Dear SPM List,
I am interested in using DCM for the analysis of some resting state data. I have been using some of my own data, and have used the nodes specified in Razi et al 2017( Large-scale DCMs for resting-state fMRI), and the batch script from here https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=SPM;fcc93eec.1709 (somewhat amended as I have already denoised the data).
As an initial check I have compared my findings to those in the paper and the distribution of in/out strength relationships are quite different from those in the paper. I was hoping to check if I have made any basic errors.
1. I have assumed that the DCM.Ep.A corresponds to the effective connectivity matrix (e.g. shown in fig 5A, fig 7B of the above paper), and that columns represent ‘out’ connections, and rows ‘in’ connections. Is this correct?
2. For the purposes of calculating the average in/out strength for a node (e.g. Fig 9,10), were self connections ignored? The text states that they were “suppressed”. If so is Figure 9 simply plot of the mean of the columns vs the mean of the rows of DCM.Ep.A (after removing the diagonal)?
3. When/how is bayesian model reduction step discussed in the paper implemented
Many thanks!
Rob
|