Synopsis

Biophysical modelling of diffusion-weighted MRI (DW-MRI) data can help to gain more insight into brain microstructure. However, models need to be validated.

This work validates a recently-developed technique for non-invasive mapping of brain cell-body (soma) size/density with DW-MRI, by using ultrahigh-field DW-MRI experiments and histology of mouse brain.

Predictions from numerical simulations are experimentally confirmed and brain’s maps of MR-measured soma size/density are shown to correspond very well with histology. We provide differential contrasts between cell layers that are less expressed in tensor analyses, leading to novel complementary contrasts of the brain tissue. Limitations and future research directions are discussed.

Introduction

Mapping brain microstructure noninvasively using diffusion-weighted MRI (DW-MRI) remains a formidable challenge due to the complexity of the underlying constituents and the relatively featureless diffusion-driven signal decay. Biophysical modelling can help to gain more insight into the microstructure^1-3. However, it is very difficult to choose the correct model⁴, and validation is necessary.

This work validates a recently-developed technique⁵ for non-invasive mapping of brain cell-body (namely soma) size and density with DW-MRI.

Advanced Monte-Carlo simulations, based on a recently-proposed framework⁶, showed that soma size/density may have a specific signature on the direction-averaged DW-MRI signal S_ave at high b-values^5,7. Simulations also suggested that at diffusion times t_d<30 ms, the water exchange between neurites and soma may be neglected, supporting the design of a simple three-compartment model to quantify the presence of soma⁵:

S_ave(b)/S(b=0)=f_sticksS_sticks(b,D_a)+f_sphereS_sphere(b,D_sphere,R_sphere)+(1-f_sticks-f_sphere)S(b,D_iso) [1]

where f_sticks represents the MR cell-fibres signal fraction, D_a axial diffusivity; f_sphere MR soma signal fraction, R_sphere soma radius, D_sphere and D_iso intra-soma and extra-cellular diffusivity.

Here we use ultrahigh field DW-MRI and histology of mouse brain to validate those simulation results⁷ and the proposed compartment model⁵.

Methods

Simulations. Detailed three-dimensional geometries were constructed to mimic realistically connected neurites, in different soma size/volume fraction (r_soma,f_soma) conditions, as proposed in^5-7(Fig.1a). The S_ave(b) signals, as measured by a Pulsed-Gradients-Spin-Echo (PGSE) sequence, were computed in CAMINO⁸ for b-values=[0:1:40] ms/μm², δ/Δ=3/11 ms and direction-average across 40 gradient directions (Fig.1b). Only intracellular water diffusion was simulated. However, the effect of extracellular diffusion and membrane permeability were shown to be negligible at very high b-values⁹ and diffusion times<20 ms¹⁰.

Specimen preparation. All animal experiments were pre-approved by the institutional and national authorities and were carried out according to European Directive 2010/63. Mice (N=2), male, 8 weeks old, were perfused intracardially with 4% paraformaldehyde. The brains were isolated and kept 48h in 4% paraformaldehyde and 5 days in PBS (changed daily), before being transferred to a 10 mm NMR tube filled with Fluorinert (Sigma Aldrich) for susceptibility matching. Upon MRI, tissue was used for microscopy as described in¹¹ to provide histological banchmark.

MRI experiments. All experiments were performed using a 16.4 T MRI scanner (Bruker BioSpin, Karlsruhe, Germany) operating at 700 MHz for ¹H nuclei equipped with a micro5 imaging probe with maximum gradient strength 3000 mT/m isotropically. The brain was kept at constant temperature of 37°C. DW-MRI were acquired using a PGSE sequence with: TE/TR=20/2500 ms; δ/Δ=3/11 ms; b=0,1,2,3,5,8,10,12,16,20,25,40 ms/μm², 40 gradient directions, resolution 50x50x250 μm³, 10 slices, 4 averages.

Data preprocessing and analysis. Data were denoised using MP-PCA¹² and Gibbs ringing corrected¹³. No artifacts from movement or eddy-current were observed. Data at b=3 ms/μm² were used to compute diffusion tensor imaging (DTI) maps in mean diffusivity (MD) and fractional anisotropy (FA), using FSL¹⁴. Relation [1] was voxel-wise fitted to the measured S_ave(b) signal to estimate f_sticks, D_a, f_sphere, R_sphere, D_iso, fixing D_sphere=3 μm²/ms, as previously described⁵.

Results

Simulations vs. experiments. Simulations predicted that, at Δ=11 ms, S(b^-1/2) shows specific curvature in the range 0.2<b^-1/2<0.5 induced by specific (r_soma,f_soma) conditions (Fig.1b). In agreement, the signal decay extracted from ROIs known to have different (r_soma,f_soma) conditions, like corpus-callosum white matter (WM) and cortical gray matter (GM), confirms that these are correct (Fig.1b).

MR-measured features. Fig.2a shows f_sphere, R_sphere and f_sticks maps, for the two mouse brains. Note the exquisite differential contrasts between layers in the olfactory bulb, cortex, hippocampus and cerebellum. These contrasts are less expressed in DTI maps (Fig.2b).

