Equipment: MRI: Track 2: Development of BrainScan: An Instrument for High-Bandwidth Real-Time Closed-Loop Neural Interfacing

About This Grant

This project develops BrainScan, a neural interface that enables direct real-time bidirectional communication between biological neurons in the brain and the digital world. The centerpiece of BrainScan is BrainCore, a bespoke chip designed for neural interfacing. BrainCore enables an energy efficiency that is orders of magnitude superior to the state of the art. Unlike current neural interfaces, BrainScan will be highly programmable to support many neuroscience studies centered on millisecond-scale closed loops involving analysis of large volumes of brain data. BrainScan’s form factor and power dissipation will make it amenable for portable use, eschewing the need for tethering. Current neural interfaces support only a small subset of these capabilities. BrainScan does not compromise on performance, timeliness, flexibility, or power. BrainScan’s development will encompass both the hardware and software. The bulk of hardware design would target the BrainCore chip, which uses multiple hardware accelerators for neural data processing, operating, unusually, in distinct clock domains, and the integration of peripheral off-the-shelf components for storage, wireless communication, and power. Software design includes an intuitive interface in a high-level language familiar to neuroscientists along with a compiler framework. Together, BrianScan will process about a hundred megabits per second of neural data below 1 watt of power. This ability will be validated in a pilot rodent study. With a transformative redesign of neural interfaces to read and electrically stimulate orders of magnitude more neurons than is currently feasible, BrainScan will enable cutting-edge neuroscience research. BrainScan will also mark a leap in the nation’s interest in building high-precision and high-bandwidth neural interfaces that may one day help augment human cognition and decision-making. Furthermore, by designing BrainScan to be flexible, BrainScan will be the foundation for a standardized computational platform for neural interface research and industry. This will help neuroscientists and clinicians transcend the limitations of current neural interfaces that support only narrow classes of experiments, and the fragmented software ecosystem and lack of standardized and portable applications that they have lead to. BrainScan will be made available to multiple labs nationwide through an inter-institutional loan program. The PIs will devise written documentation and online video tutorials, hosted on a dedicated website, to enable use of BrainScan in a turn-key manner. BrainScan’s board design files, software, and high-level simulators for the BrainCore chip will be shared with the community to build on. This project will also train PhD students and postdoctoral scholars in neurotechnology and computer science and 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.