[Freesurfer] fsfast: using an empirically derived HRF with mkanalysis-sess

hi freesurfers,

i would like to use an empirically derived HRF with mkanalysis-sess. i have derived the HRF i want [via an FIR model], but it's not obvious how i can use my custom HRF with mkanalysis-sess.

one option, it seems, is to feed in my own regressors with the -taskreg flag and then not supply the model with any other regressors of interest. i'd just convolve stick regressors that have the appropriate timing with my own HRF. but this would only work if selxavg3-sess does NOT convolve task regressors with the HRF.

so two questions:

1. is the -taskreg flag the best way to do this? is there another, better way?

2. does selxavg3-sess convolve task regressors with the HRF?

thanks, alex