To investigate the long-and short-T 1D components correlation with myelin content using inhomogeneous magnetization transfer (ihMT) high-pass and band-pass T 1D -filters and to compare ihMT, R 1 , and the macromolecular proton fraction (MPF) for myelin specific imaging. Methods:The 3D ihMT rapid gradient echo (ihMTRAGE) sequences with increasing switching times (Δt) were used to derive ihMT high-pass T 1D -filters with increasing T 1D cutoff values and an ihMT band-pass T 1D -filter for components in the 100 µs to 1 ms range. 3D spoiled gradient echo quantitative MT (SPGR-qMT) protocols were used to derive R 1 and MPF maps. The specificity of R 1 , MPF, and ihMT T 1D -filters was evaluated by comparison with two histological reference techniques for myelin imaging. Results:The higher contribution of long-T 1D s as compared to the short components as Δt got longer led to an increase in the specificity to myelination. In contrast, focusing on the signal originating from a narrow range of short-T 1D s (< 1 ms) as isolated by the band-pass T 1D -filter led to lower specificity. In addition, the significantly lower r 2 correlation coefficient of the band-pass T 1D -filter suggests that the origin of short-T 1D components is mostly associated with non-myelin protons. Also, the important contribution of short-T 1D s to the estimated MPF, explains its low specificity to myelination as compared to the ihMT high-pass T 1D -filters. Conclusion:Long-T 1D components imaging by means of ihMT high-pass T 1Dfilters is proposed as an MRI biomarker for myelin content. Future studies should | METHODS | Animal experimentsAnimal studies were conducted in agreement with the French guidelines for animal care from the French Department of Agriculture (Animal Rights Division), the directive 2010/63/EU of the European Parliament and of enable the investigation of the sensitivity of ihMT T 1D -filters for demyelinating processes.

To identify T 1D -filtering methods, which can specifically isolate various ranges of T 1D components as they may be sensitive to different microstructural properties.Methods: Modified Bloch-Provotorov equations describing a bi-T 1D component biophysical model were used to simulate the inhomogeneous magnetization transfer (ihMT) signal from ihMTRAGE sequences at high RF power and low duty-cycle with different switching time values for the dual saturation experiment: Δt = 0.0, 0.8, 1.6, and 3.2 ms. Simulations were compared with experimental signals on the brain gray and white matter tissues of healthy mice at 7T. Results:The lengthening of Δt created ihMT high-pass T 1D -filters, which efficiently eliminated the signal from T 1D components shorter than 1 ms, while partially attenuating that of longer components (≥ 1 ms). Subtraction of ihMTR images obtained with Δt = 0.0 ms and Δt = 0.8 ms generated a new ihMT band-pass T 1D -filter isolating short-T 1D components in the 100-µs to 1-ms range. Simulated ihMTR values in central nervous system tissues were confirmed experimentally. Conclusion:Long-and short-T 1D components were successfully isolated with high RF power and low duty-cycle ihMT filters in the healthy mouse brain.Future studies should investigate the various T 1D -range microstructural correlations in in vivo tissues.

No abstract