clinicians and scientists interested in macromolecules in healthy and diseased brain tissueMagnetic Resonance Spectroscopic Imaging (MRSI) at short TE and short TR allows enhanced detection of macromolecules (MM), particularly at high fields (i.e. 7T). Using either high spatial resolution [1] or a crusher coil [2], these MM signals can be obtained at short TR and TE in the human brain without contamination of lipid signals. In this study we investigate the regional, spectral and relaxation differences of macromolecular spins in the human brain using T1 mapping, metabolite nulling and high resolution MRSI with a crusher coil at 7T.

Data Acquisition:

High resolution MRSI data of the brain was acquired at 7 Tesla using a 32 channel head coil (Nova) and a crusher coil (MR Coils) inserted into the head coil as preciously described [2]. For T1 mapping of MM a single slice high resolution MRSI was obtained using different flip angles in 8 healthy volunteers. The flip angles used were 15°, 30°, 50°, 60°, 70° and 90° and a short TR of 160-190 ms was used. Other measurement parameters were: TE 2.5ms, matrix size 34x38, voxel size 5x5x10mm³, scan time: 2:42min. To create metabolic maps of individual metabolites a different single slice MRSI measurement was performed using the following parameters in a pulse-acquire experiment: TR 820ms, TE 2.5ms , flip angle of 50°, matrix size 17x19, voxel size 10x10x10mm³, acquisition time 3:28min, one average and water suppression was obtained by using frequency selective spokes pulses [3]. Furthermore, metabolite nulled MRSI data was acquired using the same setup and a double inversion recovery experiment and VAPOR water suppression. The following scan parameters were used: TR 1500ms, TE 2.5ms, excitation flip angle 30°, TI1 570ms, TI2 20ms, matrix size 23x23, voxel size 7.8x7.8x 10mm³, 3 averages, acquisition time 19:36min. For inversion an adiabatic full passage pulse was used with duration of 20ms and 95% inversion bandwidth of 1.2kHz, centered at 0ppm. In all experiments a separate non-water-suppressed acquisition at lower resolution and one average was acquired for eddy current correction and residual water removal.

Data processing:

Water unsuppressed data was interpolated to the same matrix size as the water-suppressed data. To subtract the residual water and water sidebands, in every voxel a fit of the water data to the metabolite data was subtracted from the metabolite data. Eddy current correction was performed and the data was saved for fitting purposes. In the data of the high resolution flip angle series, voxels from white matter and from gray matter were fitted with a linear combination of model signals using NMRWizard [R. de Graaf, MRRC, Yale University]. The macromolecules were fitted as one profile. The signal intensity of the MM versus the true local flip angle as assessed from a B1 map was used to obtain an estimate of the T1 relaxation time of macromolecules in white matter (WM) and gray matter (GM) (Figure 1). To create difference maps of the signal intensity of individual metabolites a WM ROI was averaged and subtracted from all individual spectra. The residual was used to create macromolecule and metabolite maps. The process is explained in Figure 2.Voxels from mainly gray and mainly white matter in the metabolite nulled data were averaged to obtain a measured MM profile at short TE and long TR (Figure 4b).

When fitting a full MM baseline to the multi angle experiments, a difference in macromolecule T1 (255 ± 40ms vs. 386 ± 82ms) between GM and WM could be observed (Figure 1). Maps created using the method shown in Figure 2 provide an insight into which macromolecules show a contrast between WM and GM (Figure 3). As can be seen from Figure 4, M1 M2, M3 and M4 show differences between GM and WM, which particularly for M4 is even more clear in the metabolite nulled spectra. Both Figure 4a and 4b show that we observe differences in MM between WM and GM, note however that the spectra in Figure 4a were obtained at a shorter TR than in Figure 4b (850ms compared to 1500ms). When combining all MM signals the apparent T1 as assessed from multi angle experiments is different. However, as observed in Figure 4, the individual MM signals seem to have different T1 contributions as well, similar as recently observed in animals at 17.2T [4].Using high sensitivity measurements, we could clearly observe differences in macromolecular signal between GM and WM. Some of these differences can be explained by differences in T1 relaxation of specific macromolecules.