Synopsis

Keywords: Relaxometry, White Matter, T1 Anisotropy WM

In general, a broad consensus exists that dipolar couplings between water protons and motional restricted macromolecules are prominent modulators of longitudinal relaxation in WM. Recent studies demonstrate anisotropic relaxation in human white matter (WM) in dependence of the fibre-to-field orientation. Our approach for further investigations of such effects overcomes previous limitations, in particular, sample reorientations for a direct demonstration of orientation effects, which were missing in earlier studies. Samples of ex-vivo WM tissue of pig brain yielded slightly reduced spin-lattice relaxation times by 1-2% around an angle of 40° between the main fibre direction and B0.

Introduction

Recent research has demonstrated an anisotropy of longitudinal relaxation times in human brain in dependence of the fibre-to-field angle (θ_FB), especially in white matter tracts with a high degree of structural order^1,2,3. Noticeably, some studies showed a broad maximum of T1 (up to 7% above the mean) around a position of θ_FB≈50°, which might be field-dependent¹, others proposed a decrease in T1 with increasing angle². This discrepancy between previous in vivo results calls for further research that (i) validates orientation-dependent T1 changes in white matter with a well-defined experimental strategy and (ii) yields further information about the underlying mechanism. Besides differences in T1 mapping protocols^4,5 restrictions to vary the head orientation in human MRI need to be considered. Consequently, conclusions about orientation effects have been based on comparisons of different fibre tracts rather than reorienting the sample in the magnetic field.

Methods

Sample preparation: Fresh pig brains were obtained from a local slaughterhouse with a short post-mortem delay (<6h) and immersed in cold phosphate-buffered saline solution (PBS, pH 7.4). Small cylindrical pieces (<25mm) of brainstem were carefully dissected and placed in 5mm NMR tubes filled with Fomblin.

MR experiments: Data were acquired on a clinical 3T scanner (MAGNETOM Skyrafit) system at room temperature and at 36°C with a custom-made TxRx Helmholtz coil (16mm loop radius and spacing; Figure 1). It permits free adjustment of the sample orientation without changing the coil properties (e.g., tuning/matching, SNR, B1 homogeneity, radiation damping). All acquisitions were performed with a 1D readout gradient (0.28mm nominal resolution) utilizing three schemes to prepare the spin system:

(i) A 5ms (soft) adiabatic 180° pulse (BIR-4) for inversion-recovery experiments (Figure 1B).
(ii) Two adiabatic 90° BIR-4 pulses for saturation-recovery experiments.
(iii) A composite, phase-modulated magnetization-transfer (MT) pulse for transient MT experiments as proposed by van Gelderen⁶ (Figure 1C). This pulse consisted of 16 subpulses (nominal flip angles of 60°, -120°, 120°, -120° … 120°,-60°) and had a total duration of 6ms. It is assumed that the MT pulse effectively saturates the macromolecular protons and has a minor effect on the hydrogen protons of water⁶.

Overall, 53 closely sampled evolution times between 10ms and 5s were recorded for all relaxation experiments (i-iii) with sufficient time (TR = 9s) to ensure equilibrium magnetisation before each measurement.

Reconstruction of the diffusion tensor for average fibre direction estimation was based on 1D Stejskal-Tanner measurements with 60 diffusion-sensitising gradient directions (b=1500 s/mm²) . In between, seven data sets were included without diffusion weighting gradients.

Results and Discussion

Fractional Anisotropy (FA): In common with previous observations⁷, FA values were lower in all pig-brain samples compared to highly-ordered human brain regions. The sample selected for extended experiments showed two largely homogeneous regions (~6mm long; Figure 2) with average FA values of 0.38 and 0.39). While this indicates larger dispersions of fibre orientations compared to selected human results, detectable changes in T1 have been proposed for FA > 0.3 in previous work³.

Orientation-dependent T1 times: A week orientation dependence of the ‘observed relaxation time’ was robustly obtained in all measurements (Figure 3) indicating a T1 minimum around 40°, which was also replicated when restricting the analysis to inversion times >300ms. Simulations assuming a two-pool model^8,9 with magnetization exchange between free water and motional restricted macromolecules suggest minimal contributions from the (faster) intrinsic macromolecular T1 in this range. Further analyses comparing one- and two-component fits of the IR experiment support the assumption of non-mono-exponential recovery (Figure 4), in line with earlier work indicating that multiple proton pools are relevant also in T1 analyses^5,10,11. The position of the T1 minimum agrees roughly with the results of Kauppinen et al.¹, who, however, reported a T1 maximum. Of note, a maximum of the apparent transverse relaxation time of the macromolecular pool was observed in previous MT experiments and explained by a distinct dipolar absorption lineshape of myelinated fibre bundles. Given that T1 and MT experiments are different expressions of the same cross-relaxation effects, we may assume a common mechanism underlying their orientation dependences. A detailed analysis of these effects requires more complex models (e.g., two or four spin pools) and acquisition strategies, leading to a number of parameters that cannot be easily fitted to the current data. An entirely model-independent verification of orientation dependence is presented in Figure 5, showing an increasing signal offset for single fibre orientations from the acquisition with parallel fibre alignment taken as a reference.

Limitations: While a consistent orientation dependence was observed in all measurements, additional experiments in a sample with lower fibre orientation dispersion (i.e., higher FA) is advocated for presumably increased effects. Further temperature-dependent measurements might also be useful; initial results indicated approx. 10% longer T1 at 36 °C compared to room temperature.

Conclusion

A strategy was implemented to investigate orientation dependence of spin relaxation in fresh or fixed tissue samples on a clinical scanner. It allows precise reorientation of the sample and minimizes confounding effects, such as B1 inhomogeneity or radiation damping.The obtained high precision achieves robust identification of subtle orientation-dependent T1 differences (order of 1-2% at room temperature) with a local T1 minimum around 40°, which are even evident in samples of only moderate fibre alignment.

Acknowledgements

No acknowledgement found.