 |
 |
Dear Becky, OK - thx for FAIR/FIAR resolution :-) You may have a local minima problem - but it may be too early to jump to this conclusion. Given that you know the true parameters you should check (i) are they within the range of the priors (See spm_dcm_fmri_priors.m) - if not the optimisation algorithm will never find them, (ii) are the parameter estimates DCM.Ep close to the true values, esp important for the connections that are different between models. It may also be the case that the models are not identifiable given single subject data. This is sometimes the case for event related designs which may lack power (same reasoning as for GLM analyses), and one can only identify the models at the group level. Its probably a good idea to use spm_dcm_generate etc as Martin suggested. Best, Will. > -----Original Message----- > From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] > On Behalf Of Becky van den Honert > Sent: 30 April 2012 20:42 > To: [log in to unmask] > Subject: Re: [SPM] DCM fails to select model that generated the data > > Thank you for your feedback! I mistakenly assumed that the self- > connection parameters defaulted to -1. In case you are wondering, the R > package I used was FIAR (not FAIR): http://www.jstatsoft.org/v44/i13. I > will try fixing the A matrix and running in SPM as well. > > If you have the time, I have a follow up question: > > That makes sense that my problem might have something to do with the EM > algorithm settling on a relatively poor local maximum for the correct > model. How can one mitigate this risk? If the EM algorithm starts at an > arbitrary point, then it would seem that increasing subject or run > number would help. Are there properties of a specific model structure > (or space) that would make one more or less susceptible to this? > > Thank you again, > Becky