zvalim@gmail.com wrote:
Dear FreeSurfer Gurus
I have two questions about surfcluster that I would appreciate some help with....
- What 'connectivity radius' is used in surfcluster? It seems to be
hardwired, because I can't find a flag to specify it. By connectivity radius I do not mean the minimum cluster size that of course can be specified, but the minimum distance in mm that two vertexes can be within in order to be considered part of the same cluster on the surface. If this value is low, then clusters are probably defined as cases of adjacent vertexes (perhaps within 1 edge distance?), but I could see uses for changing this parameter so that 'adjacency' would be defined somewhat more liberally (especially in cases when smoothing is low). SUMA's surfclust has the -rad option that achieves this (["Maximum distance between an activated node and the cluster to which it belongs.Distance is measured on the surface's graph (mesh)"] and I was wondering if FreeSurfer has this option. If not, does anyone know what the hard coded value is? Knowing it will help choosing smoothing parameters on the surface.
Two vertices must share a vertex to be considered in the same cluster.
- When seeing the textual results of surfcluster, it is indeed very
useful to see the Max T, cluster size and number of nodes in cluster. But, for purposes of some permutation algorithms, it would be extremely useful to be able to know the mean T value in each cluster. This enables calculating the *mass* of each cluster, defined roughly as the number of vertexes multiplied by the mean T value (Bullmore, 1999). This lets one normalize for T values, and define 'reliable' clusters are those that pass a certain mass, independent of their size.. Would it be possible to get the mean T value in each cluster as an output from surfcluster? Obviously one could script a utility that for each cluster, computes its mean T by cross referencing the cluster-tagged output with the T value of each vertex (pivot table by cluster), BUT, getting the mean cluster T directly from surfcluster would be a great feature.
Currently, this is not an option, but I'll consider it for future implementation.
doug
Thank you for your time and patience,
Best, Uri
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer