|Home | About | Journals | Submit | Contact Us | Français|
When two different images are presented to the two eyes, we perceive alternations between seeing one image and seeing the other. Termed binocular rivalry, this visual phenomenon has been known for over a century , and has in recent years been systematically studied at both the behavioral and neural levels . Similar phenomenon has been documented in audition . Here we report the discovery of alternating olfactory percepts when two different odorants are presented to the two nostrils. This binaral rivalry involves both cortical and peripheral (olfactory receptor) adaptations. Our discovery opens up new avenues to explore the workings of the olfactory system and olfactory awareness.
Most of our sensory organs come in pairs: eyes, ears, and nostrils. Typically, the two eyes form slightly different retinal images of the same object (binocular disparity). There are small differences in time and intensity between a sound arriving at one ear versus the other, as well as between a smell arriving at one nostril versus the other . The two nostrils are asymmetrical in air flow, which switches every couple of hours , and in their sensitivity to odorants with different sorption rates . Most of the time, our brain integrates these minor differences and generates stable, accurate representations of the environmental input (e.g. stereopsis, sound localization, odor localization [4, 7, 8]). Rivalry occurs when two distinctively different images are separately presented to the two eyes [1, 2]. Successive periods of dominance of the left-eye stimulus and the right-eye stimulus are described as unpredictable in duration, as if being generated by a stochastic process driven by an unstable time constant [9, 10]. Similarly, when alternating tones an octave apart are played out-of-phase to each ear, most listeners experience a single tone oscillating from ear to ear, whose pitch also oscillated in synchrony with the localization shift — a demonstration of rivalry between the two ears. Here we set out to test whether rivalry also exists in olfaction.
In Experiment 1, Phenyl ethyl alcohol (PEA, 0.5% in propylene glycol, 8ml) and n-butanol (0.5% in propylene glycol, 8ml), each contained in a narrow-mouth bottle fitted with a Teflon nosepiece, were simultaneously presented to a subject’s two nostrils, so that one nostril was exposed to PEA while the other was exposed to n-butanol. Subjects sampled from the two bottles intermittently (see Supplemental Experimental Procedures), instead of continuously; this is done because olfaction is especially prone to adaptation (occurring within 30–40s of odor presence) [11, 12]. The two odorants have differences in structure and smell. Both carry a hydroxide radical, but PEA has a benzene ring, whereas n-butanol has a chain structure (Figure 1A). PEA smells floral and is usually described as a ‘rose’ smell, while n-butanol has the smell of a marker pen. Across twenty samplings, all twelve subjects experienced switches between smelling predominantly the rose smell and smelling predominantly the marker smell (Figure 1A and Table S1, see Figure 1B for an illustration of the visual analogue scale used for olfactory similarity ratings). Some experienced more frequent and drastic switches than others. On average, to the same individual, the percepts of the same two odorants altered from a maximum of 79.2% like ‘rose’ to a maximum of 72.8% like ‘marker’, which is comparable to the range of similarity ratings when subjects were exposed to PEA and n-butanol alone (78.9% like ‘rose’ to 85.7% like ‘marker’, see Supplemental Experiment 1 and Figure S1). This separation is even greater across the entire sample of 12 subjects, ranging from 94% like ‘rose’ to 92% like ‘marker’. Whereas how biased a subject was towards perceiving the ‘rose’ smell or the ‘marker’ smell, as reflected by the mean of his/her similarity ratings across the 20 samplings, follows a normal distribution with the mean at 53.9% similar to ‘marker’ (Figure 1C), their similarity ratings form a bimodal distribution, with the local maxima at 66% similar to ‘marker’ and 65% similar to ‘rose’ (Figure 1D). This shows that the observed fluctuations (Figure 1A) cannot be due to large random sampling errors; rather, they reflect genuine switches in olfactory percepts within the subjects.
