Cardiac and Respiratory induced 3D Brain Tissue Strain as Marker of Physiological Blood Volume Dynamics at 7T MRI.
Jacob Jan Sloots1, Peter R. Luijten1, Geert Jan Biessels2, and Jaco J. M. Zwanenburg1

1Radiology, University Medical Center Utrecht, Utrecht, Netherlands, 2Neurology, University Medical Center Utrecht, Utrecht, Netherlands


Brain tissue deformation induced by the cardiac and respiration cycles could be a valuable source of information on the physiology of the brains tissue properties. In this work, we assess the tissue deformation by computing the tissue strain from DENSE displacement data sets and unravel cardiac and respiratory contributions by using a linear model. We observed consistent trends in the three strain components due to cardiac and respiration cycles, which agree with blood volume changes. In contrast to tissue displacement, the tissue strain may serve as a reliable novel marker of physiological blood volume dynamics in the brain.


Brain tissue deformation induced by the cardiac and respiration cycles could be a valuable source of information on the brains tissue properties such as drainage of cerebral waste1. We have shown that displacement encoding with stimulated echoes (DENSE) allows for separate assessment of cardiac and respiration induced motion2. However, these displacement maps include rotations and translations that are not considered to drive clearance. Thus, this work aims to assess tissue deformation by computing the tissue strain from DENSE displacement maps. We acquired 2D slices in both the sagittal and coronal plane to assess the strain in 3 directions, in two small regions of interest (ROIs) along their intersection line.


A cardiac triggered, single-shot 2D DENSE sequence2 was designed to measure heart beat and respiration related brain tissue motion (Figure 1, 3mm isotropic, EPI: 33, SENSE 2.5). A key feature included a time-delay of one cardiac cycle between encoding and decoding to make the sequence more sensitive for respiration induced motion contributions. Simulations of the DENSE signal showed optimal settings for Tenc=0.3mm/2π and αmax=50° (white matter: T1=1200ms and ADC=0.8·10-3mm2/s at 7T3,4). Variable excitation flip angles were used to obtain a stable signal to noise ratio over the cardiac phases5.

Written informed consent was obtained from all volunteers in accordance with the Ethical Review Board of our institution. Nine healthy subjects (3 females, age 29±3 years) were included and scanned at 7T (Philips Healthcare) using a 32-channel head coil (Nova Medical). The DENSE measurements were repeated over 100 dynamics with alternating encoding direction for every other dynamic. To allow for corrections of potential eddy current effects, a reference phase was acquired directly after encoding (Figure 1). The protocol resulted in 400 snapshots over time, with a scan duration of 2.5-5 min (80-40bpm).

A sagittal and perpendicular coronal slice were planned, and displacement maps were separately acquired with two in-plane encoding directions (Figure 2), resulting in four displacement-encoded time series. Physiological data was simultaneously recorded by using a vector cardiogram (VCG) for triggering and a respiration belt to trace abdominal breathing. Repeatability of the measurements was assessed by acquiring an additional sagittal slice with Feet-to-Head motion encoding after repositioning of the volunteer in the scanner.

Strain maps were computed from the acquired displacement maps, where phase derivatives larger than π were unwrapped. Cardiac and respiration induced strain were unraveled by using a linear model based on the physiological data, as was proposed previously for motion measurements2.


Imaging was successful in all subjects, and triggering was consistent; on average 96 dynamics could be used for analysis. Due to time constraints, we did not obtain repeated measurements in subject 7.

Regional analysis was performed in two regions of interest (Figure 2) for both cardiac (at peak systole; 25% of cardiac trigger interval) and respiratory (full inspiration; 95% confidence range of respiration trace) induced strain. Because of the noisy strain signal, the median in the ROIs were assessed, which makes the analysis less prone to outliers. Cardiac and respiration related strains were largest in the deep brain (basal ganglia). Here, we observed cardiac induced tissue expansion with median volumetric strain of 0.86·10-3 (IQR: 0.31·10-3 to 1.1·10-3). For inspiration, a slight compression was observed with median -0.32·10-3 (IQR: 0.03·10-3 to -0.41·10-3). The three 1D strain components showed anisotropy with stretch in the Feet-to-Head direction and minor shrinkage in the other two directions (Figure 4). The analysis of the repeatability of the measurements is shown in Figure 5.


The observed cardiac induced tissue expansion is most likely due to arterial blood flow to the brain, whereas the respiration induced compression in tissue strain is consistent with outflow of the venous blood due to lower thoracic pressure. Although subject 2 appears as an outlier in Figure 3, the strain in different directions is consistent with the other subjects (Figure 4). Summing over the three different strain components to a volumetric strain increases noise sensitivity, as they have different signs. For the centrum semiovale, we measured strain values comparable with values reported by Hirsch et al6. We found larger strain values in the basal ganglia, which may reflect regional heterogeneity in tissue volumetric strain (and blood volume). We were able to evaluate very small ROIs with fair reproducibility. Reproducibility can be improved by acquiring more dynamics or by averaging over larger ROIs.


