Sleeping our way to weight normalization and stable learning

Authors:
Thomas J. Sullivan;Virginia R. de Sa
Affiliations:
Department of Biology, University of California, San Diego, La Jolla, CA 92093, U.S.A. tom@sullivan.to;Department of Cognitive Science, University of California, San Diego, La Jolla, CA 92093, U.S.A. desa@cogsci.ucsd.edu
Venue:
Neural Computation
Year:
2008

Citing 4
Cited 1

Visual Discrimination Task Improvement: A Multi-Step Process Occurring During Sleep

Journal of Cognitive Neuroscience
Self-organizing neural projections

Neural Networks - 2006 Special issue: Advances in self-organizing maps--WSOM'05
Homeostatic synaptic scaling in self-organizing maps

Neural Networks - 2006 Special issue: Advances in self-organizing maps--WSOM'05
Synergies Between Intrinsic and Synaptic Plasticity Mechanisms

Neural Computation

Hebbian plasticity and homeostasis in a model of hypercolumn of the visual cortex

Neural Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

The functions of sleep have been an enduring mystery. Tononi and Cirelli (2003) hypothesized that one of the functions of slow-wave sleep is to scale down synapses in the cortex that have strengthened during awake learning. We create a computational model to test the functionality of this idea and examine some of its implications. We show that synaptic scaling during slow-wave sleep is capable of keeping Hebbian learning in check and that it enables stable development. We also show theoretically how it implements classical weight normalization, which has been in common use in neural models for decades. Finally, a significant computational limitation of this form of synaptic scaling is revealed through computer simulations.