Dear FSLers,
I’m writing with a query of FEAT vs. randomize.
I have run a FEAT analysis looking at activation in the following for 4 conditions:
1. ev1
2. ev2
3. ev1-ev2
4. ev2-ev1
For I each looked at:
Patients
Controls
Patients > Controls
Controls > Patients
The third comparison (ev1 – ev2) was the one I was most interested in. Based on hypotheses and observations of the data from ev1 and ev2, I expected significant group differences such that Controls were more active than Patients for this contrast. I did not get significant differences – so I tried randomize as I know that this approach might be more accurate in identifying significant clusters of activation.
For the randomise analysis, I used cope3 images from each subject to correspond to the condition ev1 – ev2, and the design.mat file from the FEAT analysis above. I get the clusters I expect, which is great. However, I don’t get clusters in any other condition, i.e. for the Patient and Control conditions separately. These were robust effects in FEAT so why is randomise not detecting them? Or did I miss s/t? It just seems odd that the group difference would be identified but not the individual groups’ activation clusters.
I would really appreciate if you could shed some light on what might be going on.
Thanks for your time,
Anna-Maria
|