Hello,
The FEAT documentation ( http://www.fmrib.ox.ac.uk/fsl/feat5/detail.html ) describes how these two numbers are used:
If Cluster thresholding is selected, a Z statistic threshold is used to define contiguous clusters. Then each cluster's estimated significance level (from GRF-theory) is compared with the cluster probability threshold. Significant clusters are then used to mask the original Z statistic image for later production of colour blobs. This method of thresholding is an alternative to Voxel-based correction, and is normally more sensitive to activation. You may well want to increase the cluster creation Z threshold if you have high levels of activation.
Using randomise with the option "-c 2.3" and then thresholding the output 1-p images at 0.95 ( 1 - 0.05 ) is broadly analogous with the cluster-thresholding in FEAT.
The randomise manual ( http://www.fmrib.ox.ac.uk/fsl/randomise/ ) covers a number of example of how to run randomise ( including a two-sample t-test ).
Many Regards
Matthew
> Dear FSL team,
>
> In the post-stats, there are two threshold can be chosen in FEAT, one is min Z (>2.3) and the other is cluster significance: P(<0.05). It is very clear for the P-value, but I have no idea about the Z-value. What does it mean?
>
> If I wanna use Randomise common to carry out the same post-stat as the cluster Thresholding of FEAT, which options should I choose?
>
>
> For example, I have two group of people, A and B, with their resting state fmri data. A seed was defined and the correlation between it and whole brain voxels had been computed for each subject. After applying Fisher'z transform to the correlation, each subject got a Z-maps. Then, I carried out a two-sample t-test between these two groups and got a raw test statistic image. What should I do to correct the P-value using the same method of FEAT by randomise command?
>