No predictable pattern of the switch was evident across the subjects or within the same subject, in line with what is observed in binocular rivalry [9, 10]. Nine out of the twelve subjects perceived mostly ‘marker’ at the beginning, possibly because n-butanol was a ‘stronger’ stimulus than PEA. Although rated as equally familiar to the subjects [F (1, 11) = 0.048, p = 0.83], n-butanol was perceived to be more intense [F (1, 11) = 12.13, p = 0.005] and less pleasant [F (1, 11) = 31.29, p = 0.00016] than PEA, independent of the side of the nostrils (left, right, or both) tested [F (2, 22) = 1.26, p = 0.30 for familiarity; F (1.26, 13.81) = 3.32, p = 0.083 for intensity; F (2, 22) = 0.73, p = 0.49 for pleasantness]. Such dominance of the ‘stronger’ competitor is also well documented in binocular rivalry [13–15].
The intensity of the perceived smell decreased over the twenty samplings [F (19, 209) = 1.97, p = 0.011], but its pleasantness was not affected by the number of times the odorants were sampled [F (19, 209) = 1.19, p = 0.27]. Across the twelve subjects, there was significant correlation between the pleasantness and the similarity ratings (how similar the smell is to ‘rose’ or ‘marker’) of the perceived smell [Bivariate Pearson correlations between pleasantness and similarity ratings, each obtained using a 100-unit visual analogue scale as described in Supplemental Experimental Procedures; average r = 0.40, s.e.m. = 0.12, t (11) = 3.30, p = 0.007], mirroring the pleasantness difference between PEA and n-butanol.
The intermittent nature of samplings prevents us from adequately characterizing the temporal dynamics of olfactory rivalry, as the interval between two adjacent samplings was typically around 20–30s, including the time when the subjects made the similarity, intensity, and pleasantness ratings. Still, the dispersion in the bimodal distribution of the similarity ratings (Figure 1D) suggests that the transitions between the two olfactory percepts were likely marked by mixed percepts. This is analogous to what is observed in visual rivalry , although such an analogy should be viewed with some caution due to the differences in the nature of stimulus delivery; whereas intermittent stimulus presentation, chosen here to reduce olfactory adaptation, is also used in visual rivalry, the majority of studies on the latter adopt continuous stimulus exposure.
Similar to binocular rivalry , the binaral competition observed here is related to adaption. In Experiment 2, when one nostril was adapted for 2mins to PEA, and then the same nostril was again presented with PEA while the other nostril was presented with n-butanol, subjects (N = 4) reported smelling the ‘marker’ smell. Conversely, when one nostril was pre-adapted to n-butanol, and then the same nostril was again presented with n-butanol while the other nostril was presented with PEA, the same subjects reported smelling the ‘rose’ smell. Nevertheless, Experiment 2 does not tell us whether the contribution of adaptation is due to central (adaptation occurred in the cortex) or peripheral (adaptation occurred at the peripheral receptor neurons) components.
As a preparatory step towards addressing this issue, we examined the effect of adaptation on the perceived intensity of the odorants in Experiment 3. Subjects were adapted for 2mins to an odorant in one nostril, and then rated the perceived intensity of the same adapting odorant or a different odorant in either the same or the other nostril. As would be expected from adaptation, when either PEA or n-butanol was presented to the nostril that had been pre-adapted to it, it was rated as much less intense [t (11) = −4.64, p = 0.001] than before the adaptation (Figure 2). One interesting question is whether such adaptation is purely peripheral, i.e. only due to the fatigue of the peripheral olfactory receptor neurons over prolonged exposure to the odorant. We find this not to be the case. When the same odorant was presented to the other nostril, which had not been adapted to it, there was also a significant drop of its intensity rating [t (11) = −3.57, p = 0.004]; although the effect is less drastic as compared to when it was presented to the pre-adapted nostril [t (11) = −2.66, p = 0.022]. Hence both cortical and peripheral mechanisms are involved, as previously demonstrated by Cain . This adaption is odorant-specific. The intensity rating of the odorant (n-butanol or PEA) that had not been adapted to was not affected [t (11) = −0.74 and −1.63, p = 0.47 and 0.13, respectively, for the two nostrils] (Figure 2).
