal-blocking approach using high-frequency alternating current (57, 94) (FIGURE 1, Maestro system), the parameters used do not appear to block neural signaling (88). vBloc therapy was also shown to decrease blood HbA1C levels compared with baseline and could be a useful for treatment of diabetes (95). In addition, there are reports of adverse events with vBloc therapy, including nausea, dyspepia, and abdominal pain (7, 96).

In contrast to continuous stimulation with vBloc therapy, research is being conducted on closed-loop technologies in which stimulation is triggered by detection of specific biological measures; for example, recording the nerve compound action potential and optimizing stimulation parameters for activation of specific nerve fiber types (109). One application includes the use of gastric distension from a meal, to trigger GES, which results in weight loss in obese subjects, similar to adjustable gastric band therapy (53, 75). More precise, meal-related approaches are in development; for example, using the detection of cholecystokinin (CCK), an appetite-suppressing gut hormone, to trigger abdominal VNS (74).

Inflammatory Diseases

Preclinical studies show cervical VNS-induced anti-inflammatory effects, as measured by a reduction in systemic inflammatory cytokines (87). Recent clinical testing in patients with inflammatory diseases show promising results. In five of seven Crohn's disease patients tested, VNS resulted in clinical remission, using continual stimulation at 0.25 mA, 30 s on and 5 min off, pulse width 0.5 ms, and frequency of 10 Hz for 6 mo (20). Moreover, VNS in 17 rheumatoid arthritis patients for 1 min up to four times a day using 0.25 to 2 mA, 0.25-ms pulse width, and 10-Hz frequency for 84 days reduced the disease (65). Both clinical studies have used the left cervical VNS system from Cyberonics composed of a helical cuff electrode (FIGURE 1). VNS therapy could potentially be applied broadly to many diseases with inflammatory components, including cardiac disease, obesity, diabetes, and cancer. For example, a recent study in mice reports VNS upregulates TFF2 protein in the spleen, which leads to suppression of cellular pathways that lead to gastrointestinal cancer (42). The reader is referred to recent reviews for a full discussion of VNS-induced anti-inflammatory effects (21, 71, 86).

The mechanisms of VNS on inflammation are not fully known. One theory is that VNS activates a cholinergic anti-inflammatory pathway, which is mediated by vagal efferents that synapse in the celiac ganglia to stimulate post-ganglionic parasympathetic neurons that project to the spleen (FIGURE 2); splenic activation would produce release of acetylcholine from lymphocytes to suppress release of TNF-α. Current anatomic evidence indicates that this pathway is either very sparse or does not exist (28). Other pathways for the effect of VNS on immune responses could include stimulation of the gastrointestinal immune system or reflex action through the CNS affecting sympathetic outflow to the spleen (FIGURE 2) (21). One unique aspect of VNS action is a strong memory component. Unlike most electroceutical effects, VNS anti-inflammatory actions extend for many hours after brief stimulation. For example, in humans, 30 s of VNS (1 mA, 20 Hz, and 0.5-ms pulse duration) produced a decrease in TNF-α, IL-6, and IL-1β at 4 h after stimulation (65).

Bladder Dysfunction

Although bladder dysfunction is less likely to result in death than cardiac dysfunction and diabetes, incontinence and related problems affect a large population, with as many as 17% of women reporting moderate to severe incontinence and bladder dysfunction, costing over $60 billion each year in the U.S. (47, 80). Additionally, urinary tract infections related to neurogenic bladder function are a leading cause of hospitalization and decreased quality of life in individuals with spinal cord injury (73, 85). The Medtronic Interstim system is an FDA-approved device (approved in 2011) for sacral nerve stimulation for treatment of urinary incontinence (1), with a longitudinal four-contact lead, similar to an SCS device but placed adjacent to the third sacral nerve branch (FIGURE 1). Stimulation between a pairs of contacts is typically tuned to produce therapeutic effects with 14-Hz continuous stimulation (5). As with SCS, this approach is commonly initiated using a test implant phase followed by a permanent placement in select patients. In these patients, clinical trials show sacral nerve stimulation can reduce urinary incontinence (50). A new multi-center clinical trial is currently underway to explore the efficacy of this system for neurogenic bladder dysfunction after spinal cord injury (92). Other approaches to control bladder function via peripheral nerve electrical stimulation have been explored in animal models. These include studies that rely on cuff electrodes wrapped around the pudendal nerve to stimulate afferent pathways to evoke reflexive bladder contractions. Multiple studies have demonstrated efficacy in bladder voiding by stimulating this pathway, with one study demonstrating the possibility for generating differential responses (i.e., continence vs. voiding) depending on the frequency of stimulation (19) and another demonstrating that stimulation trains with non-constant frequencies can improve bladder activation (23). Other studies have demonstrated that it is possible to both monitor bladder pressure (24, 62) and stimulate reflex pathways (25) with electrode arrays inserted into the sacral dorsal root ganglia, paving the way for systems that can achieve closed-loop control of continence and voiding.

