Dear SPMers,
I have a question regarding the error correction at the second level.
After having made an error adjustment using the false discovery rate (alpha=0.05) at the
first level analysis, would there be any need to use the FDR option again in the second level
analysis for a simple t-test?
The way I am looking at this a second-level error adjustment would be particularly
bothersome if we consider a smoothing of the contrast images. I don't see any reason for
an error adjustment but the default option in SPM is the FDR adjustment. Why?
Michael
|