Spherical convolutional neural networks can improve brain microstructure estimation from diffusion MRI data

TL;DR

This paper demonstrates that spherical convolutional neural networks can enhance the estimation of brain microstructure from diffusion MRI data, outperforming traditional methods and generalizing to different compartment models.

Contribution

It introduces a spherical CNN approach for microstructural parameter estimation that improves accuracy and rotational invariance over existing techniques.

Findings

Spherical CNNs outperform spherical mean technique in accuracy.

The method reduces rotational variance compared to multi-layer perceptrons.

The approach is adaptable to different Gaussian compartment models.

Abstract

Diffusion magnetic resonance imaging is sensitive to the microstructural properties of brain tissue. However, estimating clinically and scientifically relevant microstructural properties from the measured signals remains a highly challenging inverse problem that machine learning may help solve. This study investigated if recently developed rotationally invariant spherical convolutional neural networks can improve microstructural parameter estimation. We trained a spherical convolutional neural network to predict the ground-truth parameter values from efficiently simulated noisy data and applied the trained network to imaging data acquired in a clinical setting to generate microstructural parameter maps. Our network performed better than the spherical mean technique and multi-layer perceptron, achieving higher prediction accuracy than the spherical mean technique with less rotational variance than the multi-layer perceptron. Although we focused on a constrained two-compartment model of neuronal tissue, the network and training pipeline are generalizable and can be used to estimate the parameters of any Gaussian compartment model. To highlight this, we also trained the network to predict the parameters of a three-compartment model that enables the estimation of apparent neural soma density using tensor-valued diffusion encoding.