Creates a connectivity matrix using a 3D label image (the target image) and a set of streamlines. The connectivity matrix records how many streamlines connect each pair of targets, and optionally the mean tractwise statistic (eg tract-averaged FA, or length).
The output is a comma separated variable file or files. The first row of the output matrix is label names. Label names may be defined by the user, otherwise they are assigned based on label intensity.
Starting from the seed point, we move along the streamline until we find a point in a labeled region. This is done in both directions from the seed point. Streamlines are counted if they connect two target regions, one on either side of the seed point. Only the labeled region closest to the seed is counted, for example if the input contains two streamlines:
1: A-----B------SEED---C
2: A--------SEED-----------
then the output would be
A,B,C
0,0,0
0,0,1
0,1,0
There are zero connections to A because in streamline 1, the connection to B is closer to the seed than the connection to A, and in streamline 2 there is no region reached in the other direction.
The connected target regions can have the same label, as long as the seed point is outside of the labeled region and both ends connect to the same label (which may be in different locations). Therefore this is allowed:
A------SEED-------A
Such fibers will add to the diagonal elements of the matrix. To remove these entries, run procstreamlines with -endpointfile before running conmat.
If the seed point is inside a labled region, it counts as one end of the connection. So
----[SEED inside A]---------B
counts as a connection between A and B, while
C----[SEED inside A]---------B
counts as a connection between A and C, because C is closer to the seed point.
In all cases, distance to the seed point is defined along the streamline path.
cat tracts.Bdouble | conmat -targetfile targets.nii.gz -outputroot conmat_
This will output conmat_sc.csv
cat tracts.Bdouble | conmat -targetfile targets.nii.gz -scalarfile fa.nii.gz -outputroot conmat_
This will output conmat_sc.csv and conmat_ts.csv. The mean FA is computed for each tract; the average of these values is output in conmat_ts.csv.
1 some_brain_region
2 some_other_region
These names will be used in the output. The names themselves should not contain commas, spaces should be quoted eg "some brain region". The labels may be in any order but the output matrices will be ordered by label intensity.
If you only want to consider a subset of labels, such as a single hemisphere, you can pass a target name file containing only the labels you want to use. Other labels in the target image will be ignored.