Abstract: The seminal (and Nobel honored) discovery of a very large gene family of receptors in the nose in 1991 immediately suggested a possible mechanism for coding the vast and diverse set of molecules we perceive as odors. This project has been ongoing for more than three decades and most recently has included efforts by machine learning and AI tools to predict what odor quality a particular chemical structure might produce. However, the path from chemicals to perception, from chemistry through psychology or psychophysics, must go through biology and, as is its won't, biology complexifies the problem. Most of the fragrances we smell or flavors we ‘taste’ are in fact not mono-molecular compounds but complex blends of molecules numbering in the tens to many hundreds (coffee has 780 identified odorous molecules). We recently investigated how these complex blends of molecules may interact with the large family (more than 1000 in mouse) of receptors. We were able to do this using a new microscopy method developed in the laboratory of our collaborator Elizabeth Hillman at Columbia. Known by the acronym SCAPE (Swept Confocally Aligned Planar Excitation) this tool permits us to monitor the activity of thousands or primary olfactory neurons in the nasal epithelium responding in real time to complex mixtures of odors. As expected, we found that some molecules may act to activate some receptors (agonists) and at the same to time antagonize or inhibit other receptors. What we did not expect was other sorts of modulation, including apparent allosteric activation of receptor neurons. We also were surprised by the widespread extent of these modulatory effects. This significantly complicates any attempt to derive a linear or additive code (such as we find in tri-chromatic color vision). How the brain makes sense of these complex scents is now a greater mystery than when we started. The main interest is in the difference between the lower dimensional senses (vision, hearing, touch) and the high dimensional olfactory stimulus of complex blends. It would appear that novel computational solutions are required.
Noyce Conference Room
US Mountain Time
Our campus is closed to the public for this event.
Stuart FiresteinFractal Faculty