A fractal based model of diffusion MRI in cortical grey matter

Center of Functionally Integrative Neuroscience (CFIN)University of Aarhus / Aarhus University Hospital, DenmarkAbstractDiffusion Weighted Magnetic Resonance (DWMR) Imaging is an important tool indiagnostic neuroimaging, but the biophysical basis of the DWMR signal from bio-logical tissue is not entirely understood. Testable, theoretical models relating theDWMR signal to the tissue, therefore, are crucial. This work presents a toy versionof such a model of water DWMR signals in brain grey matter. The model is basedon biophysical characteristics and all model parameters are directly interpretableas biophysical properties such as diffusion coefficients and membrane permeabilityallowing comparison to known values. In the model, a computer generated Diffu-sion Limited Aggregation (DLA) cluster is used to describe the collected membranemorphology of the cells in cortical grey matter. Using credible values for all modelparameters model output is compared to experimental DWMR data from normalhuman grey matter and it is found that this model does reproduce the observedsignal. The model is then used for simulating the effect on the DWMR signal ofcellular events known to occur in ischemia. These simulations show that a combi-nation of effects is necessary to reproduce the signal changes observed in ischemictissue and demonstrate that the model has potential for interpreting DWMR sig-nal origins and tissue changes in ischemia. Further studies are required to validatethese results and compare them with other modeling approaches. With such mod-els, it is anticipated that sensitivity and specificity of DWMR in tissues can beimproved, leading to better understanding of the origins of MR signals in biologicaltissues, and improved diagnostic capability.Key words: Diffusion MRI, numerical simulations, stroke, grey matter, ischemia.