I expect this is the same issue as on this post? Did you look into whether there were any issues with running out of storage space or anything like that?
Also, I don’t think this has anything to do with dwi2response as such – the data is clearly affected before that step, something is going on within the dwifslpreproc stage. It might be worth editing the title of your post to reflect that.
I suggest you run your dwifslpreproc command with the -debug option and provide the full output here. It would also be good to have a look through the contents of the temporary folder created by the command, which I note you should still have given the use of the -nocleanup option. In particular, the command log, and any other relevant log files you can find, and also have a look at the various intermediate files generated by the command, to see if you can narrow down the particular step where things go wrong.
These numbers look identical to what they were in the first call at the beginning of your post, which is exactly what I’d expect. I’d be worried if that was what the output looked like for your dwi_den_degib.mif file though…?