Predictive information in the salamander retina

Abstract

Event Abstract Back to Event Predictive information in the salamander retina Stephanie Palmer1*, Greg W. Schwartz2, Michael J. Berry2 and William Bialek1 1 Princeton University, Department of Physics and Biocenter Oulu,, United States 2 Princeton University, Department of Molecular Biology, United States Prediction is important for almost all modes of behavior and our research focuses on how a population of neurons implements predictive computations. We have examined how groups of retinal ganglion cells (the output neurons of the retina) encode predictive information in their collective firing patterns. The population response is represented as a binary word, indicating a spike or no-spike from each neuron in a small time window of size Δt. We can then ask how precisely a word at time t specifies a future word at time t+Δt. This is the predictive information in the population firing. We next construct a synthetic neuron that receives these inputs, and gives as output a single bit (spike or no-spike). We ask how this downstream neuron, combining subgroups of its potential presynaptic cells, would learn to become maximally predictive of its inputs, in the sense that its output spiking would carry maximal information about its input pattern at the next time step. Can this output capture the available predictive information? What is the structure of the algorithm that maps inputs to predictive outputs? For four-cell subgroups, we can exhaustively search all possible deterministic rules for converting input responses into a binary output, and we use this to test our search strategies for larger groups. We find that the best rules can capture more than 95\% of the predictive information in the input firing. Rules which capture such a large percentage of the predictive information also retain a large amount of stimulus information about the visual world. These best rules can be reliably learned by a perceptron model, meaning that instantiating such a coding of predictive information in the brain might be biologically plausible. Conference: Bernstein Symposium 2008, Munich, Germany, 8 Oct - 10 Oct, 2008. Presentation Type: Poster Presentation Topic: All Abstracts Citation: Palmer S, Schwartz GW, Berry MJ and Bialek W (2008). Predictive information in the salamander retina. Front. Comput. Neurosci. Conference Abstract: Bernstein Symposium 2008. doi: 10.3389/conf.neuro.10.2008.01.102 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 17 Nov 2008; Published Online: 17 Nov 2008. * Correspondence: Stephanie Palmer, Princeton University, Department of Physics and Biocenter Oulu,, New Jersey, United States, sepalmer@princeton.edu Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Stephanie Palmer Greg W Schwartz Michael J Berry William Bialek Google Stephanie Palmer Greg W Schwartz Michael J Berry William Bialek Google Scholar Stephanie Palmer Greg W Schwartz Michael J Berry William Bialek PubMed Stephanie Palmer Greg W Schwartz Michael J Berry William Bialek Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.