Dear Pei,
The canonical HRF in SPM is defined as a difference of two Gamma
functions whose parameters are stored in spm_defaults.m:
defaults.stats.fmri.hrf = [6 16 1 1 6 0 32];
See spm_hrf.m for the definition of each of these parameters.
You could change these values to match the haemodynamic response in mice.
Best regards,
Guillaume.
On 14/11/17 15:28, Pei-Ching Chang wrote:
> Hello,
>
> We are analyzing mice fMRI data, and found out mice HRF is quite different from the default in the SPM, which is based on the human HRF. While the peak of human HRF is around 6 seconds, the peak of mice HRF is only 2.5-3 seconds. Although we have been able to use informed basis or FIR to catch fMRI activation in mice, this imposes many limitations of group analysis we can perform.
>
> Would it be possible to convolute our GLM design with a custom HRF function? I am thinking to replace human canonical function with mice one, but I can not find where the default is stored at.
>
> If we could change the default, would it be okey to use informed basis to generate a custom canonical model and then replace the default with it?
>
> Thank you,
> Pei
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|