Figure 7.

Result of fitting EPG and EPG‐X(MT) models to the BSA phantom data. Both values of ${\Phi }_0$ were fit simultaneously. Top row: Fits for both values of ${\Phi }_0$. Bottom row: Residuals. T₂, D, and scaling factors were held fixed at experimentally measured values. For EPG‐X(MT), the k_a, T_1,a, T_1,b, f, and G(0) were varied; the best‐fit parameters were k_a = 6.2 s⁻¹, T_1,a = 1763 ms, T_1,b = 363 ms, G(0) = 29.4 µs, f = 0.100, which would yield ${\mathrm{T}}_1^{\text{obs}}$ = 1283 ms. For single‐compartment EPG, the T₁ was varied; the best fit was 1215 ms. The NRMSE before fit = 9.8%, single‐compartment EPG fit = 6.9%, EPG‐X = 1.3%.