Hi Antonella, Christine and all FSL users,
I guessed Antonella modified my previous posted question on large cluster size means that I am not alone getting these very large clusters. I referred to other DTI studies (Giorgio et al., 2010 http://www.ncbi.nlm.nih.gov/pubmed/19679191) and the reported cluster was up to 2077 voxels maximum (the left retrosplenial cluster) at corrected p<0.005.
I see >70,000 voxels using TFCE in randomise in my cases (at p<0.05). I have tried using p<0.005, however, the large cluster remains.
My questions are:
1) I believe these clusters are all statistically significant despite the large size, so are they alright to be reported in publications?
2) If the cluster size has to reduced for a more sensitive description of the location of change, I am still not exactly sure on how to determine the threshold to put in -c <thresh> for t contrasts.
(I noted there are other posts previously covering this topic but I am still not exactly sure on how to choose this arbitrary number.)
If anyone could shed some light on this, it would be much appreciated.
Thanks,
Bonnie
|