Dear SPM experts,
I am learning the DCM method mentioned in study ‘Dynamic causal modelling revisited, Friston et al., 2017’. I have several practical questions about setting priors on connectivity and interpreting findings when I run through the demo script and data:
1. In ‘priors on connectivity’ of ‘DEMO_dcm_fmri_nnm’ in the 1st study, what’s the two 3x3 matrices in DCM.a? According to my understanding, the 1st matrix sets priors on forward extrinsic connection from V1 to V5 and V5 to FEF, and the 2nd matrix sets priors on backward connection from V5 to V1 and FEF to V5, is that correct? If so, why cannot we just use one matrix containing both forward and backward connection?
2. In addition to the 1st question about A matrix, after ‘spm_dcm_neural_priors’, there are 4 matrix in ‘nE.A’, what are they? Should we modify it?
3. In ‘priors on connectivity’ of ‘DEMO_dcm_fmri_nnm.m’ in the 1st study, may I ask what’s the two 3x3x3 matrices in DCM.b? I am not clear what does each cell in each matrix mean hence do not know how to modify it to specific hypothesis.
4. What’s the difference between conventional DCM and this laminar fMRI DCM in setting the A and B matrices?
5. About the output of ‘DEMO_dcm_fmri_nnm’, we could know the contribution of extrinsic and intrinsic connections from neural populations to BOLD response in each region via neurovascular coupling, is this the main purpose of this method and also the main finding which should be reported in routine study? But we cannot know how experimental stimuli modulate specific connection between populations of regions (B matrix in conventional DCM), correct? For example, we cannot know how ‘Attention’ modulate connection from superficial pyramidal cells in V5 to spiny stellate cells in FEF and connection from inhibitory cells in FEF to spiny stellate cells in FEF. If I am wrong, may I ask where I could find this information in the output matrices?
Looking forward to your kind suggestions. Thanks.
Best and wishes,
Zhi Li
|