A 10-year journey towards clinical translation of an implantable endovascular BCI a keynote lecture given at the BCI society meeting in Brussels

The development of the Stentrode brain–computer interface (BCI) began in 2013 at the University of Melbourne, with a small team of scientists and engineers aiming to record brain activity without the need for craniotomy. This project was inspired by advancements in interventional cardiology, leading to the idea of using stent-mounted electrodes to capture brain signals. The resulting system, known as endovascular electrocorticography (ECoG), builds on foundational work in the BCI field, including the early exploration of intracranial intravascular electroencephalography (EEG) recordings by Driller and colleagues in the 1970s [1]. Their seminal paper demonstrated the ability to record spontaneous and evoked electrical activity from regions of the brain not accessible by traditional EEG, using electrodes in intracranial vessels in both baboons and humans. This work laid the groundwork for modern, minimally invasive, endovascular approaches to brain signal recording.

An ECoG-based BCI captures and decodes cortical motor signals through intracranial electrodes to bypass impaired or non-functional limbs and directly control external devices, such as computers or robotic limbs. While this technology does not directly restore motor function, it provides individuals with motor impairments the opportunity to interact with their environment through technology. By offering a new means of communication and control, the system enables patients to regain autonomy and perform tasks that were previously unattainable due to their physical limitations.

The Stentrode technology is simple, and the goal is to keep it as simple as possible. The device resides in a blood vessel, with sensors mounted on a stent positioned against the vessel wall (figure 1). It is fully implanted, making it invisible, simple, accessible, and crucially, wirelessly connected to the digital ecosystem that we all take for granted. Over the past 10–15 years, this digital ecosystem has become an integral part of our daily lives. Prior to the 2010s, it was not so clear how BCIs would impact patients' lives, but now, with incredible supercomputers in our hands, it is obvious that restoring patients' access to this technology will be transformative.

Zoom In Zoom Out Reset image size

ECoG-based BCI technology has made significant strides in offering more precise and stable brain signal recordings compared to non-invasive methods like EEG, making it a promising tool for controlling external devices. Its ability to capture high-resolution cortical signals with minimal noise has led to advancements in motor control and communication for individuals with severe impairments. However, there are notable challenges. The invasive nature of ECoG requires surgical implantation, posing risks like infection and complications from the procedure. Additionally, long-term use raises concerns about electrode stability and potential degradation over time. Another limitation is the complexity of decoding motor intentions from brain signals in real time, which can limit system responsiveness and accuracy. Despite these hurdles, ECoG-based BCIs have demonstrated significant potential in pilot studies, offering a path toward more reliable and effective neuroprosthetic applications.

Now, transitioning from a scientific inquiry phase to a product development phase, our primary goal is to scale our concept to ensure that it is both safe and effective for widespread use. To date, ten patients with severely impaired mobility have been implanted with the Stentrode system.

As we move toward scaling our concept, the existing infrastructure for neurointervention provides a robust foundation for advancing the Stentrode system beyond feasibility. This well-established network not only supports the technical scalability of BCIs but also ensures the potential for widespread adoption in clinical practice.

The reason endovascular BCI is so scalable is the infrastructure for neurointervention that has been created over the last decade. The field of clinical science for neuroprosthetics is poised for significant growth. A new field of neuroprosthetic physicians, likely to be neurologists or rehabilitation physicians by training, will become integral to this field. The infrastructure for neurointerventional procedures is extensive, with over 2500 neurointerventionists and 1200 catheter labs in the US, performing approximately 200 000 neurovascular procedures annually. This infrastructure supports the adoption of new neurovascular technologies, with the ability to deliver them to a large patient population. With this backdrop, plans are underway to establish 30 clinical sites in the US for a pivotal trial of the Stentrode. The regulatory landscape in the EU is still being navigated, so initial efforts will focus on the US.

One of the significant challenges in BCI development is not only addressing patient needs but also ensuring that patient selection is appropriately aligned with the technology's capabilities. In our studies, patient selection criteria have been carefully considered to ensure that individuals with preserved motor cortex activity and suitable venous anatomy are chosen, as these factors are critical for successful Stentrode implantation and use. Patients diagnosed with conditions such as amyotrophic lateral sclerosis (ALS), stroke, and spinal cord injury, who are unable to perform motor tasks but retain the ability to generate cortical motor signals, form an appropriate cohort for this technology. Careful selection ensures that the system can effectively capture movement intentions and translate them into actionable commands [2, 3].

