Improved Radiofrequency Safety Modelling in MRI Using In Vivo Measurements of Brain Conductivity

Figures

• 

  FIGURE 1. Workflow for obtaining both experimental and simulated B1+ maps, including combination of the two excitation ports, interpolation/resampling, alignment, and scaling.

• 

  FIGURE 2. Comparison of simulated and experimental B1+ magnitude/phase maps in the phantom experiment (both sagittal and axial planes are shown, the black line is the mask used for quantitative assessment): (a, d) simulation with incorrect conductivity (σ=0.24 S/m); (b, e) simulation with true conductivity (σ=0.47 S/m); (c, f) experimental B1+ maps (gel conductivity σ=0.47 S/m).

• 

  FIGURE 3. Comparison of simulated and experimental B1+ magnitude/phase maps in a healthy subject (sagittal, coronal and axial planes are shown, the black line is the experimental brain contour, segmented from the MPRAGE sequence): (a, e) simulation with the worst biomodel for this subject (Hugo) and the standard brain conductivity (σexvivo=0.46 S/m); (b, f) simulation with the best biomodel for this subject (Emma) and the standard brain conductivity (σexvivo=0.46 S/m); (c, g) simulation with the best biomodel for this subject (Emma) and brain conductivity (σin vivo=0.70 S/m); (d, h) experimental B1+ maps.

• 

  FIGURE 4. Comparison of SAR maps in the phantom, with different estimation methods (sagittal, coronal and axial planes are shown, the green line is the contour for the global SAR calculation): (a) simulated SAR using exact calculation and test conductivity value σtest=0.24S/m; (b) simulated SAR using B1+‐derived estimation and test conductivity value; (c) experimental SAR using B1+‐derived estimation and test conductivity value; (d, e, f) same as (a, b, c) but using brain conductivity σtrue = 0.47 S/m.

• 

  FIGURE 5. Comparison of SAR maps in a healthy subject, with different estimation methods (sagittal, coronal and axial planes are shown, the green line is the brain contour of the biomodel, in simulated maps, or the subject, in experimental maps): (a) simulated SAR using exact calculation and standard brain conductivity (σexvivo=0.46 S/m); (b) simulated SAR using B1+‐derived estimation and standard brain conductivity; (c) experimental SAR using B1+‐derived estimation and standard brain conductivity; (d, e, f) same as (a, b, c) but using brain conductivity σin vivo=0.70 S/m. Here, the biomodel used for simulation was Emma (best match for B1+ maps).