Sex Differences in Odor Hedonic Perception: An Overview

Odor hedonic evaluation (pleasant/unpleasant) is considered as the first and one of the most prominent dimension in odor perception. While sex differences in human olfaction have been extensively explored, gender effect in hedonic perception appears to be less considered. However, a number of studies have included comparisons between men and women, using different types of measurements (psychophysical, psychophysiological,…). This overview presents experimental works with non-specific and body odors separately presented as well as experimental studies comparing healthy participants vs patients with psychiatric disorders. Contrary to sensitivity, identification or discrimination, the overall literature tends to prove that no so clear differences occur in odor hedonic judgment between men and women. On the whole, gender effect appears more marked for body than non-specific odors and is almost never reported in psychiatric diseases. These findings are discussed in relation to the processes classically implied in pleasantness rating and emotional processes.


INTRODUCTION
Chemical senses are crucial in the animal kingdom and are involved in a large variety of adaptative behaviors such as food search, predators avoidance or mates selection. In humans, the processes implied in olfactory perception are well described while the role and the perceptual characteristics of olfaction are currently discussed (Brand, 2020).
The multidimensional aspect of olfaction is related to the three main perceptual characteristics of odors : the intensity (mainly driven by the odorant concentration), the quality (mainly related to the odorant chemical structure) and the hedonic valence corresponding to the pleasant or unpleasant character of the odor. Among them, odor hedonic perception is universally considered as the first and one of the most salient dimension of olfactory perception since a long time (e.g., Moncrieff, 1966;Schiffman, 1974;Land, 1979). Different aspects are usually considered in studies focusing on odor hedonic perception: the odor valence that refers to the pleasant/unpleasant character of odor stimulus, the liking that corresponds to the perceived pleasantness, the wanting that corresponds to the wish to be further exposed to the same odor stimulus and the emotional response that refers to the influence of odor on psychological and physiological states.
Nonetheless, several questions remain unclear such as the relation between the hedonic space of odors and the molecular properties of odorants. Moreover, methodological questions regarding the hedonic assessment of odors make data interpretation difficult and the cortical treatment of odor pleasantness appears highly complex. Overall, studies on odor hedonic perception reveal a large intra-and inter-subject variability and thus an extreme flexibility of odor hedonic perception. In sensory hedonic perception, gender is usually considered as a major source of variability and sex differences have been reported in the somatosensory modality (e.g., Jönsson et al., 2017;Novembre et al., 2021), in the gustatory modality (e.g., Cristovam et al., 2000) or in the visual modality (e.g., Sorokowski et al., 2014).
With respect to olfaction, sex differences are well-known and have been largely investigated within different processes such as sensitivity, identification, discrimination or memory in healthy subjects (Brand and Millot, 2002;Doty and Cameron, 2009;Sorokowski et al., 2019) and in patients with neurodegenerative disorders (Melis et al., 2019;Solla et al., 2020). Surprisingly, gender as an influencing factor of hedonic responses to odorants is poorly documented. Because numerous studies considered male and female groups, this paper aims to propose a general overview concerning sex differences in odor hedonic perception. For the sake of clarification, non-specific odors and body odors are separately considered. In the same way, type of task (scale, ranking,...), type of recording (electrophysiological, fMRI,...) and properties of odorant (intensity, pungency,...) are also considered. Finally, characteristics of subjects are taken into account such as mood or psychiatric disorders (depression, schizophrenia,...). The observed differences in odor hedonic perception are discussed in relation to the processes classically implied in pleasantness rating and emotional treatment and could serve usefully in future experimental studies using odors as affective or emotional inducers.

SEX DIFFERENCES IN HEALTHY POPULATIONS: NON-SPECIFIC ODORS
Data from the literature are reported in two tables, when sex-related differences are demonstrated (Table 1) and when no difference are reported ( Table 2). Hedonic judgment is usually evaluated using psychometric (rating scales) and psychophysiological recordings such as electrocardiography (ECG), electroencephalography (EEG), skin conductance response (SCR), or event-related potentials (ERPs), which are sometimes associated with imaging methods. When demonstrated, sex differences in odor hedonic perception seem to be odorant-dependent, albeit not systematic.

