search

csb0 / PCDM

pupil common drive model (PCDM) toolbox.
Other
1 stars 2 forks source link

putativeIRFdur has no effect on IRF duration estimate #5

Closed nicobast closed 1 year ago

nicobast commented 1 year ago

I noticed that in the fitParametricPuRFfromData.m you define:

impulseTime = 4000; % 4 seconds. this is an assumption used throughout, based on observation that saccade-locked purf goes back to baseline after 4 sec this option is hardcoded and thus does not consider the user defined in.putativeIRFdur from dataAnalysis. I guess, all estimate will be on a 4 second time window irrespective of the in.putativeIRFdur.

csb0 commented 1 year ago

I'm considering removing the option to set the putative IRF duration yourself. It should only be adjusted if the user does principled model comparisons with different IRF lengths, and then measures the generalization performance of the model, probably using our dataset.

nicobast commented 1 year ago

Great, thanks! will test your updates next week, as I am on conferences now.

nicobast commented 1 year ago

okay, will leave the putative IRF duration at 4s.