The brain has pulsatile motion driven by the cardiac cycle. This pulsatility is important for maintaining homeostatic processes within the brain. An increase in pulsatility has been linked to diseases such as dementia, making it potentially useful for gaining insight into these diseases¹. Determining brain pulsatility is difficult due to the time durations required to measure time-resolved low tissue velocities over a large volume. Feasibility of volumetric strain rate mapping² and whole brain strain rate mapping³ was previously shown at 1.5T but not yet at 7T. At 7T, single slice tissue velocity maps showed a potential gain in SNR by a factor of 5 compared to 1.5T⁴. In this study, we demonstrate the feasibility of measuring whole brain volumetric strain with high resolution at 7T MRI in 8 healthy volunteers.Retrospectively gated, 4D phase contrast imaging was done on eight healthy volunteers (5 males, 25.3 ± 5.0 years) with a 7T MR system (Phillips Healthcare) using the sequence parameters indicated in Table 1. A phantom (head shaped, 4-5% agarose gel) was also scanned using the same parameters. For each data set, the mean velocity over the cardiac cycle was subtracted from the velocity measurements. The built-in MATLAB (V2015b MathWorks®) divergence function was used to find the divergence of the velocity field ($$$\triangledown \cdot V = \frac{\partial V_{x}}{\partial x} + \frac{\partial V_{y}}{\partial y} + \frac{\partial V_{z}}{\partial z}$$$), which was computed to determine the strain rate. Strain in the brain over the cardiac cycle was calculated by integrating the strain rate after a 2D-median filter (kernel size: 2x2 pixels) was applied to smooth the data in the transverse plane. Large ROIs (avoiding areas with high flow from CSF or blood) were drawn in 3 slices (low, middle and top of the brain) to determine the mean strain curve. The brain was considered to be fully relaxed when the strain was minimum (end diastole) so the minimum mean strain was set to zero. Two consecutive scans (separated by a 1 minute T1 scan) were also taken in an additional volunteer to compare the spatial heterogeneity of the strain maps.Figure 1 shows a representative example of the magnitude and 3D low-velocity images. Figure 2 shows the brain tissue strain for low, middle and high slices over the cardiac cycle for all volunteers. The corresponding mean maximum strain values over all volunteers and for the phantom are shown in Table 2. The minimum strain for all volunteers and slices was found to occur during 68-82% of the cardiac cycle (relative to the peak in the pulse-oximeter signal). The apparent spatial heterogeneity was not found to be reproducible in two consecutive intra-subject strain measurements, though the strain curves were very similar (Fig. 3).The general shape of the mean brain tissue volumetric strain curves from this investigation show agreement with plots reported in the study of Hirsch et al.² However, our strain values were in general an order of magnitude greater than theirs (2.3 x 10^-3 versus 2.8 x 10^-4). This difference may have resulted from differences in image acquisition and post-processing, but requires further investigation. The minimum for all slices and volunteers occurred typically near the end of the cardiac cycle, suggesting consistent relaxation of the brain during the final stages of diastole. However, the time of the peak strain appeared to be more variable between subjects. Also the maximum strain amplitudes were very variable between subjects, and even between slices of the same volunteer. It is very unlikely that the inherent noise or systematic errors within the MR acquisition could account for these heterogeneous results since the phantom measurements showed no clear spatial or temporal patterns and yielded 10 times lower strain magnitudes than the volunteer measurements. The intra-subject spatial heterogeneity seems not to be reproducible, which suggests that local brain tissue strain may vary quasi-randomly over time. However, further research in more subjects is needed to further characterise the observed intra- and inter-subject heterogeneity.We showed in this study the feasibility of measuring and quantifying volumetric strain at 7T with whole brain coverage and high (2mm isotropic) resolution, without excessive smoothing in the analysis. The method may have potential for studying the physiology of brain pulsatility in healthy conditions and disease.