CAREER: Exploring Mixed-Signal Computation for Energy-Efficient and Robust Brain-Machine Interfaces

Brain-Machine Interfaces systems have the potential to help and substantially improve lives of people that are affected by neurological disorders such as Epilepsy, Parkinson's Disease, Amyotrophic Lateral Sclerosis, as well as Brain and Spinal Cord Injuries. Across the United States, an estimated 100 million individuals grapple with such neurological disorders. This project aims to revolutionize Brain-Machine Interfaces making them more accessible and practical for everyday use. By developing new, energy-efficient technologies, the project promises to help individuals with motor disabilities or sensory impairments interact seamlessly with their environment. These advancements in Brain-Machine Interfaces will not only enhance mobility and independence for users but also reduce the costs and technical challenges associated with current systems. Beyond its scientific and technological contributions, the project is deeply committed to education. It offers unique opportunities for students to engage in pioneering research in rehabilitation engineering. Collaborations with educational and disability-focused organizations underscore its dedication to nurturing future scientists and engineers, thereby enriching the field and benefiting society at large. This project aims to significantly advance Brain-Machine Interface (BMI) technology, with a focus on creating bidirectional systems for individuals with motor disabilities or sensory impairments. The primary goal is to develop BMI solutions that are highly mobile, energy-efficient, and cost-effective, moving beyond the constraints of current laboratory-based systems. The approach includes the design of mixed-signal Spiking Neural Network (SNN)-based algorithms for efficient neural decoding, leveraging the strengths of both analog and digital circuit elements. This novel approach is anticipated to enhance the energy efficiency of BMI hardware substantially. Additionally, the project will explore the development of integrated circuits to miniaturize and reduce the cost of BMIs, making them more accessible and practical for everyday use. Another key focus is on algorithms and circuits that facilitate on-chip learning and extraction of stable latent variables, aiming to improve the robustness and adaptability of neural decoders under the inherent variability of intracortical BMIs. Ultimately, the integration of these technologies into a single System-on-Chip (SoC) and their validation using animal models will mark a significant step forward in the field. The project holds the potential for groundbreaking contributions to BMI technology, improving the quality of life for those with disabilities and advancing the state of rehabilitation engineering. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.