Next-Generation Bioelectric Medicine: Harnessing the Therapeutic Potential of Neural Implants

Abstract

Bioelectric medicine leverages natural signaling pathways in the nervous system to counteract organ dysfunction. This novel approach has potential to address conditions with unmet needs, including heart failure, hypertension, inflammation, arthritis, asthma, Alzheimer's disease, and diabetes. Neural therapies, which target the brain, spinal cord, or peripheral nerves, are already being applied to conditions such as epilepsy, Parkinson's, and chronic pain. While today's therapies have made exciting advancements, their open-loop design—where stimulation is administered without collecting feedback—means that results can be variable and devices do not work for everyone. Stimulation effects are sensitive to changes in neural tissue, nerve excitability, patient position, and more. Closing the loop by providing neural or non-neural biomarkers to the system can guide therapy by providing additional insights into stimulation effects and overall patient condition. Devices currently on the market use recorded biomarkers to close the loop and improve therapy. The future of bioelectric medicine is more holistically personalized. Collected data will be used for increasingly precise application of neural stimulations to achieve therapeutic effects. To achieve this future, advances are needed in device design, implanted and computational technologies, and scientific/medical interpretation of neural activity. Research and commercial devices are enabling the development of multiple levels of responsiveness to neural, physiological, and environmental changes. This includes developing suitable implanted technologies for high bandwidth brain/machine interfaces and addressing the challenge of neural or state biomarker decoding. Consistent progress is being made in these challenges toward the long-term vision of automatically and holistically personalized care for chronic health conditions.

Introduction

Bioelectricity is a fundamental unit of communication in the body. As bioelectric signals travel through the nervous system, they carry vital messages that control organ function. These electrical signals are rich with information that can be recorded and incorporated into neural digital therapies, where they may be manipulated for therapeutic effects.

Bioelectric medicine leverages these natural signaling pathways in the nervous system, achieving therapeutic effects through implanted devices that stimulate targeted nerves. With its many pathways and targeted effects, the nervous system provides a natural route for therapeutic interventions to counteract dysfunction within the body, and it holds real promise as an alternative or complementary therapy for conditions not adequately controlled by traditional pharmaceuticals.¹

The global bioelectric medicine market is estimated at $18.6 billion in 2020, a figure that is forecast to grow to $28.5 billion in 2026.² Modulating neural activity could provide treatments for many diseases with unmet needs, including heart failure,³ hypertension,⁴ inflammation,⁵ arthritis,⁶ asthma,⁷ Alzheimer's disease,⁸ and diabetes.⁹ New approaches also aim to interfere with the neuronal signaling that underpins cancer progression,¹⁰ and it has been suggested that neuromodulation therapy might help control the cytokine release syndrome seen in some COVID-19 patients, with potential for a cholinergic anti-inflammatory pathway to be targeted for therapeutic purposes.¹¹

Existing Neural Therapies

Neural therapeutic approaches are already clinically applied to a variety of neurological, cardiovascular, and inflammatory conditions (Fig. 1). These target either the central nervous system (CNS), the brain and spine, or the peripheral nervous system (PNS), the nerves that communicate between the brain and organs. While noninvasive neural stimulation techniques also exist,¹² this article will focus on current and future uses of neural implants.

FIG. 1.

Examples of neural therapeutic targets and applications using a bioelectric approach. Neural therapeutic targets include the brain (DBS),¹³ the vagus nerve,¹⁸ the spinal cord,¹⁶ and the sacral nerve, but this figure is by no means exhaustive and there are potential opportunities for therapeutic stimulation throughout the CNS and PNS.¹ The asterisk indicates uses which are approved by the FDA. DBS, Deep brain stimulation; PNS, peripheral nervous system; CNS, central nervous system.

In the CNS, use of implantable electrodes to treat traumatic and neurodegenerative disorders is established and continues to expand. Deep brain stimulation (DBS) was first approved for essential tremor in Parkinson's disease in 1997 and has since been used by tens of thousands of patients.^13,14 Other potential DBS targets include pain, obsessive compulsive disorder, major depression, Tourette syndrome, and Alzheimer's disease.¹⁵ The spinal cord is also a common target, particularly for relieving chronic pain. Nevro Corp's Senza II Spinal Cord Stimulation System, for example, was approved by the FDA in August 2018. This applies electrical pulses to the spinal cord through a battery-powered generator implanted under the skin to alleviate chronic intractable pain of the trunk or limbs.^16,17

The PNS can provide lower-risk surgical options and more direct access to the neural circuits that control dysfunctional organs. The vagus nerve, which descends from the brainstem and branches to organs, including the heart, lungs, stomach, and pancreas, is a popular target for stimulation-based therapies (Fig. 2). Vagus nerve stimulation (VNS) was first approved by the FDA for epilepsy in 1997, followed by approval for depression in 2005, and has been used in over 100,000 patients.¹⁸ Since then, evidence of anti-inflammatory effects has spurred interest in low-frequency VNS for diseases such as rheumatoid arthritis and inflammatory bowel disease.¹⁹ The role of the vagus in visceral control has inspired the application of VNS to heart failure, hypertension, and obesity.²⁰

FIG. 2.

