Hi Ahnate,
I guess a good way to do this is including a parametric modulator in your EV's custom txt file.
In FSL, as you'll probably know, the most common way to set the onsets and durations of your EVs is with a custom txt file with 3 columns:
1st column: onsets of the event (in your case, the balloon on the screen -I suppose! I don't know this task quite well)
2nd column: the duration of the event
3rd column: the height of the regressor. If you don't have any hypothesis here, the normal thing is to put a column of 1s. However, in your design, you could add a value for each balloon size, up to 20.
Doing so, and with a proper contrast, you'll find those voxels which are more activated with a really big balloon, and less active with a smaller one. In you're going to include a contrast of this EV with others without modulator, you should be careful with the scale you use in the modulator (you'll be comparing an activation of 20 with an activation of 1). I've never been in that situation, but I think it can be potentially important.
Some people, in addition to this modulated EV (and the other EVs you want to put into your model), also include another one with the onsets and durations of the balloon on the screen, but without modulator (3rd column as 1s). This is a kind of EV of no interest which, in theory, can help to have a better signal in your modulated EV (I really haven't experienced a great change doing this, but it could help).
Hope this helps. Warm regards,
Javier
|