Diffusion assessment through image processing: beyond the point-source paradigm

Abstract

The quantification of transport processes of different substances in the brain’s parenchyma is important in the context of understanding brain functioning. Most of the currently used methods for assessment of the effective diffusion coefficient rely on the point-source paradigm. We propose a method for the quantitative characterization of the diffusion process in the brain’s parenchyma using a set of images recorded during the spreading of a fluorescent dye. Our method uses the frame-by-frame comparison of such images with a set of computed images that would be observed for an ideal diffusion process within the same topology of the tissue’s part. We obtain this reference set of images using blurring the image with an appropriate kernel function, and the degree of such blurring correlates with the parameters of a dye spreading process and the transport coefficients. We demonstrate the applicability of the proposed method using (i) the simulated surrogate data, (ii) the set of fluorescent images of the isolated event of blood-brain barrier opening, and (iii) the images of massive multi-source spreading of fluorescent dye.