Effects of desensitization on odors varying in concentration and pleasantness

Background: Desensitization in response to repetitive odorous stimulations is a common and well-investigated process, but it has been a matter of discussion in which way this process relates to odor valence. Our goal was to investigate changes of intensity and pleasantness induced by desensitization in odors in relation to their valence and concentration. Materials and Methods: Using air-dilution olfactometry, 30 normosmic participants received two pleasant (phenylethylalcohol [PEA], mixture of pleasant odorants [mix-P]) and two unpleasant odors (hydrogen sulfide [H 2 S], mixture of unpleasant odorants [mix-NP]). At the baseline (before desensitization), four different concentrations of each odor were presented randomly for 0.2 s, with 40 s inter-stimulus interval. During the desensitization procedure, the odor used was presented continuously for 15 s. Following an odor-free interval of 4 s the same odor was presented again, in the post-desensitization phase. Odor intensity and pleasantness were rated after presenting each stimulus in pre and post desensitization phases. Results: At baseline, Mix-P was rated as pleasant and H 2 S was rated as unpleasant, as expected, while PEA and mix-NP were rated as slightly pleasant

as expected, while PEA and mix-NP were rated as slightly pleasant or neutral, respectively.For intensity, desensitization was found for all odors regardless of hedonic tone, with the effect being more pronounced with PEA and mix-NP at lower concentrations.A decrease of pleasantness/unpleasantness ("affective habituation") was present for the two lower concentrations of the pleasant mixture, while hedonic ratings of the unpleasant odors remained largely unchanged.
Conclusions: These results suggest that, in this model, desensitization in intensity between pleasant and unpleasant odors was similar and affective habituation only occurred for the most pleasant odor.

Practical Applications
Our study unraveled the link between odor desensitization and valence considering the concentration effect, where the intensity was decreased for the relatively "neutral" odors and pleasantness was habituated only for the pleasant odor.These results help to explain individual differences in the effects of olfactory loss and provide a promising perspective for personalized treatment for olfactory loss in clinics, for instance, fast adaptation and parosmia.

| INTRODUCTION
Continuous olfactory stimulation results in a reduced intensity perception (Hummel & Kobal, 1999;Pellegrino et al., 2017).Such desensitization is a well-investigated process that can be found in all sensory modalities (Rankin et al., 2009).For olfaction, it enhances the possibility to detect new stimuli from the odorant background, especially for those hazardous odors which could threaten lives (e.g., gas leakage and rotten food) (Christensen et al., 1996;Pellegrino et al., 2017).However, patients with olfactory loss frequently report a very fast desensitization, which significantly reduces their quality of life (Chen et al., 2020).For example, during eating, some patients report that they perceive a flavor at their first bite, which then is no longer present when they try to repeat that flavorful experience.In this sense, patients are tantalized by their diminished and quickly adapting odorous perceptions.These very individual experiences become of significance in personalized care.
The degree of desensitization in intensity has been shown to depend first on intensity, then on the trigeminality and to a lesser extent to pleasantness (Croy et al., 2013;Jacob et al., 2003;Sinding et al., 2017;Stuck et al., 2014).More intense stimuli desensitize less than stimuli with lower intensity, which could relate to the higher number of recruited receptor cells and glomeruli in the olfactory bulb when increasing the concentration of an odorant (Laing et al., 2003;Pellegrino et al., 2017).It is still not entirely clear how desensitization involves odor valence, with pleasantness being regarded as a primary dimension in odor perception (Khan et al., 2007).Malodors are often associated with danger or toxicity, poor air quality, and spoiled foods, and therefore play an important role as warning signals (Dalton, 2002;Jacob & Wang, 2006).Using a repetitive odor detection paradigm, Jacob and colleagues observed that the olfactory system desensitized more rapidly and to a greater degree to malodors, with an inverse relationship between concentration and desensitization (Jacob et al., 2003).Other studies found a similar desensitization effect in intensity for both pleasant and unpleasant odors, without considering odor concentration (Croy et al., 2013;Stuck et al., 2014).
Studies in relation to the desensitization in pleasantness after repetitive exposure to odor are inconclusive.Cain and colleagues stated that exposure to odor modulates its pleasantness, producing "affective habituation" where the pleasant odor turns less pleasant or the unpleasant one becomes less unpleasant (Cain & Johnson, 1978).This phenomenon was further supported through measurements of pleasantness and sniffing duration and volume, indicating that the pleasantness of an odor decreased with repeated exposure while the unpleasantness of the unpleasant odor stayed unchanged or decreased only slightly (Ferdenzi et al., 2014).
Using a computer-controlled olfactometer, the present study aimed to investigate the desensitization effect in relation to intensity and pleasantness between malodors and pleasant odors taking concentration into account.Because malodors are typically associated with hazards, we hypothesized that desensitization to unpleasant odors would be less pronounced than desensitization for pleasant odors.

