Brain network disintegration during sedation is mediated by the complexity of sparsely connected regions

I Pappas,R.M Adapa,D.K Menon,E.A Stamatakis

doi:10.1016/j.neuroimage.2018.10.078

I Pappas, R.M Adapa + Show 2 more

Open Access

https://doi.org/10.1016/j.neuroimage.2018.10.078

Copy DOI

Journal: NeuroImage	Publication Date: Nov 1, 2018
Citations: 22	License type: cc-by

Affiliation: University of Cambridge

Abstract

The precise mechanism of anaesthetic action on a neural level remains unclear. Recent approaches suggest that anaesthetics attenuate the complexity of interactions (connectivity) however evidence remains insufficient. We used tools from network and information theory to show that, during propofol-induced sedation, a collection of brain regions displayed decreased complexity in their connectivity patterns, especially so if they were sparsely connected. Strikingly, we found that, despite their low connectivity strengths, these regions exhibited an inordinate role in network integration. Their location and connectivity complexity delineated a specific pattern of sparse interactions mainly involving default mode regions while their connectivity complexity during the awake state also correlated with reaction times during sedation signifying its importance as a reliable indicator of the effects of sedation on individuals. Contrary to established views suggesting sedation affects only richly connected brain regions, we propose that suppressed complexity of sparsely connected regions should be considered a critical feature of any candidate mechanistic description for loss of consciousness.

Highlights

Neuroimaging techniques have made possible the quantification of blood oxygen level dependent (BOLD) signal correlations from brain regions of interest, known as functional connectivity, showing reproducible patterns across experiments and individuals (Dosenbach et al, 2007)
We first asked whether the global entropy of the degree distribution of functional connectivity networks changes with increasing sedation (Fig. 1)
A repeated-measures analysis of variance (ANOVA) showed a significant effect of sedation on global entropy (F3;72 1⁄4 3:39; p 1⁄4 0:0226) (Fig. 2a) and post hoc ttests showed that entropy was decreased during moderate sedation compared to the awake state (p 1⁄4 0:0137)

Summary

Introduction

Neuroimaging techniques have made possible the quantification of blood oxygen level dependent (BOLD) signal correlations from brain regions of interest, known as functional connectivity, showing reproducible patterns across experiments and individuals (Dosenbach et al, 2007). One fundamental property of brain networks is how well connected each region is to the rest of the brain as quantified by its number of edges (degree). Studies have shown that the degree sample of resting-state functional connectivity networks (where participants do not engage with any specific task) conforms to distributions that follow a non-trivial heavy-tailed pattern (Vertes et al, 2012). The complexity of this distribution is a critical trait of functional connectivity networks, emphasising the co-existence of highly connected regions and sparsely connected regions that together support local and global integration in the brain (Zamora-Lopez et al, 2016). While some propose alterations in highly connected regions or hubs (Chennu et al, 2017; Crossley et al, 2014) others have shown alterations in regions with a small number of connections (Achard et al, 2012) precluding the specific characterization of changes in the degree distribution

Objectives

Results

Discussion

Conclusion