Dear Ece,
I don’t think the DCM-IND parameters would make any difference. They are just ignored in DCM-ERP. In principle the same data should be selected independently of what the total epoch length is. I suggest that you look at DCM.xY.y fields for the two cases and compare the data to see where they differ. You can also look at the DCM.ID fields. If they are not the same it means there is something different about the data, you just need to find what it is.
Best,
Vladimir
On 31 Mar 2014, at 15:37, Ece Kocagoncu <[log in to unmask]> wrote:
> Dear experts,
>
> I have been working on a DCM-ERP analysis for some time, where I carried out the same analysis twice. I initially used the -200 to 1000 ms epoched file, and estimated my DCM models for the 1-600 ms time window.
>
> As a secondary question, we were also interested in determining the precise time window within this 600ms frame, when the effects we're seeing were happening. To find out about this I carried out the same DCM analysis in 100 ms short time windows (i.e. 1-100 ms, 1-200 ms, ... , 1-600 ms).
>
> The problem was that the script that I was using was modified from the DCM-IND batch script, and it had time-frequency decomposition related parameter settings like the wavelet cycle, and the frequency range. Unfortunately I forgot that DCM-ERP did not require these settings, and I decided to use an input file with a longer epoch (-650 to 1650 ms) to account better for the theta band power!
>
> When I compared the results from the first and second analyses (1-600 ms window only), I realized that the results were not identical. I have not changed any other DCM parameters in these analyses. The source reconstruction steps I followed prior to DCM estimation were identical (IID, inversion over -200 - 1000 ms window, contrast over 0-600 ms window).
>
> I would appreciate if you can help me understand why the results of these two analyses are different, when the only difference is the length of the input file I supplied.
>
> Thank you!
>
> --
> Ms. Ece Kocagoncu, MSc
> PhD student
> Centre for Speech Language and the Brain
> Department of Psychology
> University of Cambridge, UK
|