For the first time, a three-dimensional map of the distribution of water in a test sample has been obtained from the random radio signal (spin noise) emitted spontaneously by hydrogen nuclei in a magnetic field with varying field gradients. A special variant of a projection–reconstruction algorithm has been developed for noise data, which allows one to adjust the image quality between high resolution/low contrast and low resolution/high contrast from the same previously recorded spin noise data.

Detailed interpretation of brain MRI data is hampered by the fact that each imaging voxel comprises several types of cells and tissues. To address this, we adapt signal encoding and data inversion strategies from solid-state and low-field NMR to quantify the sub-voxel heterogeneity of the human brain with 5D relaxation–diffusion distributions wherein distinct tissue components are resolved, individually characterized, and subsequently mapped throughout the volume of the brain.