Dear SPM M/EEG experts,
I am learning the Chapter 47 of the SPM manual ‘DCM for Cross Spectral Densities: Anaesthesia Depth in Rodent Data’. However, I cannot repeat the findings in this chapter. Please see below for what I did:
‘dLFP_white_noise_r24_anaes.mat’ from https://www.fil.ion.ucl.ac.uk/spm/data/dcm_csd/ was loaded with GUI of DCM for M/EEG, time window was set as [25000, 30000], detrending parameters were set as 1 for detrend 1 for subsample and 2 for the modes, the between-trial effects were set as [0 1 0 0; 0 0 1 0; 0 0 0 1], the extrinsic modulation and intrinsic modulation were set separately in two CSD CMC models (extrinsic vs. intrinsic) and two CSD NMDA model.
Finally, the Model Posterior Probability of the extrinsic and intrinsic models were the same based on both CMC and NMDA neural masses. The trial effects of 1.4%, 1.8%, 2.4% and 2.8% are also the same in the intrinsic models. Theses findings were different from that provided in the Chapter 43. May I ask what I was missing or if I need to preprocess the data?
I am new to the neural mass models in M/EEG. May I ask how can I get the intrinsic connectivity from the estimated DCM file? For example, if I understand it correctly, for one EC which links two ROIs in a CSD CMC model, there should be modulatory effect on the intrinsic connectivity within each ROI, e.g., from inhibitory neurons to the superficial and deep pyramidal neurons in ROI A or ROI B, whereas in a CSD NMDA model there should be ion channel dynamics including AMPA, GABAA and NMDA receptor, correct? Are these parameters in the DCM.Ep? If so, what’s their names in DCM.Ep? Is there a manual about how to extract the estimated parameters?
Looking forward to your kind suggestions. Many thanks.
Best,
Li Zhi
|