White matter compartment models for in vivo diffusion MRI at 300 mT/m

摘要: This paper compares a range of compartment models for diffusion MRI data on in vivo human acquisitions from standard 60mT/m system (Philips 3T Achieva) and unique 300mT/m (Siemens Connectom). The key aim is to determine whether both systems support broadly the same or Connectom higher gradient supports significantly more complex models. A single volunteer underwent 8h acquisition each provide uniquely wide dense sampling available space pulsed-gradient spin-echo (PGSE) measurements. We select set promising possible three-compartment white matter (WM) that previous work preliminary experiments suggest as strong candidates, but extend them fit compartmental T2 diffusivity. focus corpus callosum where WM fibre architecture simplest compare their ability explain measured data, using Akaike's information criterion (AIC), predict unseen cross-validation. also stability parameter estimates presence i) noise, bootstrapping, ii) spatial variation, visual assessment comparison with anatomical knowledge. Broadly similar emerge AIC cross-validation sets. Specifically, model consisting either Bingham distribution sticks Cylinder intracellular compartment, an anisotropic tensor (DT) extracellular well isotropic CSF performs consistently well. However, various other perform no emerges clear winner. (with virtually contamination) do not partially Evaluation favours simpler than those identified by They level complexity underpinning currently popular microstructure imaging techniques such NODDI, CHARMED, ActiveAx, number free parameters about 4 5 rather 10 11, may reflect achievable useful technique current systems, although