Anomalous Diffusion and Non-Monoexponential b-decay

M. G. Hall, C. Mackay, M. D. Robson, T. R. Barrick Centre for Medical Imaging Computing, Dept of Computer Science, University College London, London, United Kingdom, OCMR, University of Oxford, Oxford, United Kingdom, Dept of Clinical Neuroscience, St Georges, University of London, London, United Kingdom Introduction Significant departures from the standard monoexponential diffusion model have been reported in experimental data acquired in rats [2] and in humans [e.g. 3-4]. One suggestion for this deviation is the biexponential model in which two populations of spins with different diffusion constants are present in imaged brain tissue. However, this model involves a complex fitting problem in several parameters [4]. Here we present an alternative interpretation of non-monoexponential signal decay for diffusion in terms of the theory of Anomalous Diffusion [1] and show that simple scaling arguments lead to an experimentally-testable hypothesis. We reproduce the non-monoexponential decay observed previously [2-4] and obtain results that provide support for the anomalous diffusion hypothesis. Furthermore, we conclude that non-mono-exponential decay may occur due to non-locally homogeneous disorder in the diffusion environment.