Efficient coding of binocular spontaneous activity for innate learning in V1 development

Abstract

Event Abstract Back to Event Efficient coding of binocular spontaneous activity for innate learning in V1 development We present a model of spontaneous activity in the developing LGN which shows how simple patterns of activity can train the early visual system for natural binocular vision before the eyes open. This model uniquely applies principles of efficient coding (sparse coding/ICA) to show that both spontaneous activity and natural stimuli can lead to a code much like that found in primary visual cortex. Spontaneous, patterned neural activity plays a necessary role in visual system development. For example, spontaneous activity has been recorded in the developing retinae of a variety of animals including turtles, chicks, cats, mice, and monkeys; these retinal waves are necessary for proper LGN layer segregation and refinement. However, there has been converging evidence both experimentally and theoretically that spontaneous activity in the developing visual system is not only permissive but also instructive. A previous model by Albert, Schnabel, and Field [1] has shown that an efficient learning algorithm used for natural scenes can be applied to spontaneous activity patterns to generate a visual code resembling V1 simple cells. However, models based on insights from retinal waves have some important limitations, such as independent activity across the eyes. Here we present a model which further demonstrates how spontaneous activity can play an instructive role in visual development without these retinal limitations. We generalize our spontaneous activity model to fit known prenatal activity statistics in the LGN. The model presented here includes the binocular correlations demonstrated in LGN activity by Weliky and Katz [2]. This work demonstrates correlations in spontaneous activity across eye-layers in the LGN, and the data indicates that such correlations may be from a traveling wavefront of activity in the LGN. The inclusion of these partial correlations generates binocular filters with a disparity distribution similar to that found in animals at eye-opening. This indicates that these patterns may be important for establishing both monocular and binocular properties of visual neurons. Although there is much less data on LGN/V1 spontaneous activity, we believe that models such as this one will help in guiding experiments in this area, and further establishing the role of this activity as instructive for visual development. Conference: Computational and systems neuroscience 2009, Salt Lake City, UT, United States, 26 Feb - 3 Mar, 2009. Presentation Type: Poster Presentation Topic: Poster Presentations Citation: (2009). Efficient coding of binocular spontaneous activity for innate learning in V1 development. Front. Syst. Neurosci. Conference Abstract: Computational and systems neuroscience 2009. doi: 10.3389/conf.neuro.06.2009.03.302 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: 04 Feb 2009; Published Online: 04 Feb 2009. 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 Google Google Scholar PubMed 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.