afni.nimh.nih.gov

Can anyone explain to me the meaning/value of using varying lags in 3dDeconvolve when your stimulus function has been generated using waver?

In my case, my stimuli are not locked to the TR, so I am hoping to use the modelled hemodynamic response functions generated by waver as an input for 3dDeconvolve.

Another way of stating my question, I suppose, is: when you are basically hoping to use 3dDeconvolve to perform multiple linear regression, what does varying the lags imply?