Psychometric Studies
From the beginning of the 20 th century, Kenneth (1924) showed that women rated camphor, menthol and ferric valerian as more pleasant than men while the reverse was found for pin oil, cedarwood oil, musk and tonka beans. Later, Wysocki and Gilbert (1989) demonstrated that women rated eugenol and rose as more pleasant than men while the reverse was noted for isoamylacetate and mercaptan. More recently (Seubert et al., 2009), the pleasant vanillin odor was rated as more pleasant by women compared to men whereas no difference was observed for the unpleasant hydrogen sulfide odor as well as for eugenol, a bimodal pleasant/unpleasant odor. In addition, it has been shown that men assessed patchouli oil as more pleasant than women (Trellakis et al., 2011) and women assessed phenyl ethyl alcohol as more pleasant than men . Interestingly, in hyposmic populations, no sex difference was observed before treatment whereas after treatment (theophylline) women assessed pyridine and thiphene as more unpleasant than men (Henkin, 2018).
More globally, it seems that women evaluate the pleasantness of perceived odors in a more extreme manner than men for both pleasant and unpleasant polarities (Thuerauf et al., 2009). This observation appears congruent with results obtained from selfrated psychological scales, especially in the case of disgust. For instance, Haidt et al. (1994) used the 32-item Disgust Scale (DS) and showed that women scored higher than men. Factor analyses revealed that DS taps three dimension of disgust: core disgust, animal reminder disgust and contamination-based disgust. It was then demonstrated that women scored higher than men for the three disgust dimensions (Olatunji et al., 2007). Using a specific body odor disgust scale (BODS), Tullio Luizza et al. (2017), revealed a reliable but small effect of gender indicating that women reported a higher level of body odor disgust than men. In another way , the CSS scores (Chemical Sensitivity Scale used to quantify affective reactions to odorous/pungent substances in the environment) were found to be higher in women than in men. However, studies carried out in different countries (Ferdenzi et al., 2013;Knaapila et al., 2017) found no sex-related differences in odor hedonic perception whatever the odorants and the country. These discrepancies may be due to several factors, especially the experimental conditions (i.e., type and number of odorants, type of rating scales,...).
Concerning the odorant properties, the perceived intensity, which is related to the stimulus concentration, is an important determinant of hedonic estimation and can reveal gender differences. For instance, women rated several concentrations of pyridine as more unpleasant than men (Broman and Nordin, 2000). These findings were later confirmed with three pyridine concentrations (Olofsson and Nordin, 2004). Additionally, Croy et al. (2017) showed a significant correlation between olfactory threshold and disgust ratings, but only in men. Furthermore, the experimental context can strongly influence many aspects of olfactory perception such as detection threshold (e.g., Rabin and Cain, 1986) discrimination (e.g., Jehl et al., 1995) and hedonicity (e.g., Djordjevic et al., 2008). First, considering repetitive exposure to three pleasant odorant stimuli (coconut, aloe and flowers), Triscoli et al. (2014) demonstrated that perceived pleasantness decreased from the first five presentations in men whereas it remained stable in women. Interestingly, the reverse occurred for "wanting" ratings (defined as "the wish to be further exposed to the same olfactory stimulus") that remained stable in men while it decreased during repetitive stimulations in women. Second, it has been shown that verbal labels associated to odors can influence olfactory perception (Herz and von Clef, 2001). Using odorants from the Sniffin'Stick Identification Test (Hummel et al., 1997), a study compared hedonicity before and after the presentation of an odor label (Seo et al., 2008). Results revealed significant changes for specific odors of cinnamon, rose, cloves and licorice in women and for specific odors of shoe leather, pineapple, lemon in men. However, for coffee odor the effect of odor label was observed in both men and women. With