| Participants
Thirty healthy normosmic subjects (14 men, 16 women; mean age 23 years; age range 18-34 years) volunteered for the experiment.We estimated the sample size utilizing G*Power software (Faul et al., 2007).Within the difference between two dependent means in matched pairs, our sample size obtained a medium effect size of 0.53 with α level set to 0.05 and power set to 0.8.Prior to the study normal olfactory function was ascertained using the odor identification part of the "Sniffin' Sticks" test battery (Oleszkiewicz et al., 2019).Pregnant women were excluded because pregnancy may change olfactory function (Hook, 1978).Participants provided written consent after having been informed about the aims and possible risks of the study.
All investigations were performed according to the Declaration of Helsinki.The Ethics Committee at the Medical faculty of the Technical University of Dresden approved the study design (EK208082007).

| Stimuli
Two pleasant and two unpleasant odors were chosen for olfactory stimulation.Phenyl ethyl alcohol (PEA; Sigma, Deisenhofen, Germany) and a mix of pleasant, flowery odors (mix-P; Givaudan, Dübendorf, Switzerland) were used as pleasant odors, hydrogen sulfide (H 2 S; airliquide, Düsseldorf, Germany) and malodorous components reminiscent of human sweat (mix-NP; Givaudan, Dübendorf, Switzerland) were used as unpleasant odors.Each odor was presented in four different concentrations (12%, 24%, 36%, 48% v/v) using air-dilution olfactometer (OM6b; Burghart-Messtechnik, Wedel, Germany).Temperature and humidity were kept constant (36 C, 80% relative humidity).The total flow rate was 7.4 L min À1 .The air stream was presented to the left or right nostril (counterbalanced order between participants) intranasally through a Teflon tube of 4 mm inner diameter.To mask switching clicks from valves of the olfactometer subjects wore headphones playing a constant white noise of 60-70 dB sound pressure level.During the experiment, participants performed a tracking task.They had to keep a white dot inside a larger, moving square using a computer mouse on a computer screen in order to stabilize vigilance and eye movements (Kobal et al., 1990).Subjects were asked to breathe through their mouth to avoid respiratory airflow in the nasal cavity during chemosensory stimulation.

| Experimental procedure
For the baseline measurement, four different concentrations (12%, 24%, 36%, 48% v/v) of each odor were presented for 0.2 s.The interval between stimuli was 40 s.The presentation of the different stimuli and concentrations was randomized across all participants.During the desensitization procedure, the odor used was presented continuously for 15 s before each stimulation with a break of 4 s before evaluation (Figure 1).The desensitizing stimulus contained the same odor as the target but the concentration was adjusted to 5% in order not to overwhelm the participants.After each target, stimulus subjects rated the odors' intensity and hedonic tone using a horizontal linear scale of approximately 10 cm length.The left endpoint anchors of the scale were defined as "no odor perceived" or "very unpleasant" and the right ones as "very strong" or "very pleasant."Desensitization measurements were repeated 3 times.The respective ratings were then averaged for further analyses.

