Skip to main content

A physicochemical model of odor sampling

Cite this dataset

Gronowitz, Mitchell et al. (2021). A physicochemical model of odor sampling [Dataset]. Dryad.


We present a general physicochemical sampling model for olfaction, based on established pharmacological laws, in which arbitrary combinations of odorant ligands and receptors can be generated and their individual and collective effects on odor representations and olfactory performance measured. Individual odor ligands exhibit receptor-specific affinities and efficacies; that is, they may bind strongly or weakly to a given receptor, and can act as strong agonists, weak agonists, partial agonists, or antagonists. Ligands interacting with common receptors compete with one another for dwell time; these competitive interactions appropriately simulate the degeneracy that fundamentally defines the capacities and limitations of odorant sampling. The outcome of these competing ligand-receptor interactions yields a pattern of receptor activation levels, thereafter mapped to glomerular presynaptic activation levels based on the convergence of sensory neuron axons. The metric of greatest interest is the mean discrimination sensitivity, a measure of how effectively the olfactory system at this level is able to recognize a small change in the physicochemical quality of a stimulus.



Mice carrying OMP-IRES-tTA:tetO-Kir2.1-IRES-tauLacZ alleles (Kir2.1 mice) and wildtype controls were used as subjects. Kir2.1 mice were weaned on P21 and then maintained on a 20 mg/kg doxycycline-inclusive diet for at least six weeks prior to behavioral testing to induce the multiple receptor type (MRT) phenotype (Figure 8A). Littermate control animals are referred to as “single-receptor type” (SRT) mice. All animals were maintained in the Stowers Institute Lab Animal Services Facility on a 12:12 light cycle, and provided with food and water ad libitum except as described for two-alternative forced choice experiments. All behavioral experiments were carried out during the animals’ dark cycle under red or infrared illumination. All experimental protocols were approved by the Stowers Institute Institutional Animal Care and Use Committee (protocol 2019-102) and in compliance with the NIH Guide for the Care and Use of Laboratory Animals.


Monomolecular odorants were purchased from Sigma-Aldrich and diluted in mineral oil to the desired concentration as described below. Maple and lemon flavors were purchased from the Frontier Co-op (; cat #23081 and #23071) and were not diluted in the liquid phase. Odor delivery was controlled by an automated olfactometer with custom LabVIEW software (National Instruments) as described previously. All odorants presented were diluted tenfold in the gas phase by the olfactometer in addition to liquid-phase dilutions as noted. All mixtures were prepared in the liquid phase. Odorants are listed in Table 1.

Table 1:List of odorants

Odorants utilized in behavioral experiments along with their CAS numbers and chemical structures. The four odorant pairs used for cross-habituation studies (at 70:30 vs 30:70 mixture ratios; Figure 8B) are identified with descriptive labels. Odorant pairs used in two-alternative forced choice tests (2AFC; Figure 8C-D) are identified with their panel letter.

Cross-habituation task

Cross-habituation tests were performed in 2-4 month old mice as previously described [70]. Each experimental group contained 6-14 animals. Each animal was tested with a total of 4 odors (2 pairs) in 2 separate experiments with at least one week between tests. Testing was performed in a 20 x 20 cm chamber to which the animals were first habituated for 30 minutes. Odorants were delivered by the olfactometer at 100 ml/min into a nose cone positioned on a side wall at 5 cm height. A vacuum tube connected to the opposite wall of the nose cone served to remove residual odors after odor delivery. Monomolecular odorants were diluted into mineral oil at 1:1000 (v/v), whereas lemon/maple flavors were not diluted in the liquid phase; for all odorants, a 10 ml/min air flow through the odorant-containing vials was mixed into 90 ml / min carrier air to yield a final nominal dilution of 10-4 saturated vapor (s.v.) for monomolecular odorants and 0.1 s.v. for lemon/maple flavors. Animals’ investigation of the odor source was registered by infrared beam breaking events and recorded by the same custom software that controlled the olfactometer.

