Title - "Full-stack support for high-performance neural experimentation"
Abstract - Modern neural experimentation suffers from severe limitations on spatial resolution, temporal resolution, and ease of use. These limitations restrict experimenters ability to design experiments, forcing them laboriously construct experiments which can tolerate larger spatial and temporal granularities such as rate-based approaches. Some of these limitations are hard to scale against. Spatial resolution in particular is limited by extracellular electrode technology, which currently covers a very small percentage of all neurons in a culture and aliases the activity from a hundred odd neurons at each electrode. Temporal resolution and ease of use are theoretically much easier to scale against; modern computers clock 5 orders of magnitude faster than neural recordings, offering plenty of margin with which to implement computing stacksī. Despite this potential, the temporal resolution and ease-of-use of modern neural experimentation stacks are dismal. Several orders of magnitude's worth of temporal resolution is left on the table while simultaneously providing no intuitive interface for designing tightly-coupled (i.e. high-res) experiments. The MARK L system is designed to reduce the limitations imposed on temporal resolution and ease-of-use by providing a simple and intuitive yet highly-performant stack. It is capable of precise millisecond-level feedback, allows all experiments to be written at the spike level in C, and can be implemented on hardware costing as little as $3000.
Bio - Jakob Arend is a third-year Computer Science Ph.D. student with an undergrad in Engineering Physics. He is advised by Josep Torrellas and has been working on the Mind in Vitro project for three years. His research interests include hardware and software support for neural experimentation, as well as theory and experimentation for neural computation.
Part of the Illinois Computer Science Speakers Series.
Food will be provided after the seminar.
If accommodation is required, please email <communications@cs.illinois.edu>. Someone from our staff will contact you to discuss your specific needs.