Thanks for this very useful information. I was able to run a similar script to the one you suggested and corrected in just the PE we collect in. I will add the -interp argument to the next scripts. I have questions about inspecting the images and ensuring the correction procedure did what we were hoping they would do. Do you have general suggestions about how to inspect the data afterwards? What should I look out for, and sanity checks can I perform?
When I inspect the corrected images alone, it is difficult for me to tell what to look out for. I created a difference map between corrected and uncorrected images, and I see that values have changed across the entire brain, and seem to be more extreme around the eye sockets, brainstem, and the cerebellum area which makes sense to me. I am most concerned about correcting distortion around the vmPFC/OFC areas. I also have questions about the set up file b02b0.cnf. Can you tell me generally how the default values are set up? Is it assumed these should suffice for most datasets, and we shouldn't edit unless doing very specific tweaking?
Thanks very much,
# Resolution (knot-spacing) of warps in mm
# Subsampling level (a value of 2 indicates that a 2x2x2 neighbourhood is collapsed to 1 voxel)
# FWHM of gaussian smoothing
# Maximum number of iterations
# Relative weight of regularisation
# If set to 1 lambda is multiplied by the current average squared difference
# Regularisation model
# If set to 1 movements are estimated along with the field
# 0=Levenberg-Marquardt, 1=Scaled Conjugate Gradient
# Quadratic or cubic splines
# Precision for calculation and storage of Hessian
# Linear or spline interpolation
# If set to 1 the images are individually scaled to a common mean intensity