Hi Rosalia,

If you used TFCE, there's no need to get worried with significance of cluster sizes; the peak significance already encapsulates both extent and height of the map at a given voxel. In fact, TFCE doesn't output p-values for cluster, but p-values for voxels individually.

All the best,
Anderson

On 11 September 2014 14:27, Rosalia Dacosta Aguayo <[log in to unmask]> wrote:

Dear FSL experts,

I have run randomise with -m option (roi of a size of 15mm) regarding a selected independent component.

I have got significant tfce corrected results: 2 voxels, p=0.043; coordinates (23,13,30) in standard space resolution of 4mm.

The question is, taking into account that I am doing t-statistics within this small ROI...the size of the cluster is really significant?

This question comes from other questions sent to the forum list regarding the significance of the size of the cluster....what to report and what no report...

Kind regards,

Rosalia.