Figure 2.

A) A two-site exchange model that includes pyruvate vascular input was applied for the simulation of spin-echo RF spoiling effects. Spoiling was modeled as magnetization loss in the input function M_pyr,A(t) with equivalent decay factor T_DSE. B) With stronger DSE spoiling (i.e. a shorter time constant), the pyruvate signal decays more rapidly than expected. This leads to poor pyruvate fits and overestimation of k_PL if a fixed AIF is used, which can be eliminated with an inputless fitting that makes no assumptions about the AIF [15,28]. Note that M_AIF(t) is the AIF magnetization curve, while S_x(t) is the signal curve for metabolite x. In fact, the magnetization and signal curves have independent scaling and were only plotted together for illustration purposes.