An ongoing goal of research is to achieve a better understanding of the contrast mechanisms associated with myelin, including inhomogeneous magnetization transfer (ihMT). IhMT requires residual dipolar couplings between motion-restricted restricted ('semisolid') protons and a sufficiently long dipolar order relaxation time (T1D). Although the measured ihMT ratio depends strongly on T1D, direct observation and quantification of the differences in T1D of biological tissues via MRI are still not feasible. Instead, complex modelling of a limited amount of experimental data is required, which is based on compartment models of at least two proton reservoirs, indicating the need for external referencing. The Jeener-Broekaert (JB) measurement is the 'gold standard' method for T1D quantification, originally developed for the application to 'pure solids'. Recently, the JB method has been used to study lipid model systems and tissues. However, the translation from 'pure solids' to heterogeneous systems characterized by protons with different mobilities proved more challenging than initially anticipated. In particular, strong unwanted signal contributions related to 'Zeeman' order, surviving established phase-cycling schemes, biased the observed signal decay. Here, by modifying the JB sequence by adding a 180° refocusing pulse between the first two RF pulses significantly improved the overall performance, as demonstrated by experiments and additional simulations. Consequently, the reliability of T1D quantification extracted from JB measurements in the investigated lipid model system could be considerably enhanced, enabling future applications to biological tissue. A detailed discussion of additional effects influencing the observed JB signal decay, including unavoidable signal contributions from other zero-quantum coherences or exchange processes, is provided. Therefore, the quantitative T1D estimates resulting from JB experiments in heterogeneous tissue should not inevitably be considered as 'gold standard' for referencing of T1D estimates of quantitative ihMT analysis. Rather, under these conditions, it is more appropriate to refer to them as 'apparent T1D' values.
Wallstein et al. (Wed,) studied this question.