On random walks and entropy in diffusion‐weighted magnetic resonance imaging studies of neural tissue

摘要: Purpose In diffusion-weighted MRI studies of neural tissue, the classical model assumes statistical mechanics Brownian motion and predicts a monoexponential signal decay. However, there have been numerous reports decays that are not monoexponential, particularly in white matter. Theory We modeled diffusion tissue from perspective continuous time random walk. The characteristic decay is represented by Mittag-Leffler function, which relaxes priori assumptions about governing statistics. We then used entropy as measure anomalous features for function. Methods Diffusion-weighted experiments were performed on fixed rat brain using an imaging spectrometer at 17.6 T with b-values arrayed up to 25,000 s/mm2. Additionally, we examined impact varying either gradient strength, q, or mixing time, Δ, observed dynamics. Results In gray matter regions, parameters demonstrated new information regarding subdiffusion produced different image contrast coefficient. choice weighting q Δ within regions interest. Conclusion We propose these potential biomarkers morphology tissue. Magn Reson Med 71:617–627, 2014. © 2013 Wiley Periodicals, Inc.