In each trial, odorants were delivered for 1 minute followed by 4 minutes of carrier air. Mice first were habituated to the protocol by delivering eight trials of plain mineral oil, after which the first odorant of the pair (habituated odor) was presented for 5 trials and then the second odorant of the pair (test odor) was presented for 3-5 trials. During each trial, the duration of the animals’ investigation of the odor port was recorded. Repeated identical trials led to systematically reduced investigation (habituation), whereas perceived changes in the delivered odor restored the investigative response. Presentation of odorants perceived as similar to the habituated odor (i.e., not discriminated) evoke no increase in investigation time [1]. A normalized port investigation time (NPI) value was calculated for each trial by dividing the investigation time by the total baseline duration of odor port investigation during background delivery. The perceptual distance between the two odorants of a pair then was assessed as the difference in NPI (ΔNPI) between key trials.

Specifically, ΔNPI was calculated as the difference in investigation times between the fifth (last) presentation of the habituated odor (thab) and the first presentation of the test odor (ttest), expressed as a percentage over the average baseline exploration time during background air presentation.Embedded Image

Values near zero therefore indicated a failure to discriminate between the paired odorants.

Two-alternative forced choice task (2AFC)

Separate cohorts of 2-4 month old mice were used for this paradigm, in which mice are motivated to perform difficult odorant discriminations. In all cases, mice were water restricted (1.5 ml of water per day for one week) prior to training. Food was available ad libitum. Testing was performed in a 20 x 20 cm chamber with three nose cones positioned at 5 cm height. Odorants were delivered by the olfactometer at 100 ml/min into the central nose cone as described above, whereas the flanking nose cones were used to deliver water rewards. Mice were conditioned to nose poke into the central port, whereupon odorant delivery was triggered. Two odorants were presented in this fashion in a pseudo-random sequence, and water reward was contingent upon the animal correctly associating the odor with the appropriate water port. If the animal chose the correct port, 0.05 ml water was released; otherwise, no water was given. Animals were trained to criterion in the basic task (80% correct), before proceeding to the experimental paradigm.

The odorant discrimination success rate (SR) was calculated asEmbedded Imagewhere P(A|a)P(B|b), and Ptotal were the number of pokes into water port A upon delivery of odorant a, the number of pokes into water port B upon presentation of odorant b, and the total number of nose pokes into the odor port, respectively. A success rate of 50% denotes chance performance.

In the first 2AFC experiment, 11 MRT and 18 wildtype mice were trained to discriminate the structurally and perceptually similar odorants methyl valerate and methyl butyrate at 1% saturated vapor (10-2 s.v.). Upon reaching criterion, discrimination performance at the testing concentration was assessed, after which the concentrations of both A and B were systematically reduced for subsequent testing at three lower concentrations (10-3, 10-4, and 10-5 dilutions). Liquid phase dilutions into mineral oil were 10% v/v for training and 1%, 0.1%, and 0.01% v/v for testing, each followed by an additional 10% dilution in the gas phase (see Odorants above).

In a second set of 2AFC experiments, mice were trained to discriminate pairs of odorants until achieving criterion (80%), after which they were systematically tested on their capacity to discriminate increasingly similar mixtures of the odor pair. For example, after being trained to discriminate amyl acetate from heptanal, 8 MRT and 8 wildtype mice then were tested on their ability to discriminate a mixture of 90% amyl acetate / 10% heptanal from a mixture of 10% amyl acetate / 90% heptanal, then 80:20 vs 20:80, and so on until reaching 50:50 vs 50:50, on which they performed at chance. All odorants/mixtures were presented at 10-4 dilution (1000x liquid phase dilution plus 10x gas phase dilution). This experiment was repeated in a separate cohort of 10 MRT and 10 wildtype mice using natural lemon and maple flavors, which were diluted only in the gas phase (10-1 overall dilution).

Usage notes

This dataset is the original data of the behavior experiments for Figure 8A-D.

.lsm files are images that can be opened with software such as Fiji or ImageJ.


National Institute on Deafness and Other Communication Disorders, Award: R01 DC014701