Revisiting double diffusion encoding MRS in the mouse brain at 11.7T: which microstructural features are we sensitive to?

TL;DR

This study investigates how double diffusion encoding (DDE) MRI signals in the mouse brain at 11.7T are influenced by various microstructural features like cell shape, size, and branching, using experiments and simulations to improve microstructure interpretation.

Contribution

The paper demonstrates that incorporating branched fiber structures into DDE models enhances the interpretation of microstructural features, moving beyond the traditional infinite cylinder assumption.

Findings

Infinite cylinder model poorly fits experimental data.

Branched fiber structures are crucial for accurate DDE signal interpretation.

Potential sensitivity to cell body diameter at short mixing times.

Abstract

Brain metabolites, such as N-acetylaspartate or myo-inositol, are constantly probing their local cellular environment under the effect of diffusion. Diffusion-weighted NMR spectroscopy therefore presents unparalleled potential to yield cell-type specific microstructural information. Double diffusion encoding (DDE) relies on two diffusion blocks which relative directions describe a varying angle during the course of the experiment. Unlike single diffusion encoding, DDE measurements at long mixing time display some angular modulation of the signal amplitude which reflects compartment shape anisotropy, while requiring relatively low gradient strength. This angular dependence has been formerly used to quantify cell fiber diameter using a model of isotropically oriented infinite cylinders. However, it has been little explored how additional features of the cell microstructure, such as cell body diameter, fiber length and branching may also influence the DDE signal. Here, we used a cryoprobe as well as state-of-the-art post-processing to perform DDE acquisitions with high accuracy and precision in the mouse brain at 11.7 T. We then compared our results to simulated DDE datasets obtained in various 3D cell models in order to pinpoint which features of cell morphology may influence the most the angular dependence of the DDE signal. While the infinite cylinder model poorly fits our experimental data, we show that incorporating branched fiber structure in our model is paramount to sensibly interpret the DDE signal. Lastly, experiments and simulations in the short mixing time regime suggest that some sensitivity to cell body diameter might be retrieved, although additional experiments would be required to further support this statement.