Odors were presented in a boxcar format (Fig. 1) during three 36 second periods, alternating with periods of sham valve openings (controls) that only shunted the same non-odorized control air stream used between odorant stimulations. Within a given 252 seconds functional time series, all three 36-second odor blocks consisted of a single stimulus class (AO, NApO, or ApCO). In the first 3 functional times series acquisitions, stimulus classes were presented in a pseudo-random sequence across subjects, either (NApO, ApCO, AO), (ApCO, AO, NApO), or (AO, ApCO, NApO). For every subject, this sequence of 3 times series acquisitions was then replicated, resulting in 6 functional time series per subject/session.
