I am trying to find a solution to downsample connectome network with 998 nodes from Hagmann 2008 to 148 number of regions according to Destrieux parcellation.
I need to label the connectome nodes to proper regions of anatomical parcellation.
xyz coordinates of connectome network are in Talairach coordinates so I think I need a coordinate conversion too.
But how?
Do you know any package (Python is preferred) or reference for that?
Thanks in advance for any guide.
You can probably do that with Nilearn, by sampling the destrieux parellation into the Hagmann atlas space.
See e.g. https://nilearn.github.io/auto_examples/03_connectivity/plot_signal_extraction.html#sphx-glr-auto-examples-03-connectivity-plot-signal-extraction-py
Nilearn won’t handle the Talairach to MNI conversion, but I doubt that Hagmann parcellation was created in Talairach space (most likely it was in MNI space ?).
Best,
Bertrand
Sorry for the delay.
I tried the nilearn. It provide the code to extract signal according to anatomical parcellation which is fine but I also need to down sample the connectome matrix as wall. Probably it is possible to follow this method by Cabral, NeuroImage 57 (2011) 130–139:
To downsample the structural connectivity to 66 regions, the connection strength between two regions was calculated by summing all incoming fiber strengths to the target region, and dividing it by its region-dependent number of ROIs, resulting in a non-symmetric downsampled connectivity matrix.
it also need to know which nodes belong to the same anatomical regions.
A list of labels like this:
[1,1,1, 2,2,2, 1, …]
nodes 0,1,2 and 6 are in region 1, and so on …
MANAGED BY INCF