General anaesthesia reduces the uniqueness of brain connectivity across individuals and species

2465 

Abstract Submission 

Andrea Luppi¹, Daniel Golkowski², Andreas Ranft², Rudiger Ilg², Denis Jordan², Danilo Bzdok³, Adrian Owen⁴, Lorina Naci⁵, Emmanuel Stamatakis⁶, Enrico Amico⁷, Bratislav Misic⁸

¹McGill University, Buguggiate, Italy, ²Technical University Munich, Munich, Germany, ³McConnell Brain Imaging Centre (BIC), Montreal Neurol, McGill Universityogical Institute (MNI), Montreal, Quebec, ⁴Western University, London, Ontario, ⁵Trinity College Dublin, Dublin, Ireland, ⁶University of Cambridge, Cambridge, Cambridgeshire, ⁷Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland, ⁸McGill University, Montreal, Quebec

Andrea Luppi, PhD  
McGill University
Buguggiate, Italy

Daniel Golkowski  
Technical University Munich
Munich, Germany

Andreas Ranft  
Technical University Munich
Munich, Germany

Rudiger Ilg  
Technical University Munich
Munich, Germany

Denis Jordan  
Technical University Munich
Munich, Germany

Danilo Bzdok  
McConnell Brain Imaging Centre (BIC), Montreal Neurol, McGill Universityogical Institute (MNI)
Montreal, Quebec

Adrian Owen  
Western University
London, Ontario

Lorina Naci  
Trinity College Dublin
Dublin, Ireland

Emmanuel Stamatakis  
University of Cambridge
Cambridge, Cambridgeshire

Enrico Amico  
Ecole Polytechnique Federale de Lausanne
Lausanne, Switzerland

Bratislav Misic  
McGill University
Montreal, Quebec

The human brain is characterised by idiosyncratic patterns of spontaneous thought, rendering each brain uniquely identifiable from its neural activity (Amico and Goñi, 2018; Finn et al., 2015). However, deep general anaesthesia suppresses subjective experience. Does it also suppress what makes each brain unique? We attack this question from three conceptual angles. First, we compare brain connectivity within and across individuals. Second, we compare brain activity against the canonical brain maps of human cognition. Third, we ask whether anaesthesia makes the human brain less distinctive from another species: the macaque.

We consider resting-state functional MRI data from N=15 volunteers at baseline and at different levels of the anaesthetic sevoflurane: at burst-suppression, 3 vol%, 2 vol%, and post-anaesthetic recovery of responsiveness (Ranft et al., 2016).

Here we use connectome fingerprinting of individuals based on fMRI scans to evaluate whether individuals become less distinguishable when under anaesthesia, based on the similarity (correlation) between functional connectivity patterns at different points in time (Amico and Goni, 2018; Finn et al., 2015). We compare the patterns of regional changes in identifiability against brain maps of evolutionary cortical expansion (Xu et al., 2020), human-accelerated gene expression (Wei et al., 2019), inter-individual variability (Mueller et al., 2013), and the archetypal sensory-association axis of cortical organisation (Sydnor et al., 2021), obtained from the neuromaps toolbox (Markello et al., 2022). To control for spatial autocorrelation, we employ autocorrelation-preserving null models (Markello et al., 2022).

We also use 123 meta-analytic patterns from NeuroSynth (Yarkoni et al., 2011), and assess how well their spatial distribution matches the spatial patterns of spontaneous brain activity. Finally, we compare functional connectivity across levels of anaesthesia, against the functional connectivity of the macaque brain (Michael Milham et al., 2018). We replicate all results in an independent dataset of resting-state fMRI from N=16 volunteers scanned before, during, and after loss of responsiveness induced by propofol.

When comparing two scans of the same individual while awake, it is easy to distinguish self from other. However, self-self correlations between awake and anaesthetised brains are significantly diminished compared to awake-recovery self-similarity (Fig.1a-d). We localise regional changes in identifiability, quantified as the mean change in intraclass correlation coefficient (ICC) across each region's edges (Amico and Goñi, 2018; Finn et al., 2015). We observe a significant spatial correlation with the brain's archetypal sensory-association axis (ρ = 0.67, p < 0.001) as well as inter-individual variability (ρ = 0.63, p < 0.001). Changes in regional identifiability correlate with molecular and morphometric markers of phylogenetic cortical differentiation between human and non-human primates: evolutionary expansion (ρ = 0.35, p < 0.001), and the regional mean expression of human-accelerated ("HAR") genes (ρ = 0.42, p < 0.001) (Fig.1e-h).

As anaesthesia deepens, the quality of cognitive matching deteriorates: the best spatial correlation between brain activity and meta-analytic brain maps from NeuroSynth is lower at deeper levels of anaesthesia (Fig.2a,b). At increasing concentration of sevoflurane, we also observe increasing similarity between the regional functional connectivity patterns of humans and macaques (Fig.2c). The effects of sevoflurane anaesthesia are reversed upon recovery, and are replicated with a different anaesthetic, propofol.

Altogether, the present results indicate that regardless of the specific anaesthetic used, anaesthetised human brains are less unique, both across individuals and even across species, with regions that are most heterogeneous across individuals and across species being especially affected.

Transcriptomics

Connectivity (eg. functional, effective, structural)

Anatomy and Functional Systems

Consciousness and Awareness ¹

Pharmacology and Neurotransmission ²

CHEMOARCHITECTURE

Consciousness

Cross-Species Homologues

FUNCTIONAL MRI

Meta- Analysis

Other - Brain fingerprinting; Anesthesia; Cross-species comparison; Evolutionary Expansion; Human-Accelerated Genes