Konrad Kording is the Nathan Mossell University Penn Integrates Knowledge (PIK) Professor of bioengineering and neuroscience at the University of Pennsylvania and a co-director of the CIFAR Learning in Machines & Brains program. He is interested in understanding the brain as a computational device and how to mine neural data for causal relations. He uses data analysis methods, including machine learning and Bayesian models, to ask fundamental questions about the brain, behavior, and disease. He is also an advocate and contributor to open science and scientific rigor.
Let us simulate a nervous system
Being able to simulate a nervous system is clearly one of the salient goals of systems neuroscience. Being able to do so we need two things: a parts list and a functional description of the neurons. We need to know all the neurons and need to know how they (nonlinearly) influence one another. And yet, we have so far always tried a very indirect approach: observe activities of neurons, and from that data estimate the influences. However, from a causal inference perspective it is clear that this problem is generally ill-posed. In my talk I will sketch what exactly it will take to simulate a complete nervous system: the nervous system of C. elegans. I view this as a crucial and necessary step towards simulating larger nervous systems.