I am very new to AFNI but trying to get started obtaining single trial betas in an event related design.
We have successfully used 3dDeconvolve with the stim_times_IM option to get “LSA” single-trial betas, and then have used 3dLSS on the -x1D output to get LSS single trial betas.
But my understanding is that this pipeline includes no autoregressive modelling at all - to get pre-whitening / AR modelling I need to use 3dREMLfit.
I guess we can still use stim_times_IM to get LSA single trial betas including pre-whitening output from 3dREMLfit. But is there anyway to get LSS single trial betas including modelling of temporal autocorrelation?
ie is there some way to pass pre-whitened data and model matrix from 3dREMLfit to 3dLSS?
Or is there any reason this would be unnecessary or even not recommended?
Thanks for any advice,