Abstract #1613

Frhlich A, stergaard L, Kiselev V

The nonlinear dependence between the logarithm of the diffusion weighted signal, lnS, and the b-value is studied. The cumulant expansion of the signal is advocated instead of using the common biexponential fit. This implies that lnS is a power series in b, which is shown to converge in a large parametric range using the basic model of a flat, impermeable boundary. To a realistic accuracy, the signal is determined by the first two coefficients, which are related to the geometry of the restrictive boundaries. In this case, the fitting parameters of the biexponential function are not defined uniquely.