Displaying the Results
Dealing with dynamic range problem
in Siemens 3D anatomic scans
3D T1-weighted anatomic scans
acquired on the Siemens scanner tend to suffer from a problem with their
dynamic range. This occurs because artifacts inherent to the MRI process
appear as a cluster of extremely bright voxels at the edge of the field
of view. These generally lie outside the head, and do not obscure the anatomy.
However, they are often five times as bright as brain tissue voxels, and
cause some visualization programs, such as register, to allocate too few
gray levels to the range of intensities arising within the brain. You can
get around this by using `mincmath' to clamp the intensity levels in the
volume. After looking at the image, you should be able to spot the artifact
and determine it's intensity. The constant value that you can use
to clamp the intensity, should be chosen in such a way that the bright
spot is cut out, but the values inside the brain are not clamped. For example
if you chose your constant value to be 300
( this is just an example,
each image has to be checked separately) you will say:
mincmath -clamp -const2
0 300 smith_john_19971024_1_2_mri.mnc.gz smith_john_19971024_1_2_mri_clamped.mnc
Initial inspection of the activation
image
Assuming that your experiment
consisted in 120 time frames and the standard gradient-echo EPI protocol
was used. To take a look at the activation maps that you just obtained,
it is suggested that you follow the next steps:
Load the anatomic and activation
images in `register,'
e.g.:
register smith_john_19971024_1_105235_mri_MC_ tstat.mnc smith_john_19971024_1_2_mri.mnc.gz &
or, if you would like to look at the images in Talairach space, you can load in register the tmaps and the anatomical file, after you resampled
both in Talairach space, using
the corresponding transformation.
set register to `synced'
make sure that activation
map is shown in `spectral' colormap
set upper window level for
activation map to +10
set lower window level for
activation map to -10
This will result in regions
of zero activation appearing green, with positive, significant, responses
showing up red to white. Motion artifacts or spurious foci will generally
be easy to spot on maps displayed in this way. It is worth making a habit
of looking at all images in this standard way. Note that the window levels
given here are specific to 120 frame acquisitions.
You can then proceed to apply
the threshold computed in MATLAB as described above.
Summary of analysis steps (example
only for one subject, one dynamic run)
mkdir /data/fmri/analysis/yourname
mkdir /data/fmri/analysis/yourname/john_smith
mkdir /data/fmri/analysis/yourname/john_smith/dynamic
cd /data/fmri/analysis/yourname/john_smith
cp ~fmrixfer/images/smith_john_19971024_1_2_mri.mnc.gz
.
cp ~fmrixfer/images/smith_john_19971024_1_??????_mri.mnc.gz
./dynamic
cd /data/fmri/analysis/yourname/john_smith/dynamic
fmr_preprocess -fwhm 6 -target
3 smith_john_19971024_1_105235_mri.mnc.gz
start MATLAB
frametimes = [0 3 6 9 12 15 18 21 24 27 30 33 ... 351 354 357];
slicetimes = [ 0.140 0.240 0.340 0.440 ..... 2.540 ];
events =[
| 1 | 9 | 9 | 1 |
| 2 | 27 | 9 | 1 |
| 1 | 45 | 9 | 1 |
| 2 | 63 | 9 | 1 |
| 1 | 81 | 9 | 1 |
| 2 | 99 | 9 | 1 |
| 1 | 117 | 9 | 1 |
| 2 | 135 | 9 | 1 |
| 1 | 153 | 9 | 1 |
| 2 | 171 | 9 | 1 |
| 1 | 189 | 9 | 1 |
| 2 | 207 | 9 | 1 |
| 1 | 225 | 9 | 1 |
| 2 | 243 | 9 | 1 |
| 1 | 261 | 9 | 1 |
| 2 | 279 | 9 | 1 |
| 1 | 297 | 9 | 1 |
| 2 | 315 | 9 | 1 |
| 1 | 333 | 9 | 1 |
| 2 | 351 | 9 | 1 | ]; |
S = [ ];
exclude = [1 2 3];
hrf_parameters = [5.4 5.2 10.8 7.35 0.35 0];
X_cache = fmridesign (frametimes, slicetimes, events, S, exclude, hrf_parameters);
input_file = 'smith_john_19971024_1_105235_mri_MC.mnc';
output_file_base =['smith_john_19971024_1_105235_mri_MC_hmw';
'smith_john_19971024_1_105235_mri_MC_hpw'];
| contrast = [ | 1 | -1 | 0 | 0 | 0 | 0 ; |
| 1 | 1 | 0 | 0 | 0 | 0 | ] ; |
exclude = [1 2 3];
fwhm_rho = 15;
n_poly = 3;
[df, p] = fmrilm( input_file, output_file_base, X_cache, contrast, exclude, which_stats, fwhm_rho, n_poly);
quit MATLAB
cd /data/fmri/analysis/yourname/
register smith_john_19971024_1_105235_mri_MC_ tstat.mnc smith_john_19971024_1_2_mri.mnc.gz &
For info regarding this web page, please contact:
-
E-mail:
-
<vali@bic.mni.mcgill.ca>
-
Office phone:
-
(514) 398-8547
-
Office fax:
-
(514) 398-2975
The information on statistical analysis, in this web page was provided by professor Keith Worsley and Chuanhong Liao. For more information regarding statistical analysis, please visit professor's Keith Worsley web page.