Brain stroke monitoring using compressive sensing and higher order basis functions

We consider the utilization of higher-order basis functions, in the sparse processing framework, for brain stroke monitoring. Instead of retrieving the permittivity of the whole brain, the goal is estimate the variation of the electromagnetic parameters of the brain between two measurements. We assume that the differences in the measured electromagnetic field indicate the stroke evolution. Using average head model and different noise levels, we show that the method yields accurate results even for low signal-to-noise ratios (SNR) and limited prior knowledge of the brain tissue parameters.