How Many Gradients are Sufficient in High-Angular Resolution Diffusion Imaging (HARDI)?

Introduction: High-angular resolution diffusion imaging (HARDI) is a powerful extension of MRI that maps the directional diffusion of water in the brain, based on applying diffusion-sensitized gradients in 30-100 or more different directions. In the Diffusion Tensor (DT) model, a tensor is fitted to the set of diffusion images, and maps of fiber integrity such as fractional or relative anisotropy (FA, RA), mean diffusivity (MD) or volume ratio (VR), may be computed from the tensors’ eigenvalues. Given the trade-off between the available signal-to-noise (SNR) and the time required to collect more gradient images, some studies argue that 20 gradient directions are sufficient to accurately compute FA [1], while tractography applications require more gradients to accurately resolve fiber directions and crossings. Here we computed how the SNR increases, depending on the number of gradients acquired, for several widely used DTI measures, to evaluate the trade-off between image quality and acquisition time.