Hi Prof.Chen,
I am interested in the HRF profile analysis. I read the article mentioned in your post. It provided evidence that HRF profile analysis has a better sensitivity and efficiency compared to the canonical HRF or sampled HRF. However, the example in the article is a fast event-related fMRI dataset. And, in the limitation part, it wrote:"one must carefully design an experiment so that the model can accurately estimate the HRF shape information at the individual level. Such designs may need to carefully consider the use of randomization to reduce collinearity through variations in inter-trial intervals and jittering of event timings across conditions."
So I have a question that whether this HRF profile analysis could be applied to block design tasks with an interval between each block. Because I try analyzing the activation of white matter in block design. But the results from "3dDeconvolve BLOCK" are not obvious. Meanwhile, to what degree of the randomization is needed for HRF profile analysis to reduce collinearity?
Thank you for your explanation.
Best regards,
Roger