Once implanted, the Stentrode's ability to accurately capture movement intent across subjects has been a key success factor [2, 3]. The recorded neural signals allowed patients to control digital devices after appropriate training, and signals remained stable throughout the study period. This ability to maintain signal fidelity, even over extended periods, underscores the robustness of the system and its potential for long-term use.

Safety has been another crucial consideration in the development of the Stentrode system. Concerns about material degradation, device migration, and vascular complications have been thoroughly addressed in clinical trials. The Stentrode's nitinol scaffold, with its platinum electrodes, has shown long-term stability, and importantly, no degradation of the device or signal fidelity was observed during the 12 month trial period [2]. Follow-up imaging confirmed that the device remained securely in place within the blood vessel, with no migration or thrombosis occurring in any patients [2, 3].

In terms of overall safety, no serious adverse events related to the device were reported, and minor side effects such as headaches or bruising resolved without intervention [2]. The favorable safety profile of the Stentrode, combined with its stable performance, provides a strong foundation for broader clinical application.

The primary focus of our current work is to address the profound unmet need for autonomy among patients with conditions such as ALS. Many current systems fail to provide continuous functionality and user autonomy, leading to significant distress and dependence among patients. Our technology aims to restore this lost autonomy by enabling continuous and effortless interaction with digital devices, which is critical for daily activities and communication.

Many patients with severely impaired mobility have shared their experiences with the limitations of eye trackers, but a particularly poignant quote comes from the son-in-law of an ALS patient we worked with: 'He'd lie there for three hours just staring. He cannot get help. He cannot do anything.' This story is heard repeatedly, where if troubleshooting is required and no one is in the room, the patient is unable to resolve issues impeding their use of eye tracking systems. Many systems have some level of troubleshooting built in, but when the patient encounters a problem, they often must sit and wait. Families struggle to sleep at night, worried their loved one cannot ask for help if needed. If the eye-tracking equipment is not aligned correctly or if the button is not in the right spot for the system to work, it causes significant distress. Some people even resort to setting up baby monitors, hoping to hear any sound indicating the need for assistance.

A 2016 paper by Vansteensel [4] was a seminal work in this field addressing this patient need, and an inspiration for our team. While the data rate from the intracortical arrays was relatively low, the critical takeaway was that achieving a fully implanted system is incredibly meaningful for patients because it allowed the patient to be mobile and use the system continuously. Importantly, the home system required minimal assistance, and the initial setup did not require any special skills for the caregiver. This is crucial when developing a system for widespread patient use and is a significant consideration for caregivers who lack the tech-savviness for managing a complicated system with frequent calibration steps.

The starting point for the system was a small number of electrodes utilizing simple, stable features, enabling a patient to control a spelling device where she would otherwise rely on an eye tracker [4]. This capability allowed her to use the device outdoors where eye tracking was not effective. Remarkably, seven years later, the patient continues to use the system. This story represents a wonderful achievement by the Utrecht group and serves as a major inspiration. It demonstrates how the simplicity of a low-bandwidth system can profoundly impact patients' lives.

The issue of continuous autonomy is a fundamental problem that BCI technology is uniquely positioned to address, and one that has not yet been tackled by other technologies. Decoding neuronal signals from brain implants has led to significant milestones, such as enabling patients with quadriplegia to control robotic or paralyzed limbs and operating computer typing programs [2, 4–10]. However, despite these breakthroughs, achieving daily, in-home use remains a challenge. BCIs offer an opportunity to restore patient autonomy by reconnecting them to the digital ecosystem, which plays a central role in self-determination and decision-making in modern life.

For BCIs to enable reliable, continuous use in a home environment, they must meet specific criteria: supporting user autonomy, ensuring continuous functionality, and reliably detecting only intentional actions [4]. The interface must remain permanently available, with sensors that stay connected to the cerebral cortex without causing discomfort or any negative aesthetic impact [11]. These design goals form the foundation for developing BCIs into empowering and life-changing tools.

One of the significant challenges in BCI development is establishing performance metrics. The FDA's guidance on neuroprosthetics [12] emphasizes the need for endpoints that reflect restored physiological functions. By focusing on motor intent transmission, and output, our approach aims to create a neuroprosthetic that effectively bridges the gap between the brain's intentions and the control of external devices, thereby restoring critical motor control function.

Importantly, since few patients will be able to pay for these devices out-of-pocket, it is essential to consider the perspectives of payors regarding how BCIs fulfill their requirements.

