A gas chromatographic (GLPC) model for the sense of smell. Variation of olfactory sensitivity with conditions of stimulation.

Abstract Computer simulation of an olfactory detector has been developed using a chemical kinetic scheme originally proposed by McNab and Koshland for bacterial chemotaxis. This model describes response as a function of two opposed reactions, both of which are activated by odorant. One reaction turns on response, while its opponent shuts it off. Net response to various stimulus profiles is compared to psychophysical experiments, with particular attention paid to simulating magnitude estimation and odor adaptation results. Effects of the access route to this detector are evaluated. Transport of odorant molecules is treated as having two sequential steps: step (i), airborne odorant is carried parallel to a retentive layer (mucus) into the detector region; step (ii), molecules diffuse through the retentive layer to the detector. Step (i) is represented as analogous to GLPC on an open tubular column. Each step has a characteristic time constant, which is proportional to (distance)2/diffusion coefficient. Response to highly volatile odorants tends to be limited by step (ii), while odorants of low volatility approach the step (i) limit. Sensitivity at both limits is attenuated by increasing the thickness of the retentive layer, but sensitivity at the step (i) limit is also affected by changes in air passageway and airflow characteristics. This picture can be used to explain variations in women's sensitivity to odorants of low volatility with the menstrual cycle, while their detection of volatile odorants fluctuates to a much lesser extent.