In the 1980s, only a handful of publications existed on neural correlates of olfaction in humans. Psychophysical and behavioral papers were more numerous, and anatomical and physiological data derived from studies carried out in “macrosmic species”, especially rats, existed, but the means with which to study brain function in humans was limited. We had the good fortune to have those means available to us. Working at the Montreal Neurological Institute, we had access to people with focal surgical excisions that had been performed as a treatment for intractable epileptic seizures. We each had ongoing research programs studying performance on various auditory, verbal, and visuospatial cognitive tasks as a function of surgical lesions in different brain regions. This paradigm allowed us to infer a significant role on the part of specific brain regions on particular tasks based on impaired performance when those regions were damaged. Neither of us specialized in olfaction at the time. Robert Zatorre’s (RZ’s) primary interest was in audition, and Marilyn Jones-Gotman’s (MJG’s) was memory, especially nonverbal memory. Curiosity about memory for odors led MJG to begin olfactory investigations, while his interest in perception was RZ’s motivation. Being each poised independently and simultaneously to launch olfactory studies in the MNI patients, we decided to join forces. We started without funding, and accumulating odorants to begin our experiments was an adventure.We acquired some in the hospital pharmacy and local grocery stores, but the more exciting acquisition came from the Brown University Smell Library, which was being closed because Trygg Engen, a longtime olfactory psychophysicist at Brown, was retiring. Because RZ had trained at Brown, he knew Trygg and his former student, Bob Mair, who kindly alerted us to this opportunity. We were given a fairly large number of odorants, which RZ transported in the trunk of his car from Brown to Montreal without being suspected by the border police or their dogs. We started with the basics: detection thresholds, quality discrimination, memory, and identification. The unilateral surgical lesions of most of our patients were in the temporal lobes and also in the frontal lobes, with a few in parietal or occipital lobes. We found that although detection thresholds were not affected by any of these surgical lesions, there were deficits on all other tasks. The biggest deficits occurred specifically with orbitofrontal lesions and not with damage to other frontal cortical areas. There were impairments also after temporal lobe surgeries. Parietal or occipital lobe lesions did not affect our tasks. These results also usually showed a right hemispheric predominance in olfaction. The temporal and frontal lobe findings confirmed our predictions, based on findings in other species, that important olfactory functions depended on cortical structures in the anteromedial frontal cortex and in the orbitofrontal cortex. A laterality difference was a novel finding, not having been observed in animal studies, but because it replicated over several studies we were tempted to believe it to be real. After working on olfaction for about 5 years with this lesion approach, we were very fortunate to have the opportunity to perform the first functional neuroimaging study of olfaction. In the late 1980s, there had been developments, at the MNI and elsewhere, in using positron emission tomography (PET) to study brain metabolism. Initial attempts using fluorodeoxyglucose as a tracer were largely unsuccessful because of the long acquisition times (>60 min) and correspondingly poor ability to resolve different events or to test M. Jones-Gotman (*) : R. J. Zatorre Montreal, QC, Canada e-mail: marilyn.jonesgotman@mcgill.ca