Computational low-frequency electromagnetic dosimetry based on magnetic field measurements

This paper compares different experimental-computational strategies for the estimation of electric fields induced in human bodies by low-frequency magnetic sources characterized by a set of magnetic field measurements. The analysis is carried out by considering three alternative procedures, which use, as the first input, the distribution of the magnetic flux density in a volume containing the studied body or on a surface surrounding the sources. The comparison is performed on a realistic model problem, related to transcranial magnetic stimulation (TMS), in which numerically simulated 'virtual measurements' are employed. The comparative analysis is developed in terms of both result accuracy and robustness against noisy input due to unavoidable experimental uncertainties. It results that by performing the measurements on a surface surrounding the sources, a significant reduction of the experimental burden is found with respect to the case of volume measurements, affecting neither the accuracy nor the robustness of the procedure. In particular, when whole body electric field evaluation must be carried out, the advantage of surface measurements with respect to volume ones becomes significant. Moreover, a preferable scheme obtained as hybridization of previously proposed strategies is identified. Besides the adoption of a TMS model problem in the comparison procedure, the achieved result can be extended to any low-frequency dosimetric assessments, where the magnetic sources are difficult to model or not completely known.