Hi everyone, as you can see in the image above I come across with this gray areas which is not good. However, I couldn’t solve it even though I used different data or made sure that my design is right. Do you have any ideas what might be the problem?
Hi Beyza,
From what I can see, the design matrix is of full rank and the three parameters are uniquely specified (the three cells are white) so I would not worry about multicollinearity in this design. If you have further concerns, it would help to give more information about the design (first/second level?) and show the design matrix.
Hope this helps,
Guillaume.
Hi, you can try my script for first-level analysis. It works for my data.
Thank you a lot for your response!
Hi Beyza,
This design matrix is of full rank, as illustrated by having all the cells underneath the image in white (and not in grey). Which contrasts are you interested in? Given that there doesn’t seem to be a baseline, this design will only be sensitive to differential effects.
You might be interested in reading this document from Rik Henson:
http://imaging.mrc-cbu.cam.ac.uk/imaging/DesignEfficiency
Hope this helps,
Guillaume.
Hi again,
I’ll check the document as soon as possible. Before that, the answers are here. This is actually a localizer scan, I’m stimulating three different areas from different visual quadrants (not the whole quadrant). At first, I got three runs with 6TR stimulus-6TR ISI for each location, however for time efficiency recently I combined three locations to one run.
Since these are in the discrete visual fields, I’m assuming that if one area is not stimulated at the moment, it is in the rest. Is this meaningful? (I consider interhemispheric effects)
According to this logic when I want to see activation from the first condition, I’m writing contrasts as 1 0 0.
Best,
Beyza
MANAGED BY INCF
