An FPGA-based platform for accelerated offline spike sorting

摘要

There is a push in electrophysiology experiments to record simultaneously from many channels (upwards of 64) over long time periods (many hours). Given the relatively high sampling rates (10-40. kHz) and resolutions (12-24. bits per sample), these experiments accumulate exorbitantly large amounts of data (e.g. 100. GB per experiment), which can be very time-consuming to process. Here, we present an FPGA-based spike-sorting platform that can increase the speed of offline spike sorting by at least 25 times, effectively reducing the time required to sort data from long experiments from several hours to just a few minutes. We attempted to preserve the flexibility of software by implementing several different algorithms in the design, and by providing user control over parameters such as spike detection thresholds. The results of sorting a published benchmark dataset using this hardware tool are shown to be comparable to those using similar software tools.