Schematic of an implanted vagus nerve stimulator. Electrical pulses originate from the pulse generator implanted in the chest. The electrical pulses pass through the electrodes to stimulate the vagus nerve. Right or left VNS has been shown to have therapeutic downstream effects for a variety of conditions, including epilepsy, depression, hypertension, and heart failure.²⁰ VNS, vagus nerve stimulation.

Open-Loop Stimulators Lack Therapeutic Precision

While existing stimulators have made exciting advancements, these do not work for everyone. Results can be variable and device adoption is limited, in part due to uncertainty about how to select stimulations for optimal therapeutic effects.^14,21,22

Instead, stimulation parameters for most existing bioelectric devices are manually programmed by the clinician. Programming is done during implantation surgery and follow-up appointments in the weeks and months afterward, which is time consuming and burdensome for both clinicians and patients. Even after this process, stimulation can become uncomfortable and lose effectiveness over time.^23,24

Such devices are considered “open loop”: they tonically stimulate according to set parameters without receiving any feedback from the stimulation's effects. This means that they may become ineffective due to changes in tissue surrounding the implant or in nerve excitability, or to patient movement. To adapt, devices must receive feedback from the patient and decode this to adjust stimulation parameters. This process is known as “closing the loop.” (Fig. 3).

FIG. 3.

Open loop neural interface system (A) compared with a closed-loop neural interface system (B). In an open-loop system, preset stimulations are consistently applied by the device to the nerve. In a closed loop system, sensor data from the nerve are fed back to the device and used to titrate stimulation parameters. “Closing the loop” has been shown to improve patient outcomes for multiple applications^28,31,32 and allows for increased personalization of treatments.

Closed-Loop Systems Provide Novel Insights to Guide Therapy

The body's natural electricity controls the activity of various organs and receives sensory information; these signals are biomarkers that can inform choices of when, where, and how to apply stimulation. Closed-loop systems take advantage of this. Closing the loop can refer to the addition of any feedback to the bioelectric system, but in the case of neural interfaces it typically suggests the decoding of neural activity into interpretable neural biomarkers and use of those decoded biomarkers to alter stimulation parameters.²⁵

All sensorimotor processing that passes through the nervous system can theoretically be decoded from neural signals. Implanted electrodes on information-rich targets, such as the vagus nerve and spinal cord, can be used to collect and decode neural biomarkers to guide therapy.

Neuropace's RNS System, approved by the FDA for epilepsy in 2013,^26,27 is one example of a system that both records from and stimulates the brain. Population-level brain recording detects epileptiform activity as a biomarker of an oncoming seizure and the RNS System applies stimulation to control it. This system showed a median 75% reduction in seizure frequency after 9 years,²⁸ and early work showed more than double the seizure reduction with responsive stimulation compared with sham.²⁹

The closed-loop Evoke spinal cord stimulation system in development by Saluda Medical³⁰ also titrates stimulation parameters based on the evoked neural response in the spinal cord, and a long-term study showed significantly greater pain relief, at more clinically meaningful levels, than open-loop stimulation.³¹ Similarly, position-adaptive spinal cord stimulation has shown significantly reduced pain intensity and increased patient satisfaction over 2 years.³²

Since closed-loop devices record chronically from the nervous system, some existing devices are being applied for clinical insight and neuroscientific research. Neuropace's RNS System and Medtronic's Percept closed-loop DBS system³³ offer clinicians a platform to collect, display, and analyze recorded patient data.

The Future of Bioelectric Medicine: A Holistically Adaptive Approach

Today's closed-loop devices hint at the future of personalized bioelectric medicine, which will use collected data for increasingly precise application of neural stimulations to achieve therapeutic effects.³⁴ The combination of neural, organ, and full-body state information to control stimulation parameters will allow devices to achieve a targeted and consistent therapeutic response (Fig. 4).

FIG. 4.

The future of personalized bioelectric medicine where neural, organ, and full-body state information are processed in real-time and at longer timescales to update stimulation parameters for optimal treatment outcomes. This approach combines interpretation of neural biomarkers and other biomarkers of patient state, moving toward holistically personalized treatment of chronic illnesses by bioelectronic therapies.

Devices that decode collected data into interpretable biomarkers and patient state estimates—which in turn control applied stimulation—are in active development. Such devices must be capable of long-term recording, decoding, and control of neural, physiological, and state biomarkers. In pursuit of this, there are numerous challenges regarding device design, implanted and computational technologies, and scientific/medical interpretation of neural activity (Table 1). Different research and commercial organizations are approaching these challenges from separate angles, in pursuit of this next generation.

Table 1.

