Direct behavioral evidence for retronasal olfaction in rats

Abstract

The neuroscience of flavor perception is becoming increasingly important to understand abnormal feeding behaviors and associated chronic diseases such as obesity. Yet, flavor research has mainly depended on human subjects due to the lack of an animal model. A crucial step towards establishing an animal model of flavor research is to determine whether the animal uses the retronasal mode of olfaction, an essential element of flavor perception. We designed a go- no go behavioral task to test the rat's ability to detect and discriminate retronasal odorants. In this paradigm, tasteless aqueous solutions of odorants were licked by water-restricted head-fixed rats from a lick spout. Orthonasal contamination was avoided by employing a combination of a vacuum around the lick-spout and blowing clean air toward the nose. Flow models support the effectiveness of both approaches. The licked odorants were successfully discriminated by rats. Moreover, the tasteless odorant amyl acetate was reliably discriminated against pure distilled water in a concentration-dependent manner. The results from this retronasal odor discrimination task suggest that rats are capable of smelling retronasally. This direct behavioral evidence establishes the rat as a useful animal model for flavor research.

Figure 1. Hypothesis and retronasal-specific setup. A: Schematic diagram of obligate nasal breathing.

Both rats and human infants are “obligate nasal breathers” where exhaled air from the epiglottis (green) does not effectively pass over the oropharynx (red). Moreover, the raised epiglottis may obstruct oro-nasal odorant passage (red stars). The question mark indicates that prior to this work it had not been directly tested if retronasal smell occurs in rats and infants. In humans the epiglottis descends around the fifth month of age. Such developmental decent does not occur in rodents. B and C: Experimental set up. Schematics of an animal performing go no-go orthonasal (B) or retronasal (C) odor discrimination task (left) and the time course of a single orthonasal trial (right). Orthonasal odorants (2% saturated vapor) were released in front of the nose by an olfactometer (B). Retronasal odorants dissolved in water were delivered (50 µl, after a lick) at the lick spout by a gustometer (C). Irrespective of the odor source, the rat could obtain water by licking the lick spout after sampling (by sniffing, B, or licking, C) an S+ odor. Vac: a vacuum tube sucking air (5 L/min) from around the lick spout to prevent orthonasal exposure. D–G: CAD flow models suggest rats were unable to sniff the lickspout. D: flow lines and flow velocity cut plot (i.e. a color coded flow rate along the median plane, see text; color bars: blue-red  = 0–3 m/s) at maximum reported sniff flow rate (1.8*10⁻⁵ m³/s). Lateral view. E: same at twice the maximum reported sniff flow rate. F: as in E, but now entire flow (3.6*10⁻⁵ m³/s) through the right naris (a 4-fold over-estimation). Isosurface plots (i.e. the “balloons” around the spout and the right naris) indicate volumes with at least 0.1 m/s flow rate. Cut plot color scale: blue  = 0 m/s, red  = 0.5 m/s. G is an enlargement of Fig. 1F minus the cut plot for clarity. Flow line color bars: blue  = 0 m/s, red  = 0.1 m/s.

Figure 2. Learning an orthonasal odor discrimination task did not help learning discrimination of the same odors retronasally.

A. Orthonasal odor discrimination by an average performing rat on the third day of training. Each block consisted of 20 trials separated by 10 s (ITI) with an additional 10–15 s (punishment) for an incorrect lick. S+ = 1% (s.v.) 2-hexanone, 2hex; S− = 1% (s.v.) vinyl cyclohexane, VC; avg  =  mean of S+ and S−. B. Orthonasal (left) and retronasal (right, from fourth block onwards) odor discrimination by a rat on day 10. Note that even though the rat was performing well for orthonasal odors (a), she failed to discriminate the same odors retronasally (b). Orthonasal odors were the same as in A. Retronasal odors were as follows: S+ = 0.01% 2-hexanone in water, 2hex; S− = 0.01% vinyl cyclohexane in water. C. Average daily performance of 3 rats for ortho- and retronasal discrimination of 2-hexanone vs. vinyl cyclohexane. Rats learned to discriminate the orthonasal odors as early as day 3 of the training, but failed to do so for orally ingested/retronasal odors.

Figure 3. Successful taste-guided learning of go-no go retronasal odor discrimination task by rats. A

. An example of a retronasal daily session for the three rats (a–c) and their average (d) (day 34 of B). S+ = 0.03% amyl acetate in water, S− =  water. B. Successful performance of go-no go retronasal odor discrimination by 3 head-fixed rats. Aqueous solutions of odorants were initially combined with tastants in order to enhance shaping (day 1–9). Tastant were gradually removed (day 10–21), leaving only retronasal odorants (day 22–29). Subsequently the same animals learned to discriminate a different odorant against water in a concentration-dependent manner (day 30–45, with cue control day 35–37). C. Color legend for B & D. D. Overall performance of the three rats. Tasteless retronasal amyl acetate was convincingly discriminated against water in a dose-dependent manner. ANOVA on concentrations across rats, F_(4,10) = 16.8, P<0.0005.** t-test: above control, p<0.005; * t-test: above control, p<0.05.

Figure 4. Overall performance of all six rats tested.

All 6 animals learned to perform the task well above chance level, and convincingly discriminated tasteless amyl acetate against water in a concentration-dependent manner. ANOVA on concentrations across rats, F_{(4, 25)} = 22.0, P<10⁻⁷.*** t-test: above control, p<10⁻⁴; ** t-test: above control, p<0.005.

Figure 5. Confirmation of strictly retronasal detection of odorants.

A–B: To ensure that rats were only depending on licked retronasal odors, we added a constant flow of clean air (5 L/min) targeting the nose. Flow models again confirmed no possible flow from lickspout to nose. Overall performance of 3 rats under these conditions averaged over 4 sessions per odor is shown (EB: 1 session). Note that rats were still able to discriminate tasteless odorants against distilled water, and their performance was not affected significantly. ANOVA on stimuli across sessions, F_{(3, 8)} = 43.5, P<0.0001. *** t-test: above control, p<10⁻⁵; ** t-test: above control, p<0.005.