| Statistical analysis
Results were analyzed using SPSS version 28.0 (IBM, Armonk, NY, USA).We compared the baseline intensity and pleasantness separately using within-subject analysis of variance for repeated measures (rm-ANOVA) with the factors "odor" (mix-P, PEA, mix-NP, and H 2 S) and "concentration" (12%, 24%, 36%, and 48% v/v), followed by the Least Significant Difference (LSD) adjusted post hoc analysis to compare variables pairwise.A series of paired-sample t-tests were used to test each concentration of the odors to investigate the desensitization effect by comparing the intensity and pleasantness between baseline and desensitization sessions.To look into the relationship between the desensitization effect (measured by desensitization minus baseline) and concentration, rm-ANOVA was made with the factors "odor" and "concentration," separately for intensity and pleasantness.A twotailed p-value of less than .05was considered significant.

| Baseline characteristics of the odors
In the baseline session, rm-ANOVA showed that there were no differences in intensities between four odors and concentrations (all ps > .05).In terms of pleasantness, we found significant main effects of odor and concentration (odor: F[3, 84] = 8.71, p < .001;concentration: F[3, 84] = 5.03, p = .003).The LSD-adjusted post hoc analysis suggested that mix-P (mean rating ± sd: 68.03 ± 4.32) was the most pleasant odor, followed by PEA (57.52 ± 5.09), mix-NP (53.84 ± 4.05) and H 2 S (35.10 ± 4.32), while the concentration of 12% was more pleasant than 36% and 48%, as well as the concentration of 24% was more pleasant than 48% (all ps > 0.05).Hence, across all odors and concentrations, mix-P was pleasant, H 2 S was unpleasant as predicted, while PEA was rated as slightly pleasant and mix-NP was neutral.

| Desensitization in relation to the intensity
Paired-t test showed that all odors were rated as less intense in the desensitization session, indicating a behavioral decrease in intensity from the baseline to the desensitization session (all ps < 0.037, Figure 2a).There was an interaction between odor and concentration (F[9, 252] = 2.05, p = .035),post hoc analysis suggested more desensitization in the concentration of 12% in PEA compared with H 2 S (p = .028),and more pronounced desensitization in lower concentration in PEA and mix-NP compared to higher concentration (PEA: 12% vs. 36%, p = .02;mix-NP: 24% vs. 48%, p = .005).Interestingly, mix-NP had more desensitization than mix-P and H 2 S in the concentration of 24% (all ps < .05).In a word, more pronounced desensitization in terms of intensity appeared in lower concentrations of slight pleasant F I G U R E 1 Graphical explanation of study design.Four different concentrations of each odor (mixture of pleasant odors, mix-P; phenyl ethyl alcohol, PEA; mixture of unpleasant odors, mix-NP; and hydrogen sulfide, H 2 S) were presented randomly for 0.2 s, with intervals of 40 s.During the desensitization procedure, the odor was presented continuously for 15 s before each stimulation of the same odor in 5% concentration as the target, with an interval of 4 s.The desensitization session was repeated three times.The presentation of the different stimuli and concentrations was randomized across all participants.After each presented odor, ratings of intensity and pleasantness were recorded.
or neutral odors, with no such change in the more pleasant or unpleasant odors.