Broman and Nordin, 2000
Data not available 1 non-specific odor Aversive detection thresholds Women present an odor aversion at lower concentrations (i.e., lower aversive detection threshold) than men, and they evaluate concentrations above this threshold as more unpleasant.
Platek In children, boys assess female sweat odor as more unpleasant than girls. In teenagers, girls rate male sweat odor to being more unpleasant than boys, and caramel odor as more pleasant than boys.
In adults, women rate both male and female body odors as more unpleasant than men. Homosexual men prefer body odors from homosexual men while both heterosexual men and women as well as homosexual women prefer body odors from heterosexual men.
For patchouli oil, male subjects showed a significantly higher odor pleasantness than female.
Triscoli et al., 2014 9/9 adults 3 non-specific odors Visual analog scale from "not at all pleasant" to "very pleasant" From the 5th presentation, the perceived pleasantness decreases in men while it remains constant in women. an other paradigm, Seo et al. (2009) asked participants to rate the emotional responses of dairy odors using 25 adjective pairs (e.g., common/rare, natural/artificial, usual/unusual, attractive, denial,...) as well as the hedonicity using a line scale. Results showed that both men and women preferred dairy odors that were characterized as more "comfortable, " "attractive " and "fragrant." However, women preferred odors characterized as more "bright, " "faint, " "warm, " "common" and "intellectual, " while men preferred odors characterized as more "refined, " "familiar" and "manlike." Finally, several comparative exposure conditions have not been yet explored. For instance, the hypothesis that gender differences in odor hedonic perception could be related to the nostril stimulated Disjkersrhuis et al., 2002) would be worth investigating. Additionally, further studies are needed to compare ortho versus retronasal stimulations or to examine the impact of the odorant qualities. For instance, the impact of trigeminal component of odorants (Brand, 2006) has never been explored, except in the study of Wallrabenstein et al. (2015) suggesting that the activation of VN1R1 by hedione might play a role in sex-specific modulation of hormonal secretion in humans. More specifically, because of the importance of pleasantness in food consumption, research should explore the role of food composition in sensory attributes evaluation. As a precursor, the work of Rosa et al. (2020) showed that women exhibited a greater ability than men to detect pleasantness of an odor rich in free fatty acid such as the mullet roes.

Psychophysiological and Imagery Studies
A number of studies have used psychophysiological measures (SCR, ECG, EEG, or ERPs) in response to pleasant and unpleasant odors (e.g., Alaoui-Ismaïli et al., 1997a;Glass et al., 2014). Independently from gender, results showed a more marked effect with unpleasant odors compared to pleasant odors. Surprisingly, gender differences in odor hedonic perception using psychophysiological measures are poorly documented, contrary to psychometric measurements. Some studies included only women (e.g., He et al., 2014He et al., , 2016Pichon et al., 2015) or did not consider sex differences (e.g., Alaoui-Ismaïli et al., 1997b;Bensafi, 2002;Djordjevic et al., 2008;Ferdenzi et al., 2017). Some other works compared men and women in skin conductance, heart rate responses or reaction time using pleasant and unpleasant odors (Brand and Jacquot, 2001;Bensafi et al., 2002a,b;Boesveldt et al., 2010). Data revealed no gender-related differences according to the hedonic valence of odorants. By contrast, in an other study (Hoffmann-Hensel and Freiherr, 2016), only men showed differences related to the pleasantess of odors (orange, cherry, vomit and spoiled fish) in skin conductance responses. Using ERPs with the bimodal odor of peppermint oil, Lundström and Hummel (2006) demonstrated that, although there were no differences in hedonic ratings between men and women, a sexrelated difference occurred in hemispheric responses, i.e., women expressed larger amplitudes and longer latencies in the left hemisphere while the same phenomenon was found in the right hemisphere in men. Another study (Olofsson and Nordin, 2004) conducted with three concentrations of pyridine indicated that the amplitude and the latency of P2/P3 component at Cz position were, respectively, larger and shorter in women than in men.
For several years, research concerning brain activity in olfactory perception, especially in hedonic perception, seeks to determine cerebral areas activation and to understand the underlying mechanisms (Zou et al., 2016) that are always currently discussed (Ruser et al., 2021). However, few studies have considered gender, probably because the number of participants recruited was small. Among them, the study of Royet et al. (2003) reported differential activation in men and in women, particularly in the left orbitofrontal cortex.

