Displaying the Results

Initial inspection of the activation image

• Load the anatomical and the t-stat magnitude images in register:
  
  register  <t-stat_file>  <anatomical> &
  
  register study_subj001_20031019_111740_3_mri_MC_mag_t.mnc study_subj001_20031019_111740_2_mri.mnc.gz &
  
  or, if you would like to look at the images in Talairach space, you can load in register the t-stat maps and the anatomical file, after you have resampled both in Talairach space, using the appropriate transformation.
  
  
• click the "Not Synced" button to set it to "Synced"
• make sure that activation map is shown in Spectral colormap
• set the upper and lower window levels for your activation map appropriately, keeping in mind the peak threshold reported by stat_threshold

Summary of analysis steps (example only for one subject, one dynamic run)

mkdir /data/scratch/scratch1/yourname/study_subj001

cd /data/scratch/scratch1/yourname/study_subj001

cp /data/transfer/minc/study_subj001_20031019_111740/* .

fmr_preprocess  -fwhm 6 -target 3 study_subj001_20031019_111740_3_mri.mnc.gz

start Matlab by typing matlab at the command line

frametimes =  [  0  3  6  9  12  15  18  21  24  ...  351  354  357  ];

slicetimes = [  0  0.100  0.200  0.300  0.400 ... 2.30  2.40  ];

events = [

	1	30	30	1
	2	90	30	1
	1	150	30	1
	2	210	30	1
	1	270	30	1
	2	330	30	1 ];

X_cache = fmridesign ( frametimes, slicetimes, events );

input_file = 'study_subj001_20031019_111740_3_mri_MC.mnc';

output_file_base = [	'study_subj001_20031019_111740_3_mri_MC_hmw';
	'study_subj001_20031019_111740_3_mri_MC_hpw' ];

contrast = [	1	-1	;
	1	1	];

exclude = [ 1 2 ];

which_stats = '_mag_t  _fwhm  _mag_sd  _mag_ef' ;

fwhm_rho = [ ];

n_trends = [ 3 1 1 ];

df = fmrilm ( input_file, output_file_base, X_cache, contrast, exclude, which_stats, fwhm_rho, n_trends );

quit Matlab by typing exit

cd /data/scratch/scratch1/yourname/study_subj001

register study_subj001_20031019_111740_3_mri_MC_hmw_mag_t.mnc study_subj001_19971024_1_2_mri.mnc.gz &

Additional Usage Notes and Tips

fmrilm
- df and fwhm values are written into the mincheaders of the output files
- Regarding the fwhm of the spatial smoothing of the autocorrelation parameters: if the fwhm_cor is negative, it is taken as the desired df of fmrilm's output, and the default is -100.
- For effective connectivity of a voxel, you will need setup your contrast variable as a struct. To do this, add a '.C' (note the uppercase) to the contrast variable name: contrast.C = [  1  ];

multistat
- DO NOT COMBINE percent of whole volume results with old voxel value based results!
- The current version of multistat now outputs a single struct to the Matlab environment variable space. Do not attempt to assign more than one value, such as [df, df_resid] = multistat(...), which will result in a Matlab error. Other values are available from the struct, such as df.resid and df.fixed. See multistat's help message in Matlab for more information.

extract
- extract has been changed so that the ordering of the voxels input is maintained in the output, instead of ordering the output according to slice number
- If you have coronal images with dimension ordering (y, z, x), you will need to re-order the image to axial (z, y, x) order to use extract properly - use mincreshape to do this:

mincreshape -dimorder zspace,yspace,xspace <input> <output>

locmax
- If you would like an output volume of cluster indices, check that the CLUSTER_FILE variable filename includes a valid image file extension, like '.mnc' for minc output
- Note however that stat_summary outputs a cluster index mask file by default.

stat_summary
- You can format the output by typing in 'format short g' before stat_summary - check 'help format' for more information

fMRI Related Frequently Asked Questions

For information regarding this web page, please contact Michael Ferreira:

E-mail:

Office phone:

(514) 398-8547