Print

Print
John,

I was just looking into this and found the same thing. Different rp file values with reoriented images.
hdr(1).private.mat
   -3.0000   -0.0000         0   97.5000
   -0.0000    3.0000    0.0000  -96.9129
    0.0000         0    4.5000  -56.4129
         0         0         0    1.0000

 hdr(1).private.mat0
   -3.0000         0         0   97.5000
         0    3.0000         0  -96.9129
         0         0    4.5000  -56.4129
         0         0         0    1.0000

and reoriented file:
hdr2(1).private.mat 
   -3.0000   -0.0000    0.0000   97.5000
    0.0000    3.0000    0.0000  -56.9129
   -0.0000         0    4.5000  -86.4129
         0         0         0    1.0000
hdr2(1).private.mat0
   -3.0000         0         0   97.5000
         0    3.0000         0  -96.9129
         0         0    4.5000  -56.4129
         0         0         0    1.0000

It looks like the .mat fields are slightly different; however, the same for all volumes. The question is thus, why are the realignment parameters different if the same data is used, just different transformation matrices.

Additionally, since the rp files are different, using the motion regressors as covariates would change the models as well.

Are the differences due to interpolation to produce the mean image resulting in slightly different values?


Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Office: (773) 406-2464
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
intended only for the use of the individual or entity named above. If the
reader of the e-mail is not the intended recipient or the employee or agent
responsible for delivering it to the intended recipient, you are hereby
notified that you are in possession of confidential and privileged
information. Any unauthorized use, disclosure, copying or the taking of any
action in reliance on the contents of this information is strictly
prohibited and may be unlawful. If you have received this e-mail
unintentionally, please immediately notify the sender via telephone at (773)
406-2464 or email.



On Fri, Nov 25, 2011 at 3:43 PM, John Ashburner <[log in to unmask]> wrote:
I don't have any ideas here.  I think you really need to take a close
look at the data.  You can read the fields in the headers by:

   P='blah.nii';
   Nii=nifti(P)

If you want to see the actual values, then you can look at a hidden
field of the nifti object:

   Nii.hdr

Computing difference images via ImCalc, and displaying them via
CheckReg can also help to give clues.

Best regards,
-John

On 24 November 2011 10:53, Joshua Balsters <[log in to unmask]> wrote:
> Dear SPMers
>
> I found one previous discussion on this point but not much else, so apologies if I'm repeating what everyone already knows.
>
> I've recently starting using 4D files straight from the scanner (Philips Acheiva) in SPM rather than splitting them into 3D files using fslsplit. I'm finding in most cases there are subtle (+/-0.5mm) to drastic (over 2mm) differences in my realignment parameters depending on whether I use the 4D or 3D versions of the same datasets. These differences are only in the translations and not the rotations.
>
> It would be great if anyone could explain why this happens. I had a quick look for differences between the 3D and 4D files and I think the issue is that fslplit is automatically changing my datatype from int16 to float32, and i read in a previous discussion that it also changes the scaling factor which is probably the bigger issue.
>
> any advice on this issue would be greatly appreciated!
>
> Best
> Josh
>