What is it about?
The shape of the Neural-spike (N-spike) signal contains vital information that provides insights to tackle problems related to brain disorders such as neural epilepsy. A key challenge is to identify the source neuron of the Neural signal which is known as spike sorting. Further to identify the source in real-time requires a small, low-power implantable system, which is a challenging problem.
Featured Image
Photo by Josh Riemer on Unsplash
Why is it important?
This work presents a system-level overview of a low power on-chip spike sorter. The proposed scheme of variation aware, adaptively trained, low-power analog spike-sorting system can be an attractive option for complex, multichannel brain-machine interfaces for emerging applications.
Perspectives
This article provides insights on approaching a biomedical application known as spike sorting in real-time. Here, we have explored the feasibility of using an analog resistive crossbar network (RCN) Spiking Neural Network (SNN) classifier. The sub-systems such as the analog front-end (AFE), V to I, CCO, and SNN(RCN) form the processing units of a channel. As the future trend is increasing the number of electrodes/channels, they must consume minimal power. Also, we propose an adaptive 2-step shared training unit which is shared by multiple channels, amortizing the area cost. We hope this article gives an overall perspective to raise many more insights and scope towards similar applications.
Anand Mukhopadhyay
Indian Institute of Technology Kharagpur
Read the Original
This page is a summary of: Power-efficient Spike Sorting Scheme Using Analog Spiking Neural Network Classifier, ACM Journal on Emerging Technologies in Computing Systems, April 2021, ACM (Association for Computing Machinery),
DOI: 10.1145/3432814.
You can read the full text:
Resources
Contributors
The following have contributed to this page