Integrated Lessons from the Electroceutical Parameter Space

Table 1 summarizes stimulation patterns used in studies discussed above. These studies span multiple organ systems, disease states, and target nerves, and the stimulation frequencies and duty cycles reflect that diversity. Some experiments rely on constant stimulation at a single frequency (104), whereas others involve short bursts of stimulation occasionally throughout the day (65) or high-frequency stimulation to block nerve transmission (95). In almost all cases, these parameters were determined empirically, with little mechanistic understanding of the underlying effects of stimulation. Often, parameters developed in animal models are applied directly to human studies, despite the sometimes vast difference in nerve anatomy between humans and animals (48). Furthermore, in nearly all of the studies described here, stimulation targeted an entire nerve (e.g., the cervical vagus) or a large region of the spinal cord. These complex structures contain thousands of both afferent and efferent neurons innervating organs throughout the body. Electrodes with contacts that wrap completely around the nerve circumference are simpler to manufacture than those with multiple independent contacts, but these simpler devices will always be limited to stimulation of large populations of neurons, with little control over selectivity. Moreover, there is a limitation in stimulating small-diameter fibers, such as C-fibers, because more current is required as fiber diameter decreases (91). These limitations suggest that future studies should involve a heavy focus on developing a mechanistic understanding of the effects of electrical stimulation on target nerves and organs. Furthermore, computational models of the electrode nerve interface may be informative in the development of new electrodes and tuning of stimulation parameters. In fact, improvements in imaging technology may allow for development of patient-specific computational models that can automatically calculate optimal stimulation parameters without the need for costly office visits to empirically adjust stimulation (32).

Table 1.

