I thought it worth starting a new reply for this:

I took a smaller subset and methodically edited the different values. It appears column 4, roll, value appeared to be the culprit. Below are values from a dfile_VolReg2_aff12.1D file:

-0.0288 -0.124 -0.005 0.8544 0.0083 -0.0323 0 0 0 0 0 0
-0.0254 -0.1439 -0.002 0.8848 0.0439 -0.0921 0 0 0 0 0 0
0.1401 -0.7076 0.0146 4.8042 0.101 0.1251 0 0 0 0 0 0
0.3368 -1.6307 0.0161 5.428 0.0873 0.2056 0 0 0 0 0 0
0.4403 -1.1964 -0.0157 6.8392 0.094 -0.2051 0 0 0 0 0 0
0.1105 -0.3726 0.0019 1.46 0.0443 -0.0523 0 0 0 0 0 0
0.0628 -0.2221 -0.0054 1.0794 0.0573 -0.1024 0 0 0 0 0 0
0.0428 -0.1325 -0.0043 0.3687 0.0532 -0.0329 0 0 0 0 0 0
0.0241 -0.0445 -0.0008 0.272 0.0817 -0.047 0 0 0 0 0 0
-0.0033 -0.0147 0.0126 0.0418 -0.0239 -0.0296 0 0 0 0 0 0
0.1256 -0.4107 -0.0024 2.4248 0.1107 0.0819 0 0 0 0 0 0

I thought it was strange that each of the roll values are typically an order or two magnitude higher than the pitch or yaw. EDITED* I looked more closely through the data set and perhaps actual data/image volume does move more so in the roll direction. BUT it also seems to apply a translation! Or something weird.

Anyway, IF I use the dfile_VolReg2_aff12.1D file, BUT make all of column 4, roll values, zero, it looks MUCH more near identical to the registered volume using 3dvolreg. Could it be a order of operation, whether translation or rotation gets applied first??

QUESTION: Is there a way to make only this column zero in AFNI? Currently can only do in matlab...



Edited 2 time(s). Last edit at 03/02/2023 04:55PM by ebaldelo.