SEX DIFFERENCES IN HEALTHY POPULATIONS: BODY ODORS
There is evidence that body odors are involved in human communication and contribute to social interactions in daily life. Indeed, body odors carry important information relating to emotional state (de Groot et al., 2015), mate selection and attractiveness (Singh and Bronstad, 2001) as well as hormonal state (Preti et al., 1986(Preti et al., , 2003. Body odors derive from volatile compounds present in urine, vaginal secretions or breath and produced by degradation of bacteria on human skin (e.g., axillary zone, feet,...). These odors are mainly composed of aldehydes, ketones and carboxylic acid (Natsch and Emter, 2020). In the current society, body odors are usually perceived as unpleasant although interindividual differences have been reported, particularly between men and women.
In a pioneer study (Doty et al., 1975), vaginal secretions were collected from four women during 15 consecutives menstrual cycle and the odor pleasantness was rated by both men and women. Results clearly showed that women assessed the odors as more unpleasant than men. In the same way (Doty et al., 1978), hedonic valence of male and female axillary odors were rated by men and women and no difference was observed. In contrast (Doty et al., 1982), when both men and women rated male and female breath odors, the breath odor of males were rated as more intense and less pleasant than the breath odor of females. In addition, women gave significant lower pleasantness ratings to the breath odors than men. Interestingly, both men and women assigned the breath odor to the correct gender class and an inverse relation between breath odor intensity and pleasantness was observed. Concerning body odors, usually collected using T-shirt, cotton or gauze pieces placed under armpits, it appears that women assessed their own body odor as less pleasant than did men toward theirs (Platek et al., 2001). A comparative study (Stevenson and Repacholi, 2003), examined the hedonic estimation of body odor from adults by children, teenagers and adults. In children, boys rated women body odors as more unpleasant than girls while in teenagers, girls rated men body odors as more unpleasant than boys. In adults, women rated both male and female body odors as more unpleasant than men.
However, several factors could influence the hedonic estimation of body odors. For instance, the hypothesis that sexual orientation of donor could induce changes in pleasantness rating has been tested (Martins et al., 2005). Results indicated that homosexual men preferred body odors from homosexual men while both heterosexual men and women as well as homosexual women preferred body odors from heterosexual men. In another way (Mutic et al., 2016), a comparative analysis assessed the pleasantness of body odors collected during an intensive physical exercise and showed that women body odor was globally rated as less pleasant than those of men without sex-related differences.
Among body odors, androstenone is probably the most studied in experimental research. Androstenone is a steroid hormone commonly found in sweat and urine of male mammalians. In humans, androstenone is present in axillary regions in greater quantity in men than in women (Gower et al., 1985). Because of a large genetic polymorphism of genes encoding the olfactory receptors (ORs) among human's population, many people and mostly men cannot detect androstenone (Wysocki and Beauchamp, 1984;Bremmer et al., 2003). Like other smells, the pleasantness of androstenone is context-dependent: perceived as unpleasant when associated with urine and sweat or perceived as pleasant when associated with floral cues (Hasin-Brumstein et al., 2009). Surprisingly, although androstenone is widely studied, few publications have been dedicated to gender differences. Elsewhere, several experiments considered additional specific parameters that are possibly related to androstenone pleasantness. For instance (Lübke and Pause, 2014), in men with higher testosterone level, androstenone was perceived as more unpleasant than in men with lower testosterone level. Similarly, in women with higher estradiol level, androstenone was perceived as more unpleasant than in women with lower estradiol level. In another way, it has been suggested (Knaapila et al., 2012) that specifically in women, hedonic rating of androstenone is related to sexual intercourse experiences. Indeed, androstenone is rated as more pleasant by women who had a sexual intercourse experience with at least one partner, compared to those who reported never had sexual intercourse. Finally, the unique study comparing the hedonic perception of androstenone in relation to gender  found no sex differences for any of the perceptual variables, including pleasantness.