Effect of electroceutical stimulation patterns on functional responses

Organ System	Disease State	Target Organ	Target Nerve	Model	Spatial Selectivity of	Frequency	Duty Cycle	Proposed Mechanism of Action	Study Results: Biomarkers/	Refs.
Brain	Epilepsy	Brain stem nuclei	Cervical vagus	Human	Non-selective	20–30 Hz	30 s on: 5 min off; no known effect of duty cycle	Modulation of brain stem and higher center neural networks	Reduction in seizure frequency	38, 84
	Depression	Multiple CNS targets	Cervical vagus	Human	Non-selective	1–30 Hz, typically 20 Hz	7–60 s on: 0.2–180 min off	Synchronization of orbito-frontal activity; frontal lobe slow waves	Improvement in depression scores	83
Thoracic Organs	Normal, HFpEF, ischemic heart disease	Heart	Cervical vagus	Dog, guinea pig	Non-selective	2–30 Hz, most effective below 10 Hz	14 s on: 66 s off	Anti-adrenergic effects, blunting of intrathoracic reflexes, changes myocyte substrate utilization	Bradycardia; tachycardia	93
	Heart failure	Heart	Cervical vagus	Human	Non-selective	10 Hz	Continuous or intermittent	Anti-adrenergic effects	Increase in left ventricular ejection fraction	90
	Atrial fibrillation	Heart	Cervical vagus	Dog	Non-selective	15 Hz	3 min on	Blunting of intrinsic cardiac reflexes	Stabilization of atrial electrical function	93
	Postoperative atrial fibrillation	Heart	Cervical vagus	Human	Non-selective	20 Hz	72 h on	Inhibition of cardiac autonomic nervous system and suppression of inflammation	Decrease in postoperative atrial fibrillation	101
	Heart failure	Heart	Spinal cord	Human	Non-selective; paresthesia covering chest	50 Hz	Continuous	Modulation of sympathetic and parasympathetic nervous system	Decrease in rate of heart failure; decrease in NYHA class	104
	Systolic heart failure	Heart	Spinal cord	Human	Non-selective; paresthesia covering chest	50 Hz	12 h on: 12 h off	Modulation of sympathetic and parasympathetic nervous system	Trial failed: no change in left ventricular end-systolic volume index	113
	Hypertension	Heart	Carotid sinus	Human	Non-selective	Unavailable	Continuous	Stimulation of baroreflex pathways	Decrease in blood pressure	37
	Normal	Heart	T3 paravertebral ganglion	Pig	Non-selective	15 kHz	30 s on	Block of sympathetic excitation	Lengthening activation recovery interval	26
	Post-myocardial infarction arrhythmia	Heart	T1-T2 paravertebral ganglion	Pig	Non-selective	Charge-balanced DC	30 s on	Block of sympathetic excitation	Lengthening of ventricular effective refractory period	26
Abdominal Organs	Obesity	Stomach	Abdominal vagus	Human	Non-selective	5 kHz	5 min on: 5 min off	Block of vagal signaling	Decrease in body weight	57, 94
	Type 2 diabetes	Stomach	Abdominal vagus	Human	Non-selective	5 kHz	5 min on: 5 min off; 14 h/day	Block of vagal signaling	Decrease in body weight and HbA1C blood levels	95
Abdominal Organs (cont.)	Obesity	Stomach	Abdominal vagus	Minipig	Non-selective	30 Hz	30 s on: 5 min off	Activation of multiple brain areas including olfactory bulb	Decrease in weight gain and food consumption	105
	Obesity	Stomach	Serosal surface of stomach	Human	Non-selective	40–120 Hz	Closed-loop, when eating is detected	Stimulation of vagal afferents	Sensation of satiety, decrease in body weight	53, 75
	Gastroparesis	Stomach	Serosal surface of stomach	Human	Non-selective	14 Hz	2 pulses on: 5 s off	Pacing of stomach activity, reflexive gastric accommodation	Increased gastric motility	59
	Gastroparesis	Stomach	Cervical vagus	Rat	Non-selective	10 Hz	20 s on: 40 s off	Activation of afferents triggering reflexive pyloric relaxation	Pyloric opening, increased gastric motility and emptying	69
	Crohn’s disease	Immune system	Cervical vagus	Human	Non-selective	10 Hz	30 s on: 5 min off	Increased vagal tone and equilibrated autonomic balance	Change in Crohn’s disease activity index	20
	Rheumatoid arthritis	Immune system	Cervical vagus	Human	Non-selective	10 Hz	30 s on: 5 min off	Increased vagal tone and equilibrated autonomic balance	Decreased TNF blood concentration	65
	Urinary incontinence	Bladder	S3 nerve root	Human	Non-selective	14 Hz	Continuous	Activation of afferents to trigger reflexive bladder control	Improved urinary continence	50
	Urinary incontinence	Bladder	Pudendal nerve	Cat	Non-selective	2–100 Hz	2–40 s on	Activation of afferent or efferent pudendal pathways	Low-frequency: continence; high-frequency: micturition	19
	Urinary incontinence	Bladder	Pudendal nerve	Cat	Non-selective	1–100 Hz	20 s on	Activation of pudendal afferents to trigger reflex bladder contraction	High frequency bursts improve voiding	23
	Urinary incontinence	Bladder	S1-S2 dorsal root ganglia	Cat	Highly selective microstimulation	1–33 Hz	5 s on	Activation of afferent pathways to trigger bladder contraction and relaxation	Microstimulation can generate bladder contractions	25

This table is not comprehensive; it only provides an overview of some the stimulation parameters used in electroceutical research.

The Future

Modern electroceutical devices use electrodes with small surface areas to achieve selective activation in the nerve interface and rely on miniaturization of all components, wireless communication and charging, low power consumption, closed-loop control, multiple contacts (often circumferential) with improved selectivity, and flexible self-sizing materials. We briefly discuss important active research topics.