In the US, a major payor for healthcare is the Centers for Medicare & Medicaid Services (CMS), a federal agency that administers the nation's major public health insurance programs, including Medicare. Medicare is a national health insurance program primarily serving individuals over the age of 65 and some younger people with disabilities or specific medical conditions. Unlike private insurance, Medicare is government-funded and plays a critical role in determining which medical devices and treatments are eligible for coverage.

For a medical device like a BCI to be widely accessible in the US, it must meet CMS criteria to qualify for Medicare reimbursement, demonstrating that it is 'reasonable and necessary' with tangible health benefits. Without CMS approval, the cost may not be covered by insurance, limiting access to patients unable to pay out-of-pocket. Therefore, clinical trials must provide evidence not only for safety and effectiveness, as required by the FDA, but also meet CMS standards to ensure coverage and widespread access. It is critical to design trials that measure the primary effectiveness endpoint of a BCI to offer clinically meaningful data on its performance and benefits.

As BCI technologies advance, there is an increasing need to define clinically meaningful endpoints that go beyond engineering metrics and directly reflect patient outcomes. A range of engineering outputs can be used to measure the elements of system performance, each telling a different story about the system in a meaningful way. Thompson et al presented a comprehensive paper at a BCI Society workshop 10 years ago that dissected BCI functionality into discrete and continuous categories, providing a useful framework for understanding decoding [13]. However, these features do not necessarily reflect clinically meaningful benefits. Although they may be insightful, they remain engineering outcomes. As the field moves into the clinical science domain, there is a pressing need to create a new framework for endpoints that are inherently clinically meaningful.

Thompson et al also offered a user-centric approach to evaluating BCI systems, incorporating four key factors: Usability, Affect, Ergonomics, and Quality of Life [13]. Usability focuses on how efficiently users can complete tasks, considering learnability and satisfaction. Affect addresses the emotional experience during extended use. Ergonomics included cognitive load reduction and ease of system control. Quality of Life involves the overall system impact on user well-being. These factors emphasize the need for a holistic view of BCI performance, including the importance of user experience in system design and evaluation [14]. Metrics such as ease of use, setup time, ability to function outside of controlled environments, and user experience (both for the patient and caregivers) are critical in determining the overall usability of these systems [13–15]. The ability for devices to operate in a plug-and-play manner, without requiring the continuous support of clinical professionals or engineers, is essential for successful long-term use.

While engineering-based performance measurements, such as decoding accuracy and bitrate, are essential for understanding the technical capabilities of a BCI system, they do not necessarily directly translate into tangible benefits for patients and caregivers. These metrics provide insight into the BCI's functioning, but to qualify for reimbursement under Medicare and CMS guidelines, it is vital to demonstrate that the device offers 'reasonable and necessary' advantages, particularly in terms of improving patient autonomy and enhancing quality of life. Additionally, caregiver needs—such as reducing the burden of assistance—must be addressed to ensure practical, sustained use. Therefore, it is critical to incorporate endpoints that reflect patient-centered outcomes, such as the ability to regain independence in daily activities. By aligning clinical trial designs with these regulatory expectations, for successful FDA approval and broader accessibility through Medicare reimbursement, thus increasing the likelihood of widespread clinical adoption.

The FDA guidance document on BCIs [12] will play a critical role in shaping considerations for clinical guidelines for BCI neuroprosthetics. It defines implanted BCI devices as 'neuroprostheses that interface with the central or peripheral nervous system to restore lost motor and/or sensory capabilities in patients with paralysis or amputation.' While the BCI definition is broader, this focus on implanted and sensory motor neuroprostheses is significant. Considerable attention has been paid to what constitutes a neuroprosthesis, considering its important clinical implications.

There have been extensive discussions about the definition of a BCI within the BCI Society, with two crucial definitions central to this discussion. The BCI Society needs to take a position, but currently, Wolpaw's 2002 paper remains a key [16]. A BCI is defined as 'a system that records and decodes brain activity to produce an output to directly control a computer or other external device.' This differs from a neuroprosthesis, a medical device that restores function, with the term 'restore' being vital. A neuroprosthesis is defined as 'a medical device that can restore the function of a motor, sensory, or cognitive modality that is damaged due to injury or disease.' It encompasses motor, sensory, and cognitive functions, anticipating the development of both motor and speech neuroprostheses. These devices fall under the same category as those that restore brain function. The Cochlear hearing device exemplifies a neuroprosthesis. Whether stimulating or sensing, the restoration of brain function qualifies it as a neuroprosthesis. Krucoff et al [17] noted, 'In addition to BCI function—which connects the brain to a computer—a neuroprosthesis replaces a missing biological functionality'. This distinction is critical for clinical medicine, as demonstrating the restoration of physiology is essential. Therefore, developing endpoints that measure physiology directly will enable the quantification of restored performance and meet the criteria for clinical success.

