 |
 |
Hi, To answer your other specific questions: > - Does it mean that FSL uses for eddy current correction a slice by slice correction (not a whole brain global affine FSL (i.e. eddy_correct) does _not_ do slice by slice correction. It _does_ use a whole brain global affine correction. Essentially it just runs FLIRT with 12 DOF, so referencing FLIRT is also worth doing. > - Mangin’s paper describes the Geman-McLure M-estimator algorithm for the diffusion tensor estimation. Does FSL use this algorithm? Within the eddy current correction there is no diffusion tensor estimation. It does weighted linear least squares inside dtifit. All the best, Mark