Differential effects of aquaporin-4 channel inhibition on BOLD fMRI and diffusion fMRI responses in mouse visual cortex.

Denis Le Bihan,Eliana Scemes,Boucif Djemai,Yuji Komaki,Clement Debacker,Luisa Ciobanu,Tomokazu Tsurugizawa

doi:10.1371/journal.pone.0228759

Denis Le Bihan, Eliana Scemes

Open Access

https://doi.org/10.1371/journal.pone.0228759

Copy DOI

Abstract

The contribution of astrocytes to the BOLD fMRI and DfMRI responses in visual cortex of mice following visual stimulation was investigated using TGN-020, an aquaporin 4 (AQP4) channel blocker, acting as an astrocyte function perturbator. Under TGN-020 injection the amplitude of the BOLD fMRI response became significantly higher. In contrast no significant changes in the DfMRI responses and the electrophysiological responses were observed. Those results further confirm the implications of astrocytes in the neurovascular coupling mechanism underlying BOLD fMRI, but not in the DfMRI responses which remained unsensitive to astrocyte function perturbation.

Highlights

Diffusion functional MRI (DfMRI) has been proposed as an alternative to blood oxygenation level dependent (BOLD) fMRI to monitor neural activity noninvasively [1]
While BOLD fMRI relies on the indirect neurovascular coupling mechanism [8,9] the current hypothetical mechanism of DfMRI is thought to be related to the neural activation triggered cell swelling, for which there is a large body of evidence [10,11]
local field potentials (LFP) are correlated with BOLD fMRI signal responses in normal conditions compared to multiunit activity (MUA) [26] and we have previously shown that LFPs are well correlated with DfMRI responses [17,20]

Summary

Introduction

Diffusion functional MRI (DfMRI) has been proposed as an alternative to blood oxygenation level dependent (BOLD) fMRI to monitor neural activity noninvasively [1]. Beside the established fact that water diffusion as monitored with MRI decreases in tissues undergoing cell swelling in pathological, extraphysiological and physiological conditions [12,13,14,15,16] several preclinical studies relying on pharmacological challenges interfering with neurovascular coupling or cell swelling have confirmed that (1) the DfMRI and BOLD fMRI responses could be decoupled, confirming their differential mechanisms; (2) the DfMRI response is not dependent on neurovascular coupling, but, instead, sensitive to underlying neural swelling status [17,18,19] and (3) the DfMRI response follows neural activity status closely and more accurately than BOLD fMRI, especially under anesthetic or vasoactive drug conditions [17,20].

Methods

Results

Discussion

Conclusion