19 February 2016
David Zwicker
Brenner Group, School of Engineering and Applied Sciences
Harvard University
Natural odors typically consist of many molecules at different concentrations, which together determine the odor identity. This information is collectively encoded by olfactory receptors and then forwarded to the brain. However, it is unclear how the receptors can measure both the composition of the odor and the concentrations of its constituents. I will present a theoretical model for which we derive design principles for optimally communicating the odor information. These principles can be summarized as two possibly conflicting goals: (i) each receptor should respond to half of all odor mixtures; (ii) activity patterns of different receptors should be orthogonal. We use this result to discuss the properties of optimal receptor arrays as a function of the statistics of natural odors. Additionally, we use our framework to discuss other receptor properties, like the accuracy of concentration measurements and the capability for discriminating mixtures. Taken together, we can predict the performance and properties of receptor arrays based on a few, measurable quantities. Our work can thus be used to infer information about the receptors from physiological measurements. Moreover, we can use our results to improve artificial sensor arrays.