How do I...
- Specify a T-contrast in SPM?
- Specify a T-contrast in AFNI?
- Specify an F-contrast in SPM?
- Specify an F-contrast in AFNI?
- Construct a contrast for an FIR model? Or a basis function model?
- Construct a contrast for a parametric modulation?
- Carry out a conjunction analysis?
* Compare the activation in one contrast to the activation in another contrast? (i.e., find differences/similarities in activation between two contrasts)
One quick way to visually compare activations between 2 contrasts is by displaying the results of 2 contrasts and their overlapping regions. The following steps allow you to show the thresholded results of 2 different contrasts (A and B) in two different colors; overlapping regions between A and B are displayed in a third color.
1. bring up your results from one contrast (contrast A)
2. hit the "write filtered" button to create an image of the activations
3. bring up your results from the other contrast (contrast B)
4. hit the "write filtered" button to create an image of the activations
5. To create the "overlap" image of A and B, hit "imcalc", select the 2 images that were created using "write filtered", then enter "i1.*i2" in the calculation field and save the image.
a. Select the background anatomical image first (could be "single_subject_T1" the "canonical" directory)
b. For number of images to display, hit "3"
c. Select the 3 images you created above (A, B, and overlap)
d. Choose you want each image to display in (I recommend, yellow, blue, and green - in that order)
NOTE: The numbers contained in the overlap image shouldn't be used for quantitative analyses; this image is just used for display purposes. If you want to do a legitimate conjunction analysis, you should take the minimum T-value across image A and B - I have a program to do this which I can give more info about if needed. -Dara
- Make contrasts for a group ANOVA?
- Make contrasts for a group multiple regression?