> 1) Does DICOM Import flip (x-y flip) the images
> while converting the images?
Sometimes, but this should not be important and can not easily be explained
(if you need to know exactly what happens, then see spm_dicom_convert.m).
The important thing is that the "mm space" that SPM and other NIFTI
compatible software report is consistent with the patient coordinate system
of the scanner. Note that these coordinate systems do not have the same
axes...
% DICOM patient co-ordinate system:
% x increases right to left
% y increases anterior to posterior
% z increases inferior to superior
% NIfTI co-ordinate system:
% x increases left to right
% y increases posterior to anterior
% z increases inferior to superior
> 2) After the second level analysis in SPM, I
> obtained the coordinates of activation. I then converted them to talairach
> coordinates. Talairach coordinates are in opposite direction to MNI
> coordinates, I have learnt. Is this true?
It is not true.
>
> 3) I checked each of the Talairach coordinates in MRICro to obtain the
> exact location of areas. For example, if my coordinate is -4,-56,9
> (talairach); should I look for the following coordinate in MRICro:4,-56,9,
> i.e., the opposite x-direction? Is this the best way to check for exact
> location of coordinates? What other methods are used (if any)?
If you are using SPM5, then the headers are all in NIfTI format. MRIcro was
written before NIFTI was specified, so this software is not NIfTI compatible,
and the coordinates it reports will not be based on information from the
headers. You should try MRIcron, which deals properly with these headers...
http://www.sph.sc.edu/comd/rorden/mricron/
See http://www.cma.mgh.harvard.edu/iatr/display.php?spec=nifti-1 for a list of
NIfTI compatible software.
>
> 4) What is the measurement that is used to show activation? Is it the
> T-value in each coordinate? What other measures are used?
When you click on the Results, then SPM will show you t statistics (or F
statistics) above some threshold, where the threshold is typically derived
using Random Field Theory. Many of these questions should be answered in the
SPM5 manual (...\spm5\man\manual.pdf).
>
> 5) Once we obtain coordinates of activation, should we check all the
> coordinates or should we check only those coordinates that have T>a certain
> value. In a paper, for example (Frenck et al., 2005), they have used
> T>5.13.
>
> 6) On what basis should we choose the T-value?
See the examples in the manual, or "14. Introduction to Random Field Theory"
in http://www.fil.ion.ucl.ac.uk/spm/doc/books/hbf2/
Best regards,
-John
|