The brain functions being discussed pertain to the motor system and are topics seldom addressed because they are invisible and generally untreatable. Neuroprosthetics represent a new treatment paradigm for the motor system. The FDA's BCI Guidance Document uses the terms 'paralysis' and 'amputation,' but the real issue is motor impairment, which the guidance document refers to as motor capabilities.

Brain function includes volition, free will, and the intention to move, collectively known as motor intent. The cortex activates an area due to the intention to move a muscle in the body. This signal is transmitted from the cortex down through the brainstem, spinal cord, and peripheral nerves to a muscle, facilitated by motor neurons. ALS, for example, is a motor neuron disease. Motor output is the process by which the signal reaches the muscle, resulting in movement. Using fingers to control digital devices exemplifies the function of motor output.

Many conditions cause motor impairment, and not just paralysis or amputation. Any condition that disrupts the signal transmission from the motor cortex to the muscle can impair motor function. These conditions include stroke, spinal cord injury, muscular dystrophy, arthritis of the joints, peripheral nerve diseases such as Guillain–Barré syndrome or chronic inflammatory demyelinating polyneuropathy, multifocal motor neuropathy, and ALS. All of these conditions result in impaired motor function. A motor neuroprosthesis aims to restore the transmission of motor intent. This technology will not assist individuals whose cortex is non-functional; the brain must be capable of generating intention. With preserved volition, this technology can bypass impaired pathways and create a direct interface to restore the ability to engage with the digital ecosystem.

The Stentrode's array consists of 16 electrodes that capture neural activity adjacent to the motor cortex and transmits the ECoG signals through 16 channels [2, 3]. The device records neural features associated with motor intent (e.g. figure 2). A custom decoding algorithm is used to decode movement attempts and convert these into digital motor outputs (DMOs) for device control. The signal output consists of commands such as short or long clicks, which enable users to perform tasks like typing or navigating digital environments [2, 3].

Zoom In Zoom Out Reset image size

Sawyer et al published a comprehensive paper conceptualizing DMOs as a framework for restoring muscle function [18]. I propose that the concept of DMO can serve as an endpoint for defining the function of an implantable BCI for a motor neuroprosthesis. The DMO encompasses various metrics, including bitrate, information transfer rate, and accuracy, forming a framework for evaluating the efficacy of the device. DMOs must achieve specific elements to facilitate computer actions. For instance, a signal processed through a decoder can produce a discrete or continuous output, leading to a computer action (figure 2). Building a library of these DMOs, categorized by different elements, provides a language for discussing their efficacy, both in terms of both internal processing and external application. This framework will be instrumental in describing what is beneficial to patients.

Conceptualization of this framework will be essential for clinical trials. Neurologists and clinical physicians who will care for patients with BCIs will need a clinical science and a language to understand and explain why a system is working or not. A DMO should be objectively measurable, functionally generalizable, and capable of demonstrating restored motor intention. We now have motor intent, in our case a device that can transmit the motor signal, we have a system that creates a network which can be updated on the backend and can also talk to any number of generalizable devices, and we have computer actions that can be controlled by those DMOs in a range of different ways.

The future trajectory of this technology is exciting and profound. Focusing on the neuroprosthetic performance of a DMO to control systems, even a simple low DMO input, when combined with language models or predictive technologies such as Chat GPT, can have a significant impact. These technologies can learn a patient's needs over time and transform discrete inputs into powerful outputs.

We are building a system that we hope is generalizable across a range of technologies, using commonly available standards to create a completely wireless and invisible system for the patient. The CMS will likely require proof that these applications have a clinically meaningful impact and health benefits for patients. However, it is important to separate secondary applications from the initial use case of the system.

Consider a stroke as an example. A breakthrough occurred over the past decade with the opening of blocked blood vessels in the brain. Previously, medications were used to 'clot bust,' but they were not very effective. A new device was developed that opens the blocked blood vessel and restores brain function. Initially, the goal was simply to prove that the blood vessels could be opened. Over time, it became crucial to show that opening a vessel after one or two days was ineffective, as the brain was already damaged. However, the functional outcomes could not be measured until the device's primary function of opening the vessel was demonstrated. It is important to focus on the ability of a prosthesis to restore motor system function before establishing outcomes that demonstrate functional benefits.

