 |
 |
Dear Dr. Penny and SPM list I am currently simulating fMRI data and try to see how much SNR can DCM handle. What I have done so far is, 1- I created a model based on my specified structure (let us call it original model). 2- I inverted this model using spm_dcm_estimate. After this, I assumed that the ground truth values of the parameters are in DCM.Ep 3- I generated 11 models with different SNR using the original model (after inverting it); for example SNR = [0.001, 0.01, 0.1, 1, 10, 30, 50, 70, 80, 90, 100] 4- I inverted these 11 models using spm_dcm_estimate. The estimated parameters of each model are in DCM.Ep. When I compared the ground truth values of the parameters from step 2 with the estimated parameters from step 4, the parameters of models with high SNR is close to the ground truth. Is this procedure correct? Thanks in advance, Shamil,