Comparison against histology. A direct comparison of the MR maps with histology of the same brain is shown in Fig.3b and 4. We observe very good correspondence between soma and cell-fibres spatial distributions from histology and f_sphere, R_sphere and f_sticks maps. Quantitative analysis in Fig.5 shows higher correlation of f_sphere with DAPI staining than MD. Comparison against literature-derived histology¹⁵ is also provided in Fig.3a. Overall, the predicted brain’s composition in terms of MR-measured soma size/density corresponds very well with ground-truth histological measurements.

Discussion and Conclusion

Using high resolution DW-MRI data and histology, this work confirms the predictions of numerical simulation^6,7(Fig.1), and validate the source of new contrasts specifically related to soma size/density⁵ (Fig.3-5). These are dissimilar to the simple tensor analyses (Fig.2-5), and therefore represent new complementary information on the brain.

Future work will investigate possible bias on parameters estimates due to cross-correlation and degeneracy¹⁶ as well as multi-compartmental T₁/T₂ effects¹⁷, cell-fibres curvature¹⁸, branching¹⁹ and exchange²⁰. However, the present suggests that these novel contrasts in soma size/density may provide a new set of biomarkers of potential great value for biomedical applications and pure neuroscience.

Acknowledgements

References

1. Novikov D S, Kiselev V G, Jespersen, S N. On modeling. Magn. Reson. Med., 79(6), 3172-3193.

2. Alexander D C, Dyrby T B, Nilsson M, et al. Imaging brain microstructure with diffusion MRI: practicality and applications.NMR Biomed. 2017, e3841.

3. Jelescu I O, Budde M D. Design and validation of diffusion MRI models of white matter.Front Phys 2017, 5, 61.

4. Fieremans E, Lee H H. Physical and numerical phantoms for the validation of brain microstructural MRI: A cookbook. Neuroimage 2018, 182, 39-61.

5. Palombo M, Shemesh N, Ianus A, et al., Abundance of cell bodies can explain the stick model’s failure to describe high b-value diffusion signal in grey matter. Proc. Int. Soc. Magn. Reson. Med. 2018, #1096.

6. Palombo, M., Alexander, D. C., Zhang, H. A generative model of realistic brain cells with application to numerical simulation of diffusion-weighted MR signal. Neuroimage under review 2018, arXiv preprint arXiv:1806.07125.

7. Palombo, M, Shemesh N, Ianus A, et al., A compartment based model for non-invasive cell body imaging by diffusion MRI. Proc. Int. Soc. Magn. Reson. Med. 2018, #0892.

8. Cook P A, Bai Y, Nedjati-Gilani S K K S, et al. Camino: open-source diffusion-MRI reconstruction and processing. In 14th scientific meeting of the international society for magnetic resonance in medicine 2004 (Vol. 2759). Seattle WA, USA.

9. Veraart J, Fieremans E, Novikov D S. On the scaling behavior of water diffusion in human brain white matter. Neuroimage accepted 2018, DOI:10.1016/j.neuroimage.2018.09.075.

10. Yang D M, Huettner J E, Bretthorst G L, et al. Intracellular water preexchange lifetime in neurons and astrocytes. Magn. Reson. Med. 2018, 79(3), 1616-1627.

11. Nunes D, Cruz T L, Jespersen S N, et al. Mapping axonal density and average diameter using non-monotonic time-dependent gradient-echo MRI. J. Magn. Reson. 2017, 277, 117-130.

12. Veraart, J., et al. Denoising of diffusion MRI using random matrix theory. Neuroimage 2016, 142, 394-406.

13. Kellner E, Dhital B, Kiselev V G, et al. Gibbs‐ringing artifact removal based on local subvoxel‐shifts. Magn. Reson. Med. 2016, 76(5), 1574-1581.

14. https://fsl.fmrib.ox.ac.uk/fsl

15. BrainMaps: An Interactive Multiresolution Brain Atlas; http://brainmaps.org

16. Jelescu I O, Veraart J, Fieremans E, et al. Degeneracy in model parameter estimation for multi‐compartmental diffusion in neuronal tissue. NMR Biomed. 2016, 29(1), 33-47.

17. Veraart J, Novikov D S, Fieremans E. TE dependent Diffusion Imaging (TEdDI) distinguishes between compartmental T2 relaxation times. Neuroimage 2018, 182, 360-369.

18. Özarslan E, Yolcu C, Herberthson M, et al. Influence of the size and curvedness of neural projections on the orientationally averaged diffusion MR signal. Frontiers in physics 2018, 6, 17.

19. Palombo M, Ligneul C, Najac C, et al. New paradigm to assess brain cell morphology by diffusion-weighted MR spectroscopy in vivo. PNAS 2016; 113(24): 6671-6676.

20. Veraart J, Fieremans E, Rudrapatna U, et al., Biophysical modeling of the gray matter: does the “stick” model hold? Proc. Int. Soc. Magn. Reson. Med. 2018, #1094.