Dear Jack and fellow SPMers,
> Masking per se will not implement a restriction of the search volume
> (it simply allows you to select a subset of voxels). To effect a
> constrained search you have to explciitly perform a small volume
> correction using the cluster defined by a mask. Simply select SVC and
> then the cluster option and the table will change to give you corrected
> p values for that cluster.
This thread has prompted me to update my vol_corr software. vol_corr uses
very similar algorithms to those in SPM99b, implementing the small volume
corrections of Worsley et al (1996). However, it may be of some added use
because it:
a) has an interface to allow you to specify your anatomical hypothesis in
terms of an image, so that you can use results from previous experiments
and other literature to define your correction, and
b) will give you t/Z/F thresholds for given p values such as p = 0.05, 0.01, and
c) works with matlab 4.2 and SPM96.
It is still in beta, but I have tested it pretty carefully. There is a
little tutorial on how to use it and install it at:
http://www.mrc-cbu.cam.ac.uk/Imaging/External/vol_corr.html
I would be very grateful for any feedback,
Thanks,
Matthew
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|