When saying voxel-level, I mean using voxel correction by means of GRF-theory based maximum height thresholding..
Diederick
On Jun 19, 2012, at 3:01 PM, Diederick Stoffers wrote:
> Hi all,
>
> We performed a mixed effects group comparison on fMRI task data; a patient group showed hypoactivity in a 50-voxel cluster (Z=3.1 cluster corrected) within an overall task effect mask relative to controls.
>
> We than treated half the patient group and compared it to the untreated half after seven weeks of waitlist or treatment using a fixed-effects model with an interaction between time and group and looked within the 50-voxel mask of the initial group difference using voxel-corrected statistics at p < .05.
>
> A reviewer is questioning our use of voxel-level statistics and this threshold, he would like to see a reference to a study employing similar methodology. When using voxel-level statistics within a mask, p < .05 is not an overly liberal threshold to use, right? Could someone provide a good ref?
>
> Cheers,
>
> Diederick
|