The Computing Community Consortium (CCC) is pleased to announce an upcoming workshop focused on exploring the role of computing research in neural interfacing. This workshop, co-organized by Abhishek Bhattacharjee (Yale University), Raghavendra Pradyumna Pothukuchi (currently an Associate Research Scientist at Yale and incoming William R. Kennan Jr. Fellow Assistant Professor of CS at UNC Chapel Hill), and Nishal Shah (Rice University), with support from CCC Council member Weisong Shi (Alumni Distinguished Professor of CIS at University of Delaware) and Jojo Platt (Platt and Associates), will be held on April 22-23 in Washington, D.C.
Understanding Neural Interfaces
Neural interfaces are systems that create a direct communication link between the nervous system (typically the brain or spinal cord) and computers or machines. By reading and modulating brain activity, neural interfaces enable the exchange of information between biological neural circuits and technological systems.
Through the ability to control, monitor, or manipulate neural activity, these devices hold the potential to understand and treat neurological disorders like epilepsy, paralysis, and blindness, and may one day augment healthy brain functions. Futurists project that neural interfaces will also eventually enable fusing the best of artificial and natural intelligence.
The Evolving Landscape of Neural Interfaces
Advancements in sensing technologies and neural signal processing methods have driven rapid progress in neural interfaces. Breakthroughs have led to exponentially growing interfacing data rates, now in the hundreds of megabits per second (Mbps), fostering compelling applications and sparking a burgeoning startup and industry ecosystem to bring this technology to users. Given the growing data volumes and computational demands of neural interfacing applications, computing on the neural interface has become vital. However, supporting the large, complex and changing landscape of algorithms at these massive data rates while being safe for in- or near-brain operation is hard, and requires a re-think of the entire computing stack. In addition to this technological challenge, the regulatory processes, funding models, and industry-academia partnerships to develop and safely translate any novel computing neural interfaces into positive impact for users are evolving.
A Roadmap for Computing Research Neural Interfaces
The workshop aims to envision a clear roadmap of computing research for neural interfaces, bringing together computing researchers across the stack, neuroscientists, clinicians, key industry players, and funding agencies. A key goal is to identify the research directions across the computing disciplines on developing systems and design methodologies that rescue the field of neural interfaces from the “hardware/software lottery.”
The hardware/software lottery refers to research success being dictated by compatibility with available hardware and software rather than the intrinsic merit of the idea. Lotteries become more prominent when computing needs to embrace specialization to achieve acute levels of efficiency. Because neural interfaces must achieve a unique combination of performance and low power, they risk over-specialization resulting in a lottery for users.
Additionally, the workshop would identify the new ecosystems that must be created to spur, sustain, and translate the full impact of computing research in this space. This would help channel the collective efforts to build next-generation neural interfaces.
Balancing Flexibility and Efficiency
Establishing the right balance between flexibility and efficiency is one of the key aims of this workshop.
“This will be achieved by bringing together researchers across the many diverse fields that are involved in developing neural interface technology and establishing a community,” says Abhishek Bhattacharjee, workshop co-organizer and A. Bartlett Giamatti Professor of Computer Science at Yale University.”
“So rarely do we have opportunities to discuss the future of this technology with experts from every step of the development process, in software, hardware, neuroscience, and neurosurgery, and to hear what are their priorities and their pain points. We hope that through this workshop, participants will make connections that cascade into collaborations and partnerships down the road, and that these connections will enable a virtuous cycle of research between computer system and neural interface design, accelerating the pace of innovation in each field.”
Establishing Standardized Benchmarks
Part of this endeavor is to arrive at standardized a set of benchmarks for hardware development.
“Clear benchmarks for neural interface hardware would benefit scientists across the entire neural interfaces research community,” says Nishal Shah, workshop co-organizer and Assistant Professor of Electrical and Computer Engineering and McNair Scholar in Neuroengineering at Rice University.
“For neuroscientists, a community-wide standard for neural decoding pipelines and data sets would enable consistent comparisons of different decoding methods and create a universal platform for research and clinical use. For computer systems developers, it would provide a clear and stable computational target, helping to design better hardware. Standardized hardware would also boost software development by allowing developers to focus on coding without needing to worry about hardware specifics, ensuring compatibility across different platforms.”
Defining Research Priorities for Neural Interfaces
In addition to establishing a cohesive community, the organizers hope to produce a community-wide document that outlines possible research directions to pursue based on discussions at the workshop.
“With so many research teams seeking to innovate on every step of neural interface design, from developing new algorithms to utilizing novel materials in the implants, to even discovering less invasive implantation techniques, it is difficult to keep up with all of the innovation that is happening, and even harder to compare metrics to measure how impactful new discoveries are,” says Raghavendra Pradyumna Pothukuchi, workshop co-organizer and incoming Assistant Professor and William R. Kenan Jr. Fellow at University of North Carolina, Chapel Hill
“During the workshop, we intend to discuss how best to enable the design and comparison of various neural decoding pipelines and hardware platforms, and we hope the final report will include a clear roadmap for identifying future requirements for neural interface capabilities.”
Expanding Engagement Beyond the Workshop
Following the workshop in April, the organizers will also host virtual discussions to allow more members from the computing research community and the wider neural interfaces community to share their thoughts as well. If you are interested in participating in one of these virtual roundtables, please indicate your interest here, and stay tuned for more details.
