At 13:56 08/02/2006, you wrote:
> > Dan - I agree that most fmri studies are probably underpowered, but
> > I don't beleive that loosening our Type I error rate is the way to
> > fix this problem. Rather, people need to run enough subjects that
> > they have power to find a sufficient effect size at a reasonable
> > Type I error rate.
This is not just a question of the power of our fMRI studies, but also the
sloppiness with which authors and reviewers sometimes draw inferences from
supra-threshold voxels. The inference that can be legitimately drawn from a
voxel surviving FWE p<0.05 is different from that which can be drawn from a
voxel surviving FDR p<0.05. It is not a simple matter of greater type I
error. We have read alot on the SPM list in the last couple of years
about the ease with which one can make erroneous inferences from
conjunction analyses, or about regional differences in
activation. FDR-abuse can be added to the list of well-intentioned
pitfalls. larger studies are preferable in many ways, but they are not a
panacea.
James
|