Features of Next-Generation Closed-Loop Devices for Control of Neural Activity

Categories	Challenges
Design objectives25,44	Device miniaturization
	Improved sensing capabilities at a variety of locations
	Precise neural targeting
	Physiological compatibility, adaptability, and feedback controlled
	Privacy and security of data collection and storage
	Evidence of long-term safety, reliability, and comfort.
Technology enablers25	Scale and materials: flexibility, stability, biocompatibility, robustness
	Electrodes and sensors: overcome limits to scale, number, and longevity
	Recording quality: improved robustness and processing speed
	Computation: improved machine learning and adaptive algorithms
	Power: reliable, wireless, sufficient
	Multiscale signal processing, modeling, and control
	Communications: secure, stable, reliable, wireless, integrated
	Regulatory requirements, including regulatory processes and clinical trial design
	Ethical: relating to data access, sharing or selling; invasive and noninvasive technologies
	Translation into medical practice: cost, availability of investment capital, and standardization.
Scientific/medical application45	Understanding of the neural biomarker signals through mining and decoding of neural data
	Understanding of neural circuitry's involvement in disease pathology to identify therapeutic interventions
	Systems level understanding of neural circuitry to make more complex targets accessible.

With a configurable approach to research into system design, the Dynamic Neuro-Modulator (DyNeuMo) is a flexible device for adaptive neural stimulation that was created as a collaboration among Bioinduction, Imperial College, and Oxford University. This system allows titration of parameters in response to multiple levels of patient information, combining population-level neural activity with higher-level state information, such as circadian patterns or accelerometer-derived patient state changes.³⁵

Other companies, such as Neuralink, have focused heavily on developing suitable implanted technologies. Neuralink, which aims to develop high bandwidth brain/machine interfaces that can be generalized to many medical or nonmedical applications, has developed a high-density electrode array and surgical robot to make high bandwidth, long-term implantation both scalable and precise.³⁶

Meanwhile, the challenge of neural biomarker decoding from high bandwidth neural recordings in the PNS is being tackled by companies such as BIOS Health, which is developing technology to treat chronic health conditions through neural interfaces using closed-loop approaches. Their approach uses machine learning techniques to take continuous multimonth neural recordings to discover and decode neural biomarkers that indicate bodily functions and organ dynamics, including mobilization³⁷ and respiration,³⁸ among others. These techniques allow the isolation of neural biomarkers from continuous neural recordings to create targeted therapies for diseases, which in the future will enable real-time response to that specific neural event as well as providing higher-level insight into patient status.

In parallel, efforts to make use of state biomarkers from wearable biosensors and inertial measurements are under way by BIOS and others,^35,36,39 allowing integration of closed-loop therapies with long-term changes in patient outcomes. This multifaceted approach, inferring higher-level outcomes as well as neural and organ-level information, provides complementary information to the neural biomarkers and ultimately progresses toward holistically personalized treatment of chronic illnesses by bioelectronic therapies (Fig. 4).

Conclusion

Advances in neural implants, wearables, and machine learning deliver exciting potential to radically shift the treatment of chronic conditions to an automated and holistically personalized bioelectric approach.

As this new paradigm for medical treatment develops, researchers and commercial bodies alike have taken notice. Major pharmaceutical companies are investing in bioelectric medicine: GSK and Verily Life Sciences established Galvani Bioelectronics in 2016 with an investment of up to £540 million over 7 years,⁴⁰ and Johnson & Johnson's portfolio includes several bioelectronic devices.¹ In addition, the US National Institutes of Health is funding the Stimulating Peripheral Activity to Relieve Conditions (SPARC) program to accelerate development of therapeutic devices that modulate electrical activity in nerves to improve organ function.⁴¹ The US Defense Advanced Research Projects Agency (DARPA) has an Electrical Prescriptions (ElectRx) program, which aims to deliver nonpharmacological treatments for pain, inflammation, post-traumatic stress, severe anxiety and trauma, using closed-loop, noninvasive modulation of the patient's PNS.⁴²

Achieving this vision requires major technological advancements, including innovations in neuroscientific understanding, accelerated computation for real-time machine learning, miniaturized electronics, and increased battery life (Table 1). Against this rapidly progressing backdrop, there is also a pressing need for regulations to ensure safety and consistency in adaptive closed-loop systems.⁴³

In the long term, unlocking the potential of the nervous system through the data contained in it may hold the key to new treatments for challenging chronic conditions, including hypertension, diabetes, rheumatoid arthritis, Parkinson's, and Alzheimer's. For a person with a severe chronic condition, neural therapies may provide personalized, adaptive treatments directly through the nervous system, so that the burden of pills and doctors' visits become a second resort rather than a daily reality.

Disclaimer

The article has been submitted solely to this journal and is not published, in press, or submitted elsewhere.

Author Disclosure Statement

S.K.L., G.S.J., and A.T. are employees of BIOS Health Ltd; E.H. is CEO and co-founder and O.E.A. is CSO and co-founder of BIOS Health Ltd.

Funding Information

The authors and this work were funded by BIOS Health Ltd.