Additionally, there are no validated tools to measure what the impact of losing the ability to use a smartphone or the consequent impact of getting that control back. Fry et al recently proposed that digital instrumental activities of daily living (IADL) might be useful [19]. A digital IADL scale would be a way to validate and assess the resulting impact of restoring the ability to engage with the digital ecosystem. One challenge here is the rapid pace of technological innovation, which makes it difficult to create a validated scale for a system with a 12 month lifecycle (smartphone). The speed of technological advancement is astonishing, suggesting the need for a dynamic system like a Diagnostic and Statistical Manual of Mental Disorders that updates continuously.

The endovascular approach to BCIs offers several features that make it an appealing option for both clinicians and patients. By utilizing the body's natural blood vessels to access the brain, this technique eliminates the need for open-brain surgery, significantly reducing the risks associated with more invasive procedures. In addition to being minimally invasive, the approach is aesthetically favorable, enabling the system to be fully implanted and wirelessly connected. Patients benefit from short recovery times and the enhanced long-term stability of the device, which becomes embedded within the blood vessel walls. As discussed above, the extensive infrastructure for neurointervention already in place ensures that this technology is scalable, paving the way for widespread clinical adoption.

Implementing the endovascular approach involves accessing the veins in the brain, starting from the jugular vein. The catheter containing the Stentrode is threaded from the back end of the vein which reduces the risk of stroke because it is where blood drains from the brain. This pathway provides access to intricate networks of blood vessels in various brain regions. The primary focus is on the largest blood vessel running down the middle of the brain, the superior sagittal sinus. A stent, which posed significant manufacturing challenges owing to its construction from two types of metals with insulation between them, is fully implanted. This wireless, ECoG-based BCI resides within the blood vessels. The sensors are mounted on a self-expanding stent positioned within the blood vessel wall, situated between the areas of the M1 motor cortex, with a cable extending out of the brain through a naturally existing hole at the base of the skull.

The initial funding for the endovascular BCI came from DARPA's Reliable Neural Interfaces Technology Program, led by Jack Judy and subsequently by Doug Weber. During this time, the various grantees were working on intracortical arrays, EEG, and ECoG, while our group in Australia was on the fringes not only by being in Australia but with our focus on blood vessel-based approaches. We used a 1992 General Electric angiography machine that resembles an oscilloscope. The quality was poor, but it cost only $5000 and we made it work. We spent many hours fashioning the devices by hand followed by even more hours in the angiography room, refining the technique for navigating into the vein (figure 3).

Zoom In Zoom Out Reset image size

We did not publish our results for more than four years because we wanted to present a story that was holistic and meaningful. The goal was to achieve long-term cortical recordings from within a blood vessel and compare them to subdural and epidural ECoG arrays. Our study showed rough equivalence and that we could separate signal from noise at 220 Hz, which later improved to approximately 500 Hz in human systems.

We then wanted to understand how the device integrates into the sheep's blood vessel wall, which is crucial because if bare metal is exposed in a blood vessel, it attracts clots, posing a significant risk. To ensure that the device is fully encapsulated inside the vessel wall, a process called endothelialization is a critical safety measure. We drew upon the extensive literature on heart stents, where this problem has been addressed for decades, and found that within a few months, the device becomes fully encapsulated, allowing for a reduction in blood-thinning medications and lowering the risk of clot formation.

A key unknown was the impact of endothelialization on signal stability, a question raised by intracortical evidence showing signal deterioration over time. In contrast, the Stentrode, using a 500 micron disk electrode to record ECoG, showed improved signal stability and improved recordings of somatosensory potentials once the device was incorporated into the vessel wall. This improved conduction is due to the enhanced stability of electrical potentials moving through the tissue of the vein wall, compared to the shunting of the signal through direct contact with blood. Once implanted, the device cannot be removed, but experiences with cardiac stents and brain stents for treating idiopathic cranial hypotension suggest that the body can tolerate such devices within a blood vessel for a lifetime.

The first complete human safety data for the Stentrode was published in 2023 [2]. While the technology has since evolved, initial implementations involved a device in the superior sagittal sinus, recording from both sides and connecting to an implantable receiver under the skin. Early iterations required a portion of the device to sit over the skin such that while that component was fully implanted, it resembled a wearable device.

