But if I ever wanted to do an event-related design where measuring timings precisely is more important then I'm not sure if simply offsetting the timeseries would still be considered acceptable.
Simple offsetting is generally "considered acceptable", even for event-related designs. Is is "considered optimal"? No. But it might have less of a risk of introducing artifacts or distorting signal than using a poorly-fitting (or too-complex) model.
Another relevant reference is Akama et al 2012 (DOI:10.3389/fninf.2012.00024): figure 4 compares the canonical HRF and single-volume classification accuracies. They're close, but not identical, another example of how the information in fMRI is really blurred over time.
Jo -- Joset A. Etzel, Ph.D. Research Analyst Cognitive Control & Psychopathology Lab Washington University in St. Louis http://mvpa.blogspot.com/ _______________________________________________ Pkg-ExpPsy-PyMVPA mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
