Novel MR imaging technologies

The study of brain disease with MRI requires images of the highest quality. However head motion, or physiological effects such as cardiac pulsation, lead to degradation of image quality. We develop technologies that enable the acquisition of high-quality MR images in non-compliant patient populations.

Minimizing the effects of cardiac pulsation on brain MR images

SlicesMovie2.mp4

Changes in brain maps of the MRI parameter R2* across the cardiac cycle. Taken from Raynaud et al., 2023.

The physiological manifestations of cardiac pulsation in the brain include pulsatile blood flow, brain tissue deformation, and cerebrospinal fluid (CSF) pulsation. These physiological mechanisms lead to signal instabilities in MR images of the brain, and to increase noise levels that reduce sensitivity in analyses of disease-related brain change.

We have conducted an extensive characterization of cardiac-induced noise in brain relaxometry data. In particular, we found that cardiac-induced noise accounts for∼35% of the overall variability of R2*/QSM estimates in inferior brain regions (Raynaud et al., 2023). Cardiac-induced noise is therefore a major noise component in brain relaxometry data. However, suitable data acquisition strategies that minimize this noise source remain largely unavailable.

ISME reduces the variability of R2* and QSM estimates by ~25% (A, taken from Raynaud et al., 2025). R2* maps acquired with ISME are visually sharper and allow the delineation of fine anatomical details such as U-fibers, that are not visible with standard acquisition techniques (B).

From this fingerprint of cardiac-induced noise, we have designed acquisition strategies that mitigate cardiac-induced noise in brain relaxometry data. In particular, the ISME strategy enables the acquisition of multi-echo images at different phases of the cardiac cycle (Raynaud et al., 2025). ISME reduces the variability of R2* and QSM estimates by ~25% across repetitions (A). Reduced variability across repetitions indicates improved sensitivity in analyses of disease-related brain change.

ISME also minimizes aliasing artefacts due to pulsatile flow, and blurring of the data due to physiological noise. The resulting R2* maps are visually sharper and allow the delineation of fine anatomical details such as U-fibers, that are not visible with standard acquisition techniques (B).

References

Raynaud Q, Di Domenicantonio G, Yerly J, Dardano T, van Heeswijk R, Lutti A. A characterization of cardiac-induced noise in R2* maps of the brain. Magn. Reson. Med. (2023). DOI: 10.1002/mrm.29853

Raynaud Q, Oliveira R, Corbin N, Balbastre Y, van Heeswijk R, Lutti A. ISME—incoherent sampling of multi-echo data to minimize cardiac-induced noise in brain maps of R2* and magnetic susceptibility. Magn. Reson. Med. (2025).

Raynaud Q, Dardano T, Oliveira R, Di Domenicantonio G, Kober T, Roy CW, van Heeswijk R, Lutti A. K-space sampling strategies to reduce noise induced by cardiac pulsatility in brain maps of R2* and magnetic susceptibility (preprint).

Correction of head motion effects

Motion artefacts may be present in MR images despite the use of motion correction techniques (left). The suspension technique developed in our laboratory allows for the acquisition of high quality MR images in non-compliant patient populations (right). Taken from Castella et al.

Prospective head motion correction during MRI examinations

Head motion during MRI examinations leads to a degradation of image quality. The LREN laboratory is equipped with an optical camera system (Kineticor) that allows the prospective correction of head motion, in real-time during MRI examinations. With this technology, the position of the head captured by the camera is passed on to the MRI scanner which is adjusted in real-time during the scans to account for the motion.

Despite this technology, strong motion artefacts may remain in MR images. In our laboratory, we address this by suspending the acquisition of data during periods of head motion. This approach is time-efficient: the suspension preferentially targets the most sensitive periods of data acquisition, to minimize the resulting prolongation of scan time. This approach is also flexible: the suspension of the acquisition is triggered above a threshold that can be adjusted by the user to accomodate patients' tolerance of scan duration.

Reference

Castella R, Arn L, Dupuis E, Callaghan MF, Draganski B, Lutti A. Controlling motion artefact levels in MR images by suspending data acquisition during periods of head motion. Magn. Reson. Med. (2018) DOI: 10.1002/mrm.27214

In standard analyses, the loss of image quality due to head motion leads to a non-uniform distribution of image noise across datasets ('heteroscedasticity'), undermining the validity of statistical tests (figure A). The QUIQI method developed by our group restores homoscedasticity. With QUIQI, the strongest improvements lie in frontal regions consistently with the supine position of patients during MRI examination (figure B). Taken from Lutti et al. 2022

Retrospective correction of motion degradation in analyses of MRI data

The prospective correction technology introduced above is only available in a few research centres worldwide: most MRI examinations are conducted without motion correction technology. As a result, the data analysed in neuroscience studies likely exhibit a loss of quality due to head motion. In standard analyses, this results in a non-uniform distribution of image noise across datasets ('heteroscedasticity') which undermines the validity of statistical tests (figure A).

To address this issue, our group has developed a method that accounts for the loss of image quality in the analysis of the data. This technique, called QUIQI for ‘analysis of QUantitative Imaging data using a Quality Index’, restores the homoscedasticity assumption of statistical tests (figure A). This method performs optimally in all brain regions, although head motion effects are local. With QUIQI, the strongest improvements lie in frontal regions (figure B), consistently with the supine position of patients during MRI examination, with the back of the head resting on the scanner table.

Reference

Lutti A, Corbin N, Ashburner J, Ziegler G, Draganski B, Phillips C, Kherif F, Callaghan MF, Di Domenicantonio G. Restoring statistical validity in group analyses of motion-corrupted MRI data. Human Brain Mapping (2022). DOI: 10.1002/hbm.25767

Corbin N, Oliveira R, Raynaud R, Di Domenicantonio G, Draganski B, Kherif F, Callaghan MF, Lutti A. Statistical analyses of motion-corrupted MRI relaxometry data computed from multiple scans. Journal of Neuroscience Methods (2023). DOI: 10.1016/j.jneumeth.2023.109950

Google Sites

Report abuse