With DTI data only?

Submitted by lechap on

Hello everyone,

I have seen several published articles that used diffusion tensor imaging (DTI) data and the Brain Connectivity Toolbox to study graph theory (so no fMRI data whatsoever). Does Graphvar allow for studying graph theory using DTI data only or it necessarily requires some resting-state timeseries? It seems to me, according to the user guides and the recently published article by Kruschwitz et al (2015), that it sounds possible if some matrix were to be created first. It's only that most of the tutorials are explained through resting-state example, so I'm not sure.

Thank you for shedding light

Shady

Johann Kruschwitz

Thu, 06/04/2015 - 07:32

Permalink

Hi Shady,

You can very well use GraphVar for graph theoretic analyses of structural connectivity data. However, the GraphVar “generate conn matrix” function is specifically written for generating connectivity matrices (nxn matrix) from time series data. Thus, you cannot use GraphVar for generating nxn matrices from structural data. Best would be that you use some other program (e.g. DoDTI or ExploreDTI) to construct a structural connectivity matrix (e.g. a matrix capturing the amount of fiber tracts from region to region) that you subsequently load into GraphVar (for GraphVar the kind of matrix does not really matter as long as it is a symmetric nxn matrix). Once you have defined your nxn matrices you can make use of all the graph theoretic functions in GraphVar.

 

I know that ExploreDTI (http://www.exploredti.com/) has some specific output functions that allow exporting various nxn matrices that can be used for graph theoretic analysis (e.g in GraphVar). Alternatively I think that Mnet (http://neuroimage.yonsei.ac.kr/mnet/) or GAT (http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0040709) specifically entail pipelines for graph theoretical analysis of structural data.

 

I hope this helps,

Johann

Hi Johann,

Thanks for making GraphVar available to the community. I am just learning and starting out on some graph theoretical analyses using DTI data (3 groups of interest) and I would like to use GraphVar for that purpose. I hope you can help with some basic questions. I have already created the individual association matrices using the PANDA toolbox. Could you suggest the next proper steps for this approach? First, I load the individual subject matrices (.mat)  by "Select Subjects (Conn Matrix). What should I do next?  

Thank you very much.

Elijah

 

Forums
GraphVar