Dear Lizhi,
Thank you. This is probably a way to get around the dimension limits
with NIfTI-1. Would you be able to name which FreeSurfer function is
generating these images or if there is some documentation about how it
exactly performed?
Are you using DCM for resting state fMRI or do you create a dummy SPM.mat?
Best regards,
Guillaume.
On 08/03/2019 21:16, Zhi Li wrote:
> HI Guillaume,
>
> Thank you very much. There should be 160k+ vertices per hemisphere which
> equals to 27307 X 6. Maybe it is the way how FreeSurfer saves the data.
>
> I am wandering that if I can whiten the signals with AR(1) first, then
> feed the DCM with these whitened time series. Give the whitening matrix
> in X0 was one time series(1X303), it seems like DCM in SPM controls the
> effect of nuisance in X0 which includes whitening noise and filter
> information (through regression maybe?). Without the X0 matrix, I can
> run the DCM scripts smoothly without any errors. So the main question is
> how to control the nuisance in the times series, is that correct? I
> think process the time series before applying DCM maybe a compromised
> method, such as whitening ,filtering or even wavelet despiking. Any
> suggestions would be appreciated.
>
> Thank you and best regards,
>
> Lizhi
>
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|