Embedded Tools for Processing Neural Ensembles

DESCRIPTION (provided by applicant): Multi-electrode recording from ensembles of neurons in awake, behaving animals is a central technique for understanding the neural basis of behavior and nervous systems disorders, and a prerequisite to the creation of neural-controlled prostheses. However, for these techniques to reach their full potential for processing neural data in real time, there are both computational and the hardware barriers to be overcome. The overall goal of this project is to provide software and hardware tools for overcoming these barriers and to encourage significant advances in multi-electrode recordings. Phase I will design a fast spike-sorting algorithm for real-time use, embed critical portions in reprogrammable hardware (FPGAs), and construct an overall plug-and-play architecture that integrates FPGA capabilities and includes a high-level FPGA compiler. Phase II will produce a beta version of Neural Arts' Chorus Recording System that incorporates software and hardware designed in Phase I. The software and hardware resulting from this project will give researchers the first real-time processing platform using reprogrammable FPGAs that is low-cost, lightweight, and can take full advantage of multi-wire electrodes and advanced spike sorting algorithms. The principal scientific benefit of this research will be to make possible significantly more powerful real-time systems for use with awake, behaving animals. The principal health-related benefit will be to facilitate increased understanding of neural functioning and disorders, and to provide a development platform for accelerated progress toward neural-controlled prostheses. Multi-electrode recording from ensembles of neurons in awake, behaving animals is a central technique for understanding the neural basis of behavior and nervous systems disorders, and a prerequisite to the creation of neural-controlled prostheses. However, for these techniques to reach their full potential for processing neural data in real time, there are both computational and the hardware barriers to be overcome. The overall goal of this project is to provide software and hardware tools for overcoming these barriers and to encourage significant advances in multi-electrode recordings. Phase I will design a fast spike-sorting algorithm for real-time use, embed critical portions in reprogrammable hardware (FPGAs), and construct an overall plug-and-play architecture that integrates FPGA capabilities and includes a high-level FPGA compiler. Phase II will produce a beta version of Neural Arts' Chorus Recording System that incorporates software and hardware designed in Phase I. The software and hardware resulting from this project will give researchers the first real-time processing platform using reprogrammable FPGAs that is low-cost, lightweight, and can take full advantage of multi-wire electrodes and advanced spike sorting algorithms. The principal scientific benefit of this research will be to make possible significantly more powerful real-time systems for use with awake, behaving animals. The principal health-related benefit will be to facilitate increased understanding of neural functioning and disorders, and to provide a development platform for accelerated progress toward neural-controlled prostheses.