The gradient direction files in here are specific to the 
Hedehus/Bammer DTI sequence used at the Stanford Lucas Center.

Random notes about specific grads files: 

* dwepi.65.grads is derived from the FSL bvec file. 
I just reformatted it to Roland's format and rounded to 
four decimal places. It has 60 directions + 5 b=0.

* dwepi.25.grads is from Roland. I think he got these 
directions from Derek Jones. It has 23 dirs + 2 b=0.

* dwepi.44.grads is a old file that I made using a crude
hack. Note that the directions might not be optimally
distributed.


HARDI notes

White and Dale (HBM 2009) suggest that the optimal b-values
for spherical deconvolution analysis of GE 1.5T Signa Hdx
scanner data are b=2200 for L=4, b=3600 for L=6, and b=5200
for L=8.


