Introduction

During the last decade, paralleled with hardware/sequence improvement and clinical availability of high field and ultra high field MR scanners¹, brain sodium (²³Na) MRI has regained interest for the neurological community to characterize dysfunction of ionic homeostasis in various diseases². However, the poor sensitivity of the technique has not allowed yet to access one major phenomenon of ionic homeostasis involved in brain function, i.e. the ability to monitor sodium dynamics following cortical activation. Here we propose a dynamic ²³Na MRI approach enabling to non invasively observe the relative sodium signal changes following a simple right hand movement within the different brain compartments.

Methods

Multi-echoes 3D- radial density adapted ²³Na-MRI ⁴ was acquired in 12 right handed healthy controls on a Magnetom 7T MR scanner (Siemens, Erlangen, Germany) using a dual tuned QED ¹H/²³Na birdcage coil with the following acquisition parameters (TE_1/2/3=0.2ms/10ms/19ms; TR=120ms, 10080 spokes, isotropic voxel = (3mm)³, BW 280 Hz/pixel, density adapted (t0 = 500 us), TA=20min) (Figure 1). Successive spoke angles were incremented by the golden angle and interleaved by a factor of 42 to give a smooth azimuthal variation through k-space and full polar coverage every 25s ⁵ (Figure 2). This allowed to reconstruct temporal series of 42 ²³Na MR volumes. An opposition finger task of the right was performed with 21 alternations of rest/active periods during this dynamic acquisition (one volume acquired per condition every 25s). After denoising (rician filter), spatial realignment and normalization, smoothing (FWHM=9mm), 3D volumes were entered into a GLM to detect signal changes during right hand movement for each TE. Second level group analyses were done for each individual TE, and also a within subject ANOVA was performed to account for all TEs of individual subjects (SPM12, p<0.02, k=5, FDR corrected). BOLD ¹H-fMRI (5 alternations of rest/active periods of 20 measurements per condition) using multiband EPI sequence (TE=20ms, TR=1s, 85 slices, MB=3, grappa 2, isotropic voxel = (1.6mm)³) was used as reference for activation patterns and conventional GLM post-processing was applied onto the 200 measurements (SPM12, p<0.02, k=5, FDR corrected).

Results

BOLD activation patterns and variations of ²³Na signals for each TE are displayed on Fig 3 (increases) and Fig 4 (decreases). Significant sodium signal variations were observed within the motor network. The best correspondence of ²³Na signal variations with the BOLD activation pattern during right hand movement was observed for the decrease of ²³Na signal at TE=19ms.

Discussion

Imaging Na⁺ changes dynamically is a breakthrough in brain functional imaging. So far, such information has not been addressed in vivo and non-invasively. We demonstrate here that non-invasive observation of sodium dynamics following cortical activation is feasible in the human brain using ²³Na fMRI. The best patterns of ²³Na signal changes corresponded to a decrease of the long T₂ component sodium, mostly reflecting a decrease in extracellular concentrations in accordance with invasive studies in animal models^6,7.

Conclusion

This study may open a new era in the better understanding of the normal activity-dependent metabolic coupling in the neuro-glia-vasculature ensemble³, and decoupling in neurological diseases.

Acknowledgements

This work was supported by the following funding sources: Investissements d’Avenir 7T-AMI-ANR-11-EQPX-0001, A*MIDEX-EI-13-07-130115-08.38-7T-AMISTART. Aix-Marseille Université, AP-HM and CNRS (Centre National de la Recherche Scientifique).

References

1. Shah NJ et al. NMR Biomed. 29(2):162-74, 2016. 2. Thulborn KR Neuroimage S1053-8119(16)30674-7, 2016. 3. Jolivet R et al. PLoS Comput Biol. 26;11(2):e1004036, 2015 4. Nagel A et al. Magn Reson Med. 62(6):1565-73, 2009 5. Bydder M et al. ISMRM #4312, 2017 6. Dietzel et al Exp Brain Res 46(1):73-84. 1982. 7. Haack et al. J Vis Exp 5; (103). 2015.