Dear SPM list,
I wish to analyse the effects of one parametric modulator (Pmod) on the HRF informed basis set (i.e. three basis functions). At the 1st level, individual design matrices included condition regressors and one parametric modulator regressors (mean-corrected reaction time, pmod_RT), both convolved with the three basis functions (i.e. conditon HRF, condition temporal derivative (condition TD), condition dispersion derivative (condition DD), Pmod_RT HRF, Pmod_RT TD, Pmod_RT DD). Basic t-contrasts on the Pmod_RT HRF, Pmod_RT TD, Pmod_RT DD regressors resulted in three con images for each subject. Then, at the 2nd level, I used a 3-level Full factorial design with Pmod_RT HRF as level1, Pmod_RT TD as level2, Pmod_RT DD as level3.
As I know, for condition regressors, the earlier/later peak of HRF is determined by the conditon HRF (beta1) and condition TD (beta2) (Henson, 2002)
beta2 + / beta1 + = early peak
beta2 - / beta1 - = early peak
beta2 - / beta1 + = later peak
beta2 + / beta1 - = later peak
Therefore, for parametric regressors, what's the relationship between the earlier/later peak of HRF and Pmod_RT HRF (para_beta1), Pmod_RT TD (para_beta2), is the following right?
para_beta2 + / para_beta1 + = as RTs increase, a earlier peak HRF
para_beta2 - / para_beta1 - = as RTs increase, a earlier peak HRF
para_beta2 - / para_beta1 + = as RTs increase, a later peak HRF
para_beta2 + / para_beta1 - = as RTs increase, a later peak HRF
Besides, can you suggest some references about this issue.
Thanks a lot in advance!
Best Regards,
Yizhou Jiang
|