Dear Yumeng
The generative model in DCM for fMRI is implemented in two functions: spm_fx_fmri.m , which contains the neural and haemodynamic model, and spm_gx_fmri.m, which translates from haemodynamics to the BOLD response. For a detailed description of the model, please see https://doi.org/10.1016/j.neuroimage.2019.06.031 . In particular, see appendix 5 which details the haemodynamic / BOLD model. Note that this is a mechanistic model detailing how the BOLD response is generated, unlike SPM_get_bf, which gives a canonical haemodynamic response, with no mechanistic detail.
As you'll see, CBV is inferred from the BOLD data in DCM for fMRI, whereas I think you have the ability to measure it using MION, which might require a change to the model. It would be an interesting (and technical challenging) project.
To get started with DCM, I recommend that you do specify a GLM, as this will get the data into the right format for DCM. Feel free to get back to us if you need help with that.
Best
Peter
-----Original Message-----
From: SPM (Statistical Parametric Mapping) <[log in to unmask]> On Behalf Of Xin yumeng
Sent: 15 March 2022 05:45
To: [log in to unmask]
Subject: {SPAM?} [SPM] Questions about using DCM with monkey data
⚠ Caution: External sender
Dear SPM experts,
I was a new biginner to the SPM DCM analysis. I have questions about using DCM on my own monkey data.
Our data was collected with MION, making the HRF was different from that implemented in SPM12.
If I only want to use the DCM module (without preprocessing or GLM), should I need to add the MION HRF to the SPM function (Like SPM_get_bf.m)?
And would the differences between the traditional BOLD and MION HRF have an obvious influence on the DCM results (Since the DCM was established from human BOLD signals)?
Thanks for your help in advance.
Best regards,
Yumeng
|