 |
 |
Hi! If you have a close look at your data (e.g., with Fslview), you will usually find that there are connecting voxels between your "isolated" clusters, thus making it one big cluster. For a separation of clusters, increase your constituting z-value (e.g., 2.3 -> 3.1), but be prepared to lose some clusters (or all). Alternatively, use TFCE and decrease your p-value (0.05 -> 0.01). Be advised, though, that cluster inference and its increased sensitivity to activation comes at a price, namely reduced spatial "resolution" (i.e., you cannot "pick" one of your sub-clusters if FEAT says it's only one - all sub-areas are equally valid in terms of localization power!) Cheers, Cornelius Am 27.09.2010 um 19:15 schrieb Ignacio Letelier: > Hi experts. > > Is there anyway how to set and/or increase the number of reported clusters of the cluster list in the FEAT report? > Why sometimes the FEAT reports only one cluster in its cluster list whereas images have more than one? > > thanks in advance > > best regards > -- > Ignacio Letelier Aguirre > Ingeniero Civíl Biomédico > Clinica Santa María. > Teléfono: 97139955 >