Statistics and Data Science Seminar
Abstract: As improved recording technologies have created new opportunities for neurophysiological investigation, interest has shifted toward dynamic evolution in neural activity across multiple populations that form circuits. In this lecture I will present some results from several studies of cross-population interactions that have behaviorally-relevant timing, and will quickly review the general ideas behind the methods (along with a few salient details). I will also provide comments on four ubiquitous issues: the definition of neural populations, trial-to-trial variability and Poisson-like noise, time-varying dynamics, and causality. Although my examples involve neural spike trains and local field potentials, many of the comments, and methods, are also relevant to other kinds of data arising in neuroscience (and elsewhere). The majority of my presentation should be accessible to audience members without advanced statistical training, but I will also include some remarks aimed at statistically-oriented data scientists.
Bio: Robert E. Kass is the Maurice Falk Professor of Statistics and Computational Neuroscience in the Department of Statistics & Data Science, the Machine Learning Department, and the Neuroscience Institute at Carnegie Mellon University. Kass's early work was on Bayesian inference, and on differential geometry in statistics; since 2000 his interest has focused on statistical methods in neuroscience. He received the Outstanding Statistical Application Award from the American Statistical Association and the Distinguished Achievement Award and Lectureship from the Committee of Presidents of Statistical Societies. He is an elected Fellow of the American Statistical Association, the Institute of Mathematical Statistics, and the American Association for the Advancement of Science, and an elected member of the National Academy of Sciences.
Host: Debashis Mondal