I'm relatively new to FSL (used SPM and afni). We were just looking at group FEAT results (the thresh_zstat1.nii image where we used the default z threshold of 2.3 and a cluster threshold of p<.05). We have a robust result that ends up getting us three clusters, the largest of which includes has multiple subregions all over the brain. For example the cluster includes amygdala, basal ganglia, VLPFC....it's all over but visually it looks nonrandom and like multiple functional subregions.
It's a similar situation to this one
Re: Cluster index with 1 cluster. Output image shows several clusters.
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=fsl;6da2dcb6.1512
In SPM there are typically subpeaks reported.
In afni it was recognized that sometimes are a few contiguous voxels that "connect" multiple regions so it's possible to apply connectivity thresholds to clip these subregions to make sense of them.
Is FSL actually using something like TFCE where there is a more liberal threshold to join subregions?
https://users.fmrib.ox.ac.uk/~steve/ftp/HBM07_TFCE_poster.pdf
Does FSL have something like these other programs to identify functional subregions to make sense of these large clusters?
Or are we using too liberal thresholds and should we be doing more corrections beyond this? Someone told me that people use AFNI's 3dClusterSim or FSL's PALM.
Is there a tutorial I should be reading?
Thanks,
Allyson
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|