Dear Peter Thanks again for your quick reply. 4x4x1 is essentially 4x4. Please correct me if I am wrong. With 4x4 it is still the same error. I dig deep in to the code, and found that for some unknown reason, kl_weights=spm_kl_eig_normal(w_mean(:),w_cov,prior_cov); is coming out as NaN which is then leading to the same error. Dont know if that helps Is it possible that you may send me a working DCM.mat so I can get to know in which format the inputs are required ? That will be great help and hopefully will solve my problem Thanks and best On Fri, Sep 9, 2016 at 1:42 PM, Zeidman, Peter <[log in to unmask]> wrote: > Dear Rabia > > Your b-matrix should be of size [n x n x u], where u is the number of > inputs. As can be seen from DCM.U, you have one input (‘null’). At the > moment your bmatrix is of size 4x4x4. So you need to change it to be 4x4x1. > I have now added code to give clearer error messages, which will be in the > next release. > > > > Note that as this is resting state, you can use spm_dcm_estimate(DCM) for > stochastic DCM (which operates in the time domain) or spm_dcm_fmri_csd > (which operates in the frequency domain). The latter is a new approach > (Razi et al. 2016), which is quicker and potentially more accurate. > > > > Best > > Peter > > > > *From:* SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] *On > Behalf Of *Rabia Mazhar > *Sent:* 09 September 2016 12:00 > *To:* [log in to unmask] > *Subject:* Re: [SPM] DCM Failed Estimation error > > > > Dear Peter > > Thanks a lot for your suggestion and your quick reply. But the error still > continues. Here is my attached new DCM with connectivity matrices as 4 > dimensions. > > I will be grateful if you can suggest a solution > > > > Thanks and best > > Rabia > > > > On Fri, Sep 9, 2016 at 12:21 PM, Zeidman, Peter <[log in to unmask]> > wrote: > > Dear Rabia > > Your DCM has 4 regions (DCM.n), as well as timeseries for 4 regions (in > DCM.xY and DCM.Y), and timing offsets (delays) for 4 regions (in > DCM.delays). However, all of your connectivity matrices are of dimension 3 > rather than 4 (DCM.a, DCM.b, DCM.c, DCM.d). I suspect that is causing the > problem. > > > > Best > > Peter > > > > *From:* SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] *On > Behalf Of *Rabia Mazhar > *Sent:* 08 September 2016 13:43 > *To:* [log in to unmask] > *Subject:* [SPM] DCM Failed Estimation error > > > > Dear SPM users > > I m getting the following error while estimating DCM > > Error in inline expression ==> spm_fx_fmri(x,v,P)*2 > > Index exceeds matrix dimensions. > > > > I already have the time series extracted from the ROIs, and I m using this > to create the DCM structure. I have attached my DCM variable probably I > have some mistake in defining it. Can somebody please look in to it and > help me what I have made wrong here ? > > > > Thank you > > > > best regards, > > Rabia > > >