Dear SPM experts,
I'd appreciate if you could help us resolve a methodological question concerning a group-level analysis.
In our study, we are interested in looking at the activations within the singing network in stroke patients performing a singing task. The singing network was determined using an independent dataset of demographically-matched healthy participants based on our a priori hypothesis that the patients will show different activation within this network.
Our first approach was to use this singing network from healthy controls as a small volume correction in the results from the patients group. The volume of this singing network represents 12.5% of the total number of voxels.
The second approach was to use the singing network as an explicit mask during the 2nd level specification.
Are both approaches statistically comparable? Which one would be more suitable to test our hypothesis?
Thank in you in advance for your help.
Best wishes,
Noelia
|