Dear Donald, dear Helmut,
Thank you very much for your fast and clear answers, they have been very helpful. However, we are concerned about the need to adjust of the events to the TR for FIR analyses, because our events are not. We've found some papers where they combine FIR(events) + HRF(blocks) in a mixed design and they don't seem to perform this adjustment either. For example, Dumontheil et al. (2011) [http://www.ncbi.nlm.nih.gov/pubmed/20146600]; and the fMRI study in Visscher et al. (2003) [http://www.ncbi.nlm.nih.gov/pubmed/12948724]. In Dosenbach et al. (2006) [http://www.ncbi.nlm.nih.gov/pubmed/16731517], the jittered inter-trial interval is not specified, so we're not sure if their trials' onsets are fixed to their TRs.
Could you please explain which are the consequences of performing FIR with events not fixed to the TR? Would it cause a significant loss of statistical power? And/or a large misatribution of the signal to the event and block regressors?
Thank you
Paloma
|