Diffusion Weighted Imaging correlates with cell numbers in healthy and pathological human brains

Background

Diffusion-weighted imaging (DWI) is an MRI technique that leverages the natural thermal Brownian motion. This technique provides insights into tissue microstructure. The absolute number of cells is an important quantitative measure of brain tissue microstructure. We studied if the structural findings of total cellularity in postmortem brains of controls and subjects with Alzheimer’s Disease (AD) are related to the observed changes in MRI parameters.

Methods

All procedures were approved by institutional ethics committees. Brains of the Biobank for Aging Studies were scanned using postmortem structural and DWI MRI protocols. The Medial Temporal Lobes (MTL, n=17) and Olfactory Bulbs (OB, n=18) were regions of interest. After imaging, these regions were dissected from typical controls and subjects with dementia, and the tissue was processed with the Isotropic Fractionator for cell quantification. We also analyzed in vivo images to investigate the structural and diffusion MRI changes in MTL in a cohort of AD patients (n=14) compared to Mild Cognitive Impairment (n=31) patients and controls (n=39). 

Results

We found moderate correlations between diffusivity metrics (MD, AxD, RD), cerebral cellularity and the burden of AD neuropathological changes (amyloid plaques and neurofibrillary tangles). Using in vivo MRI, we also demonstrated structural and diffusion differences in the MTL of subjects with AD pathology. 

Conclusions

We suggest that DWI is more sensitive than conventional volumetric techniques to evaluate microstructural alterations such as changes in cell number. Among DWI parameters, MD could become a possible marker of brain cellularity by neuroimaging. However, a single MRI parameter cannot fully explain the changes in cell number; in the models tested to predict this number, it is necessary to consider sex, age, presence of cognitive decline, volume, and diffusion parameters to predict more accurately the number of cells in a given region of interest.