A model of the role of adaptation and disadaptation in olfactory receptor neurons: implications for the coding of temporal and intensity patterns in odor signals.

Natural odors occur as turbulent plumes resulting in spatially and temporally variable odor signals at the chemoreceptor cells. Concentrations can fluctuate widely within discrete packets of odor and individual packets are very intermittent and unpredictable. Chemoreceptor cells display the temporally dynamic properties of adaptation and disadaptation, which serve to alter their responses to these fluctuating odor patterns. A computational model, modified from one previously published, was used to investigate the effect of adaptation and recovery of adaptation (disadaptation) on the spike output of model olfactory receptor cells under natural stimulus conditions. The response characteristics of model cells were based upon empirically determined dose-response, adaptation, disadaptation and flicker fusion properties of peripheral olfactory cells. The physiological properties of the model cell (adaptation and disadaptation rate and the dose-response relationship) could be modified independently, allowing assessment of the role of each in shaping the responses of the model cell. Complete adaptation and disadaptation time courses ranged from 500 ms (rapid cells) to 10 s (slow cells). The stimuli for the model cells were quantified odor plume recordings obtained under a variety of biologically relevant flow conditions. As expected, the rapidly adapting model cells displayed different response characteristics than the slowly adapting model cells to identical temporal odor profiles. Responses of the model cells depended upon their adaptation and disadaptation rates, and the frequency characteristics of the odor presentation. These results indicate that adaptation and disadaptation determine the range of concentration fluctuations over which a particular cell will respond. Thus, these properties function as an olfactory equivalent of a band-pass filter in electronics. This type of filtering has implications for the extraction of information from odor signals, such as the coding of temporal and intensity features.