Towards Chip-on-Chip Neuroscience: Fast Mining of Frequent Episodes Using Graphics Processors

TL;DR

This paper introduces a GPU-based method for real-time mining of neural spike train data, enabling efficient detection of coordinated neuronal firing patterns from multi-electrode arrays.

Contribution

It presents a novel GPU algorithm with a new computation mapping scheme and elimination approach for fast, real-time analysis of neural event streams.

Findings

Achieves real-time processing of large neural datasets

Successfully detects coordinated firing patterns in synthetic and real data

Demonstrates scalability and efficiency of the GPU solution

Abstract

Computational neuroscience is being revolutionized with the advent of multi-electrode arrays that provide real-time, dynamic, perspectives into brain function. Mining event streams from these chips is critical to understanding the firing patterns of neurons and to gaining insight into the underlying cellular activity. We present a GPGPU solution to mining spike trains. We focus on mining frequent episodes which captures coordinated events across time even in the presence of intervening background/"junk" events. Our algorithmic contributions are two-fold: MapConcatenate, a new computation-to-core mapping scheme, and a two-pass elimination approach to quickly find supported episodes from a large number of candidates. Together, they help realize a real-time "chip-on-chip" solution to neuroscience data mining, where one chip (the multi-electrode array) supplies the spike train data and another (the GPGPU) mines it at a scale unachievable previously. Evaluation on both synthetic and real datasets demonstrate the potential of our approach.