Some questions about getting the multi-shell diffusion in scan

[Wang Derek]

Hi experts,

I have some questions about multi-shell diffusion image scanning.

1.
In our pilot,  I scanned the diffusion image with b1000, 2000, and 3000 separately (ie. I get three sets of single-shell bvals and bvecs documents, which correspond to b1000, 2000, and 3000 separately. However, I have seen scans from other studies that randomly mix different b-values in the same sequence (e.g., 5 1000 1995 3005 995 2995 2005.....).
I wonder, for getting the multi-shell data, is it necessary to set different b values in one sequence and include multiple b=0 s/mm2 scans interspersed throughout (please see the attached example data-multishell.bvals)? How to set up for this kind of scanning? If both scanning ways are possible, which one might be better?

2.
In addition, if we separate the different b values and scan them independently as a single shell scan, can we further combine them and do a multi-shell analysis later?

3. We already have the single shell (b=1000) scan for our previous study and want to combine the future scanning data with our existing data. To combine the new data with our existing data, do we need to scan b=1000 separately if different b values must be scanned in one sequence for multi-shell analysis? In the following analysis, can we extract only those scans in b=1000 in the same scan sequence and combine them with our existing data? (eg. 5 1000 1995 3005 995 2995 2005, 1000...... -> only extracts 1000, 995, 1000....).

4. We will do the scan in the AP and PA phases for distortion correction. To save the scanning time, I wonder if we can only acquire a b=0 s/mm2 reference scan with the opposite phase-encoding direction (posterior to anterior), not the whole scan, to correct phase-encoding direction-induced distortions.

I have been struggling with these problems for a long time, and I really appreciate any help any expert can provide.