Sleep-like slow oscillations improve visual classification through synaptic homeostasis and memory association in a thalamo-cortical model

Cristiano Capone,Bruno Golosio,Pier Stanislao Paolucci,Elena Pastorelli

doi:10.1038/s41598-019-45525-0

Abstract

The occurrence of sleep passed through the evolutionary sieve and is widespread in animal species. Sleep is known to be beneficial to cognitive and mnemonic tasks, while chronic sleep deprivation is detrimental. Despite the importance of the phenomenon, a complete understanding of its functions and underlying mechanisms is still lacking. In this paper, we show interesting effects of deep-sleep-like slow oscillation activity on a simplified thalamo-cortical model which is trained to encode, retrieve and classify images of handwritten digits. During slow oscillations, spike-timing-dependent-plasticity (STDP) produces a differential homeostatic process. It is characterized by both a specific unsupervised enhancement of connections among groups of neurons associated to instances of the same class (digit) and a simultaneous down-regulation of stronger synapses created by the training. This hierarchical organization of post-sleep internal representations favours higher performances in retrieval and classification tasks. The mechanism is based on the interaction between top-down cortico-thalamic predictions and bottom-up thalamo-cortical projections during deep-sleep-like slow oscillations. Indeed, when learned patterns are replayed during sleep, cortico-thalamo-cortical connections favour the activation of other neurons coding for similar thalamic inputs, promoting their association. Such mechanism hints at possible applications to artificial learning systems.

Highlights

Human brains spend about one-third of their life-time sleeping
During sleep external perceptions are, at least, strongly attenuated, and the majority of the motor system is blocked[15]. For this reason in our model the local interaction between cortex and thalamus is crucial during sleep rather than contextual signal coming from other cortical modules and sensory input coming from thalamic pathways
We propose a minimal thalamo-cortical model that classifies images drawn from the MNIST set of handwritten digits

Summary

Introduction

Human brains spend about one-third of their life-time sleeping. Sleep is present in every animal species that has been studied[1]. Having survived the evolutionary selection in all species, sleep must provide strong advantages Another notable fact is that newborns’ human brains occupy the majority of their time asleep, they learn at a very fast rate. In4 Watson et al propose a novel intriguing experimental evidence They used large-scale recordings to examine the activity of neurons in the frontal cortex of rats and observed that neurons with different pre-sleep firing rate are differentially modulated by different sleep substates (REM, non-REM and micro arousal). High-frequency bursts of spikes are emitted when the coincidence is detected Relying on these observation we introduced in our model external stimuli mimiking contextual information which changes the effective firing threshold of specific subsets of neurons during the presentation of examples in the training phase. Due to the change in the perceptual effective firing threshold, spike-timing-dependent-plasticity (STDP) creates stronger bottom-up (thalamo-cortical) and top-down (cortico-thalamic) connections between a subset of cortical and the thalamic neurons

Objectives

Methods

Results

Conclusion