Subsequently in Experiment 4 and 5, we set forth to assess whether both cortical adaptation and adaptation of the olfactory receptors contribute to the alternations in olfactory percepts observed in Experiment 1. We hypothesized that if cortical adaptation is an important component of binaral rivalry, alternating olfactory percepts would be experienced independent of adaptation in the olfactory epithelium (mononaral rivalry) (Experiment 4), as in monocular rivalry . Indeed, ten out of the twelve subjects (83%) experienced switches between smelling predominantly ‘rose’ and smelling predominantly ‘marker’ when they intermittently sampled from two bottles, each containing a 1:1 mixture (8ml) of PEA (0.5% in propylene glycol, 4ml) and n-butanol (0.5% in propylene glycol, 4ml) (Figure 3 and Table S1). On average, for the same individual, the percepts altered from a maximum of 70% like ‘rose’ to a maximum of 78.7% like ‘marker’. Across the 12 subjects, the similarity ratings ranged from 90% like ‘rose’ to 92% like ‘marker’. Similar to the aforementioned binaral rivalry situation, subjects experienced a decrease in the intensity of the perceived smell [F (19, 209) = 2.19, p = 0.004] over time. Their pleasantness ratings, again significantly correlated with the similarity ratings across subjects [average r = 0.44, s.e.m. = 0.11, t (11) = 3.94, p = 0.002], were not affected by the number of times the odorants were sampled [F (19, 209) = 1.11, p = 0.35].
Concerning the peripheral adaptation at the olfactory epithelium, we hypothesized that if it also plays a significant role in binaral rivalry, a swap of the sides of the two olfactory stimuli would render the previously suppressed smell perceivable again (in parallel to what is observed in binocular rivalry [20, 21]). To test this idea, in Experiment 5, subjects were instructed to simultaneously and continuously sniff from two bottles, one containing PEA (0.5% in propylene glycol, 8ml) and the other containing n-butanol (0.5% in propylene glycol, 8ml), until they can no longer detect the smell they firstly did (e.g. if a subject firstly smelled ‘marker’, he was instructed to keep sniffing until he did not smell the ‘marker’ smell). Then unknown to the subjects, the two bottles were either quickly swapped or not swapped and re-presented to the two nostrils. Consistent with our hypothesis, ten out of the twelve subjects tested (83%) reported smelling the same smell again (e.g. marker) when the bottles were swapped, but not when the bottles were not swapped.
It is worth noting that although the mononaral rivalry (Experiment 4) resembles binaral rivalry (Experiment 1) in perceptual experience (Figure 1A & Figure 3), the two recruit different mechanisms. Whereas mononaral rivalry is independent of adaptation in the olfactory epithelia located in the two nostrils (Experiment 4), there is a significant peripheral component in binaral rivalry, as shown in Experiment 5. These results are consistent with what has been found in visual rivalry [22, 23].
In the visual system, inhibitory interactions could take place among both monocular neurons (binocular/interocular competition) and binocular pattern-selective neurons (monocular/pattern competition), and the persisting neural signals could be passed on to higher stages of processing, where visual competition can continue . Anatomical parallel exists between the olfactory system and the visual system. Olfactory system is largely ipsilateral . Odorants entering one nostril is detected by the olfactory epithelium, from which the olfactory information is conducted to the ipsilateral olfactory bulb. Axons of the mitral and tufted cells of each bulb coalesce and form the olfactory tract, one on each side, which conveys olfactory information ipsilaterally to the primary olfactory cortex (anterior olfactory nucleus, olfactory tubercle, anterior and posterior piriform cortex, amygdala, and rostral entorhinal cortex). There is inhibitory interaction between the two olfactory bulbs . In addition, there is inhibitory interaction among olfactory bulb glomeruli , which receive olfactory inputs from different types of odorant receptors . The two olfactory tracts are nevertheless connected to each other via the anterior olfactory nuclei and the anterior commissure [28, 29]. Such anatomical substrates possibly contribute to the binaral and mononaral rivalries observed here, yet the neural mechanisms of olfactory rivalry await to be elucidated.
We have shown alternating odor percepts when two different odorants are presented to the two nostrils, thereby demonstrating, for the first time, perceptual rivalry in the olfactory system. The binaral rivalry involves adaptations at the peripheral sensory neurons and in the cortex. Our work sets the stage for future studies of this phenomenon, which will further characterize its perceptual properties, delineate the neural correlates of cortical and peripheral adaptation, and elucidate the mechanisms of olfactory awareness .
We thank Weiji Ma for helpful discussions. This research was in part supported by NIH R03DC4956.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.