An Efficient Deterministic-Stochastic Model for the Homogeneous Human Brain Model Dosimetry: ANOVA Approaches for Sensitivity Analysis of Model Parameters

Abstract

In high frequency dosimetry of human brain it is necessary to develop computational models to estimate the temperature distribution inside the brain, since its direct measurement is not possible. In order to describe an interaction between the high frequency electromagnetic field and the human brain, models may become very complex with numerous input variables whose values are rarely exactly known. It is important to have a good understanding of the relationship between the input and output parameters. In this paper ANalysis Of Variance (ANOVA) approach coupled with the stochastic collocation (SC) expansion method is applied on homogeneous human brain model in order to obtain a sensitivity analysis (SA) of input parameters. The results enable the ranking of the parameters from the most to the least significant ones. Moreover, the impact of the mutual interaction of input variables is obtained.

Mario Cvetković

Associate Professor | Department of Electrical Engineering Fundamentals

Associate professor at FESB in Split, with a research focus on numerical modeling including finite element and moment methods, computational bioelectromagnetics and heat transfer related phenomena. He is involved in IEEE’s ICES Technical Committee 95, various international projects and is committed to advancing both knowledge and practical applications in electromagnetic safety and biomedical engineering.