| DISCUSSION
The present study focused on the desensitization to pleasant and unpleasant odors with respect to different concentrations.We found that (i) desensitization in intensity between pleasant and unpleasant odors was similar; (ii) affective habituation only occurred for the most pleasant odor (mix-P), and (iii) desensitization in terms of intensity was dependent on the concentration of the odor, being more pronounced for the lower concentrations.
Following continuous stimulation, all stimuli were perceived as less intense, and the pleasant stimuli at lower concentrations were perceived as less pleasant.Regarding intensity, these findings are in agreement with previous data (Croy et al., 2013;Stuck et al., 2014) stating that exposure to odors induces a decreased intensity for both pleasant and unpleasant odors.In addition, the present study showed that more pronounced desensitization in intensity appeared for the lower concentration of the slight pleasant (PEA) or neutral odors (mix-NP), without such differences found for the pleasant or unpleasant odors.Low concentration of odors generally led to rapid desensitization and a strongly decreased perception (Cain & Polak, 1992;Jacob et al., 2003;Pellegrino et al., 2017).Hypothetically, the more pronounced desensitization for odors rated as neutral could be explained by their lower significance for survival (in terms of food seeking and harm avoidance) in comparison to pleasant or unpleasant stimuli with a higher intensity.
Affective habituation occurred in the present study only for a pleasant odor (mix-P), especially at the lower concentration.To our knowledge, the literature on affective habituation in relation to the sense of smell is still scarce.The pioneering work conducted by Cain and colleagues observed reduced pleasantness after exposure to a pleasant lemon odor, with reduced unpleasantness after exposure to an unpleasant rancid-smelling odor as well (Cain & Johnson, 1978).In addition, another study reported similar affective habituation for the odor initially perceived as pleasant.More specifically pleasantness   decreased for the pleasant odors, while unpleasantness for unpleasant odors remained stable (Ferdenzi et al., 2014), which in is line with the present findings.Unpleasant odors may serve as warning signals (Dalton, 1996).For example, this was described in a study by means of the shorter reaction times to detect the unpleasant odor of valeric acid compared to the pleasant amyl acetate (Jacob & Wang, 2006).
The present observations support these ideas, emphasizing that unpleasant odors are a warning toward which we exhibit little affective habituation.On the other hand, the affective habituation in the lower concentration could have potential usage in product design and marketing.For instance, a relatively higher concentration of pleasant odors could attenuate the diminution of its pleasantness although these effects may be contextual and would warrant further investigations.
In the present study, one of the supposed "unpleasant" stimuli (mix-NP) was surprisingly rated by the participants as neutral.In addition, PEA was rated to be only slightly pleasant.The prolonged stimulation with mix-NP/PEA had different effects than what was seen for the other pleasant/unpleasant stimuli.Ferdenzi and colleagues showed that the influence of repeated odor exposure on affective habituation is different according to the individual initial pleasantness.This effect was stronger when participants liked the odors at first rather than when they did not (Ferdenzi et al., 2014).Several studies revealed that cognitive factors have a major influence on odor perception (Dalton, 1996(Dalton, , 2002;;Kobayashi et al., 2008).For instance, Kobayashi et al. (2008) reported that, following repetitive olfactory stimulation, odors were rated more intense and more unpleasant when they were labeled as "hazardous" compared to the situation where they were labeled as "healthy."Taken together, initial pleasantness or unpleasantness of the odor contributes a lot to the affective habituation effect which should be considered carefully in future studies.
There are several limitations in the present study.First participants usually tend not to use the Visual analogue scales at their edges and also could be confused without intermediate values being indicated-which is an inherent issue of genuine visual analogue scales (Rousseau, 2004).Another limitation is that there was no huge variation of the intensity perception between the different concentrations which could to some extent confound results in relation to concentrations and future studies should take these into consideration.

| CONCLUSION
Taken together, the present results showed that desensitization in terms of intensity was most pronounced for "neutral odors" at lower concentrations, possibly because of their lower significance.Furthermore, affective habituation only happened for the lower concentration of the pleasant odors, while the unpleasant odors stayed more or less unchanged, indicating that adaptation to malodors is less pronounced, possibly due to their higher behavioral significance.These findings might contribute to personalized treatment in the clinical context, such as fast adaptation and parosmia, which refers to the distorted perception of odors.
Comparison of the intensity (a) and pleasantness (b) of each odor before and after desensitization sessions based on mean ratings ± standard deviation.B = baseline session; D = desensitization session; numbers refer to concentrations.*p < .05,**p < .01.