Selectivity

An ongoing challenge is to refine electrode position in anatomical space and temporal stimulation patterns to achieve increased selective targeting of physiological pathways to treat disease progression. This is especially important in functionally heterogenous nerves, such as the vagus, which possesses bi-directional communication with numerous organs. Using more electrode contacts on a single device probe substantially increases the stimulation parameter space. Each of these electrodes can have unique stimulation amplitude, pulse duration, duty cycle, and stimulation frequency, and combinations of electrodes can be used in a multipolar configuration to improve selective targeting of sub-regions of the nerve. Additional electrodes lead to an exponential expansion in the possible stimulation parameter space, requiring advanced analytics and machine learning approaches to determine optimal stimulation parameters and to control stimulation in real-time for closed-loop applications (78).

Closed-Loop Control

Closed-loop control depends on a dynamic readout from which neuromodulation interventions can be adjusted. One simple example of this is the implantable cardio-defibrillator. When an abnormal electrical signal is detected from the heart, a current discharge is delivered to the heart in an attempt to “reset” cardiac electrical stability. The process repeats itself as necessary until normal rhythm is restored, the patient is transferred to in-patient care, or attempts at resuscitation are insufficient. As one contemplates future directions for closed-loop control, there are multiple potential sites for sensing that need to be form-fit to the homeostatic mechanisms that are to be targeted. These include 1) sensors designed to quantify neural activity from axons of passage or peripheral ganglia, 2) organ-specific readouts for electrical and/or mechanical function, or 3) chemical sensors deployed to the interstitial or vascular space to assess indexes of metabolism, neurochemical release, or markers of the immune response. A variety of these sensors could be used to “tune” the level of neuromodulation during initial treatment phases and then exert a sustained function during ongoing therapy. The biological systems themselves exhibit plasticity with time; neuromodulation therapies need to transition to closed-loop control so that they to can exhibit plasticity. Additionally, advancements in recording technologies are beginning to be applied and ultimately should be included in closed-loop device applications. Figure 3 shows two recent examples of recording whole organ function of the heart and stomach for closed-loop control of stimulation.

FIGURE 3.

Representative approaches to physiological sensing of heart and stomach organ function

These technologies can be used in closed-loop applications. A: stretchable thin-film array form-fit for recording multi-point unipolar electrograms from porcine cardiac surface with capabilities for delivering point-specific pacing anywhere within the array surface. Right: activation sequence in response to focal stimulation, as indicated in schematic. B: multi-site gastric planar electrodes. Each planar probe contains four electrode sites for microenvironment sensing or for use to compare with other recording sites across the serosal surface of the stomach (78). Right inset: from the ferret stomach.

Wireless Power Delivery

Recently, it was demonstrated that focused beam-forming energy can deliver wireless power to implanted electrodes (70). This technology displays little attenuation of signal strength using depths that are well beyond superficial implantation, such as intra-gastric placement in a porcine model (70). This potentially removes the need for an implanted battery, with possible follow-up surgery for device replacement. In addition, this approach, using a smaller form-factor device, could require less complex surgery and reduced tissue trauma. In addition, there is the recent use of transcutaneous electrical stimulation (67), but this is likely to be limited in specificity and depth of penetration.

Form-Fitting Electrodes

It is often difficult to size electrodes for close nerve contact, which is critical for reducing impedance to tissue while avoiding traumatic nerve compression. With lower impedance, less current is required to produce selective biological effects, which also reduces off-target responses. Cuff electrodes are manufactured in a variety of standard sizes and are generally nontraumatic if a proper size is chosen; but individual anatomy can make this difficult to accomplish in practice. Recently it was shown, using tract tracing techniques in a rat model, that cuff electrodes can produce trauma to efferent fibers if improperly sized (98). Newer materials can stretch and stick to nerves with improved contact with less trauma, e.g., hydrogel (56).

Other Technologies

Although electrical stimulation with implanted devices has been the dominant approach for modifying neural function, there are devices delivering different energy sources, such as various light wavelengths; for example, infrared and optogenetic approaches (58, 68). Additional methods include transcutaneous magnetic stimulation (79).

Sustained progress in electroceutical approaches should be closely connected to developing a deeper understanding of the underlying biological effects (108). Traditional device development, such as early efforts for VNS, focused on clinical end points, with much less research on biological mechanisms; however, these biological insights, including precise molecular and neurophysiological changes, are the golden keys to designing highly selective and effective electroceutical therapies.