Assessing the reliability of diffuse correlation spectroscopy models on noise-free analytical Monte Carlo data

It is shown that an analytical noise-free implementation of Monte Carlo simulations [Appl. Opt. 54, 2400 (2015)] for diffuse correlation spectroscopy (DCS) may be successfully used to check the ability of a given DCS model to generate a reliable estimator of tissue blood flow. As an example, four different DCS models often found in the scientific literature are tested on a simulated tissue (semi-infinite geometry) with a Maxwell- Boltzmann probability distribution function for red blood cell speed. It is shown that the random model is the best model for the chosen speed distribution but that (1) some inaccuracies in the DCS model in taking into account red blood cell concentration and (2) some inaccuracies, probably due to a low-order approximation of the DCS model, are still observed. The method can be easily generalized for other speed/flow probability distribution functions of the red blood cells.