Velocity Spectrum Imaging using velocity encoding preparation pulses

Abstract: The movement of water in the human body is a very complex phenomenon encompassing both diffusion and convection. These different physical principles coexist within the same voxel with dominant sub-mechanisms within different intra-voxel structures. While diffusion imaging techniques can separate diffusion populations, convective flow imaging techniques usually measure an average velocity over the voxel. In this article, we present, test and implement a new technique to measure the velocity distribution of water inside each voxel of an MR image. This approach is completely non-invasive and requires no contrast agents. Modified velocity-selective RF pulses can be used to encode velocity information analogously to k-space encoding. The velocity distribution can then be decoded via the Fourier transform. This approach yields a three-dimensional velocity vector distribution of convective flow in each voxel, analogously to Diffusion Tensor Imaging. We present the theoretical principles of this technique, and demonstrates its use on a simple flow phantom with known flow characteristics. We also demonstrate the technique on human participants used to collect the velocity distribution along the three laboratory axes, and discusses its challenges and potential applications. In addition to a useful tool for validating computational fluid dynamic models in vivo, velocity spectrum imaging can be a powerful tool to study the complex movement of water in the glymphatic system and its involvement in neurodegenerative disorders.