help
help > RE: Comparing graph metrics across toolboxes
Sep 18, 2015 01:09 AM | Roger Beaty
RE: Comparing graph metrics across toolboxes
Hi Alfonso,
It turns out that my matrices hadn't been binarized after doing the previously mentioned transformations (my original goal was to analyze weighted networks). After binarizing, the global efficiency values are now exactly the same as those produced by the BCT.
Thanks again for your help,
Roger
Originally posted by Alfonso Nieto-Castanon:
It turns out that my matrices hadn't been binarized after doing the previously mentioned transformations (my original goal was to analyze weighted networks). After binarizing, the global efficiency values are now exactly the same as those produced by the BCT.
Thanks again for your help,
Roger
Originally posted by Alfonso Nieto-Castanon:
Hi
Roger,
That seems strange, I did a quick test just to double-check and I am seeing exactly the same global/local efficiency values when using CONN's conn_network_efficiency.m compared to BCT's efficiency_bin.m functions. Could you please give me more details about how exactly you are performing this comparison to see if that helps me figure out the source of this discrepancy?
Thanks
Alfonso
Originally posted by Roger Beaty:
That seems strange, I did a quick test just to double-check and I am seeing exactly the same global/local efficiency values when using CONN's conn_network_efficiency.m compared to BCT's efficiency_bin.m functions. Could you please give me more details about how exactly you are performing this comparison to see if that helps me figure out the source of this discrepancy?
Thanks
Alfonso
Originally posted by Roger Beaty:
Hi all,
I've been running graph analysis in CONN but would also like to compare results across other toolboxes (e.g., Brain Connectivity). I transformed the z matrices to r, thresholded them at cost (k = .15), and removed NaN values and negative correlations. However, the global efficiency values are noticeably smaller in the Brain Connectivity package compared to CONN. I'm fairly positive that the same equations are used across toolboxes, so I'm not sure why the results would be different. Any insight would be greatly appreciated.
Thanks,
Roger
I've been running graph analysis in CONN but would also like to compare results across other toolboxes (e.g., Brain Connectivity). I transformed the z matrices to r, thresholded them at cost (k = .15), and removed NaN values and negative correlations. However, the global efficiency values are noticeably smaller in the Brain Connectivity package compared to CONN. I'm fairly positive that the same equations are used across toolboxes, so I'm not sure why the results would be different. Any insight would be greatly appreciated.
Thanks,
Roger
Threaded View
| Title | Author | Date |
|---|---|---|
| Roger Beaty | Sep 17, 2015 | |
| Alfonso Nieto-Castanon | Sep 17, 2015 | |
| Roger Beaty | Sep 18, 2015 | |