In functional neurosurgery, the field of DBS began by precisely placing probes. Understanding what is on the other side of the blood vessels is crucial, necessitating new imaging techniques to accurately localize within the blood vessel's domain. Developing markers to define the M1 boundaries from the blood vessel perspective is complex but essential, with ongoing evaluations of these techniques. Yoo et al's work utilizing MRI and fMRI to identify the motor cortex in relation to nearby blood vessels [20–22], formed the basis for current techniques used to localize M1 angiographically.

During the implantation procedure, the neurointerventionist uses an angiogram to snake a catheter through the blood vessels of the brain (figure 3). When the catheter reaches the target region near the motor cortex, the Stentrode, which is prepackaged in the catheter, is left lodged in the blood vessel where it opens like a flower when the catheter is retracted. The first generations of Stentrode devices were challenging to open, with the surgeon relying heavily on tactile feedback. Although technical hurdles are being effectively managed, system improvements are ongoing. From a surgical perspective, the procedure is straightforward compared with stroke interventions, which require navigating multiple bends in the veins. Here, the device is delivered in a straight shot.

After implantation, the device is connected to an implantable unit under the skin of the chest and a signal check ensures system functionality before closing the site. Patients are typically discharged within 48 h and placed on dual antiplatelet therapy for three months, followed by aspirin for 12 months. Brain stents, used in transverse sinus stenting, have shown that patients can eventually discontinue antiplatelet therapy, although aspirin may be continued indefinitely for arterial stents.

It has been nearly 20 years since Kennedy first reported an implantable BCI device in a human [23]. By 2008, Braingate was flourishing, and some of the most remarkable and ambitious demonstrations of what can be achieved with intracortical devices have been witnessed since then. However, this technology has yet to advance through the critical phase of clinical translation, which requires a pivotal trial. Although numerous feasibility trials have shown proof of concept, the next steps will involve replicating these results in a large population, obtaining marketing approval from the FDA or other regulatory authorities, and making the technology available and reimbursable for a broad patient base.

Synchron will soon be targeting an FDA submission for a pivotal trial. During the presubmission process, safety has been extensively discussed with the FDA. In the first-in-human study reported in 2023 [2], the primary endpoint was stroke or permanent morbidity related to the device, which fortunately has not occurred. Secondary endpoints included device migration and blood clots leading to vessel occlusion, none of which were observed. Some patients experienced non-serious adverse events such as headache and bruising, which resolved over time [2]. These results are encouraging, although much larger sample sizes are needed for a full assessment of safety.

One questions we plan to address is how well we are positioning these electrodes over the motor cortex. Some electrodes are positioned over the sensory cortex or supplementary motor area, but all of them produce signals. Over time, the importance of precise electrode location will become clearer, which will necessitate that surgeons be trained in methods for achieving high positional accuracy.

Identifying motor-related features has revealed several fundamental insights. Encoding and training pose challenges for clinical translation because extensive training periods are impractical for patients. Instead, identifying stable neural features linked to simple, familiar actions, such as tapping a foot, results in high accuracy DMOs that can be immediately effective. We have observed high-frequency events, particularly in the 300–400 Hz range, that correlated well with movement onset and were not noise. These findings are being further explored to enhance decoding accuracy.

The real-world impact of our technology is best illustrated through the stories of our patients. For instance, one of our early patients, who was almost completely locked in due to late-stage ALS, demonstrated the profound impact of our system by his smile when he completed his first successful computer action with the system. Using a low-discrete function DMO on an iPhone, he was able to report pain issues with an app, a task previously impossible for him. While the integration of continuous decoding is progressing, the current focus is on discrete decoding due to its simplicity and stability.

With the remarkable trajectory of advances in the last decade, BCI technology is poised to change the world, and the world is watching, with the FDA and CMS closely monitoring and guiding developments. This is an opportune time for the BCI Society to step up and take a global leadership role in guiding the release and implementation of this transformative technology. I envision the BCI Society assuming a guardian role and providing critical scientific, ethical, and patient-centered guidance on all aspects of its deployment. I encourage the BCI Society to embrace and adopt strong, considered positions to help usher in a remarkable era, ensuring that the wellbeing of patients remains at the forefront of these efforts.

The author(s) have confirmed that any identifiable participants in this study have given their consent for publication. The author discloses stock ownership in Synchron and extends gratitude to Natalie D DeWitt (Accendo Scientific) for her assistance with manuscript preparation and to Adam Fry (Synchron) for his valuable input and review.

The data cannot be made publicly available upon publication because they contain commercially sensitive information. The data that support the findings of this study are available upon reasonable request from the authors.