We have successfully measured local cardiac and respiratory induced strain in small ROIs, despite the amplified noise induced by the use of spatial derivatives. The strain measurements are consistent with blood volume change and may serve as a novel physiological marker on blood volume dynamics in the brain.


The research leading to these results has received funding from the European Research Council under the European Union's Seventh FrameworkProgramme (FP7/2007-2013) / ERC grant agreement n°337333.


1. Mestre H, Kostrikov S, Nedergaard M, et al. Perivascular spaces, glymphatic dysfunction, and small vessel disease. Clinical Science. 2017;131(17):2257–2274

2. Sloots J, Adams A, Zwanenburg J, et al. Unraveling cardiac and respiratory contributions to brain tissue motion using single shot 2D DENSE at 7T MRI. ISMRM Proceedings. 2018;2:368

3. Rooney, W. D. et al. Magnetic field and tissue dependencies of human brain longitudinal 1H2O relaxation in vivo. Magnatic Resonance in Medicine. 2007;57(2):308–318

4. Helenius, J. et al. Diffusion-weighted MR imaging in normal human brains in various age groups. American Journal of Neuroradiology. 2002;23(2):194–199

5. Fischer S, McKinnon G, Boesiger P, et al. Improved myocardial tagging contrast. Magnetic Resonance in Medicine. 1993;30(2):191–200

6. Hirsch J, Klatt D, Sack I, et al. In Vivo Measurement of Volumetric Strain in the Human Brain Induced by Arterial Pulsation and Harmonic Waves. Magnetic Resonance in Medicine. 2013;70(3):671-683


Schematic representation of the DENSE sequence, covering two dynamics and four cardiac cycles. In the first cardiac cycle of each dynamic, an encoding gradient G­ is placed in the measurement direction (example: Feet-to-Head for sagittal slice). The actual cardiac phases are acquired during the second cardiac cycle to increase the sensitivity to respiratory motion. Decoding gradients are applied before each readout (green gradients). G is inverted between dynamics (dashed lines), to distinguish between phase errors and motion. A reference phase is acquired after encoding to correct for potential eddy current effects from the decoding gradients, which alternate with the encoding direction, thus mimicking motion.

Slice planning and analyzed ROIs of the 2D DENSE scans with respect to the whole-brain 3D T1-wighted TFE planning scan. a) The coronal slice was planned through the brain stem (angle varies between 12-15 degrees, subject dependent). The sagittal slice was planned perpendicular to the coronal slice, 10mm from the center of the brain. b,e) The obtained 1D strain maps for the slices, both in the Feet-to-Head direction. c,f) The associated DENSE magnitude images. The red line and boxes in Figure b-c,e-f indicate the intersection line and 2 ROIs (3x3 voxels) that were manually selected for strain evaluation, respectively.

Volumetric strain analysis for nine subjects. The volumetric strain was obtained by summing the 1D strains from all 3 encoding directions. For each 1D strain direction, the median over the associated ROI indicated in Figure 2 was used. (For the Feet-to-Head direction, the median of the combined ROIs of the coronal and sagittal slice was used). The left two boxes represent the results in the basal ganglia (BG, green ROI in Figure 2), whereas the right two boxes represent the results in the centrum semiovale (CSO, purple ROI in Figure 2). In each region, both the cardiac and respiratory related contributions are represented.

1D strain (median of ROI) shown for each direction in the basal ganglia (BG) and centrum semiovale (CSO). Figure a and b show the cardiac and respiration contributions, respectively. Please note the difference in scale along the y-axis. For cardiac induced strain, the Feet-to-Head direction shows relatively large expansion, whereas the other two directions show some compression. The effect is most pronounced in the BG, yet is of the same trend in the CSO. Interestingly, the respiration induced strain shows a similar but opposite trend, consistent with outflow of (venous) blood with inspiration, due to lower thoracic pressure.

Repeated measurement correlations (a,c) and Bland-Altman plots (b,d) for cardiac and respiratory induced 1D strain measurements in the Feat-to-Head direction in the sagittal slice. Results for the two ROIs are shown on top of each other. The reproducibility is best in the BG for the cardiac induced strain (Figure a and b). Measured strain values are lower in the CSO and, consequently, the noise is more dominant in this ROI. Similarly, respiration induced strain contributions are smaller compared to cardiac induced strain, resulting in reduced reproducibility. CoR (%): absolute coefficient of repeatability (and percentage relative to the average of all individuals).

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)