Robustness of diffusion scalar metrics when estimated with Generalized Q-Sampling Imaging acquisition schemes

Introduction: Generalized Q-Sampling Imaging (GQI) has recently been introduced by Yeh and colleagues [1], and was shown to have comparable accuracy to other well established q-space methods when it comes to resolving crossing fibres. In addition, this is achievable with as little as 102 points on a grid sampling scheme, bringing the total acquisition time down to a clinically acceptable level. Another advantage of GQI is that it is also applicable to a shell sampling scheme. Despite their successes in tractography applications, q-space techniques have until now failed to produce scalar metrics that could replace the ones derived from the diffusion tensor model (e.g. mean diffusivity, MD, and fractional anisotropy, FA) in terms of their multi-subject comparability and specificity to pathology. The data acquired with a grid sampling scheme can still be used to estimate a diffusion tensor and respective scalar parameters, but the effects of the high b-values required for q-space imaging (>2000 s/mm) in the accuracy of the resulting DTI-based parameters has not been well characterized. The authors of GQI have also proposed a new scalar metric called quantitative anisotropy (QA), but its properties have not been compared to FA’s. In this study we will compare the estimated values of MD, FA and QA0 (first component of QA) obtained with grid and shell sampling schemes, in terms of their precision and ability to differentiate between different brain fibre populations.