Characterization of open issues in low-frequency computational dosimetry

Two international organizations, namely the International Commission on Non-Ionizing Radiation Protection (ICNIRP) and the Institute of Electrical and Electronics Engineers International Committee on Electromagnetic Safety (IEEE ICES), have established exposure criteria and safety limits for human protection to electromagnetic fields. In the low-frequency range, both organizations recognize that the main adverse health effects are represented by the induction of retinal phosphenes, the alteration of synaptic activity and the stimulation of nerves. On this basis, the exposure limits were derived from threshold data of internal electric fields with the purpose of avoiding such adverse effects. Since direct measurement of the induced electric field is not feasible, both the standard and guidelines have introduced limits for external electric and magnetic field strengths. In this context, computational dosimetry was used to relate the internal induced quantities with the external field strengths. However, low-frequency dosimetry suffers from various sources of error and uncertainty.The main aim of the present thesis is to lessen such uncertainty, as well as further characterize computational artifacts in the evaluation of the induced electric fields. Investigations were carried out using state-of-the-art methods based on physiological measurements, high-resolution realistic anatomical models, individualized electric field computations and biological axon models. Several open issues affecting low-frequency dosimetry have been characterized with the aim of producing quantitative data useful for the harmonization and revision of current exposure standard and guidelines. Our findings showed a large margin of safety in the current exposure limits established by both international organizations. In this regard, the obtained results represent a solid basis for deriving safety levels that offer acceptable protection for the human population without being overly conservative. In addition, the present work improves the reliability of human exposure assessment at low frequencies.