Detecting dendritic spine density with double diffusion encoding magnetic resonance spectroscopy

Motivation: Studies shown that single diffusion encoding measurements of metabolites diffusion may be sensitive to dendritic spines, while Double Diffusion Encoding (DDE) increases sensitivity to microstructure features like dendritic length. Goal(s): Investigate whether DDE measurements can be sensitive to spine density and used to map it in-vivo. Approach: Simulated DDE signals from models of spiny dendrites were used to train an artificial neural network to predict spine density from DDE signals. Results: Spine density estimated from in-vivo data in mice expressing amyloid precursor protein match the known ~50% decrease in spine density from this animal model. Impact: Using numerical simulations and in-vivo mouse experiments we showed that DDE-MRS may be a promising non-invasive method for estimating dendritic spine density in-vivo, providing a new avenue for in-vivo studies of healthy and pathological brain gray matter microstructure.