SEX DIFFERENCES IN PSYCHIATRIC DISORDERS
Olfactory deficits have been reported in psychiatric diseases, particularly in schizophrenia, depressive and bipolar disorders (e.g., Atanasova et al., 2008;Turetsky et al., 2009;Lahera et al., 2016;Brand and Schaal, 2017;Kamath et al., 2017;Kiparizoska and Ikuta, 2017). Patients usually presented lower olfactory sensitivity, discrimination and identification scores than healthy people (e.g., Postolache et al., 2002;Chen et al., 2018). Moreover, it has long been shown that in these diseases, olfactory impairment is stage-dependent and treatment-dependent in these diseases as it has been shown for a long time (Gross-Isseroff et al., 1994;Sirota et al., 1999). Olfactory dysfunction could also be disease-specific as suggested in a comparative study between bipolar, depressive and schizophrenic patients (Li et al., 2021). However, some brain areas such as amygdala, hippocampus, insula, anterior cingulate cortex and orbitofrontal cortex, involved in the above-mentioned psychiatric disordersespecially mood disorders -are linked to the olfactory system in relation to odor hedonic perception (e.g., Soudry et al., 2011). This relationship warrants further experimental studies on odor pleasantness ratings.
In schizophrenia, both men and women overevaluated the pleasantness of isoamyl acetate at high concentrations and underrated the lower concentrations compared to healthy subjects (Kamath et al., 2013). By contrast, another study (Walsh-Messinger et al., 2018) showed that schizophrenic men rated pleasant odors (orange, apple, mint,...) less pleasant than healthy men whereas no difference was observed between healthy and schizophrenic women. When studies related to this topic focused on gender differences in patients with schizophrenia, some of them showed that schizophrenic men rated pleasant odorants as more pleasant than schizophrenic women and healthy participants whatever the concentration level (Moberg et al., 2003;Robabeh et al., 2015). Other studies found no gender differences in odor hedonic judgment (Hudry et al., 2002;Lui et al., 2020).
In depression, the disease effect on odor hedonic ratings is not clearly demonstrated (e.g., Lombion-Pouthier et al., 2006;Swiecicki et al., 2009;Atanasova et al., 2010). When gender effect was considered in depression (Clepce et al., 2010), no difference between men and women was found in odor hedonic judgment. Similar data were obtained in bipolar disorder patients (Cummings et al., 2011) although the authors noted that women tended (i.e., in a non-significant way) to rate pleasant and unpleasant odors as more pleasant than men.

CONCLUSION
Contrary to sensitivity, identification or discrimination, this overview tends to prove that not such a clear differences occur in odor hedonic estimation between men and women. Gender effect appears more marked for body than for non-specific odors and is almost never reported in psychiatric diseases. It must be noted that no published study has yet focused on gender effect in odor hedonic estimation in neurodegenerative diseases (Alzheimer, Parkinson,...) while olfactory capabilities have been largely assessed (e.g., Doty, 2012;Woodward et al., 2017). When differences are reported, albeit not systematic, it seems that women overrated pleasant odors and underrated unpleasant odors compared to men. In accordance with the view that the hedonic processing of odor stimuli is an emotional rather than an analytical task and because sex differences occur in emotion (Wester et al., 2002), gender should logically affect clearly both the polarity and the extremity of the odor hedonic estimates. Thus, these findings regarding odor hedonic perception are rather surprising insofar as women are considered more responsive to emotional stimuli than men (Lithari et al., 2010) including auditory stimuli such as music and sounds, emotional words or emotional tones of voice (e.g., Schirmer et al., 2002Schirmer et al., , 2005Bachorowski and Owren, 2003) visual stimuli such as pictures and films (see Cahill, 2006 for a review) or danger-related stimuli (Williams and Gordon, 2007).
Additionally, except for pyridine and androstenone, there is not enough consistent data to draw conclusions on particular odorants. Moreover, most of studies focus on several parameters and rarely only on hedonic estimation. Thus, the odorants, tasks and participant characteristics are generally heterogeneous from one study to another (Clepce et al., 2014) which induces discrepancies and makes an overall conclusion difficult.
The main hypothesis related to this tenuous difference between men and women probably concerns the great intraand inter-variability in hedonic responses to odors. Indeed, numerous factors can influence odor perception (Greenberg et al., 2013) and certainly can influence the hedonic estimates and leveled the scores into groups of participants. Interindividual differences can occur in relation to sensitivity, familiarity, experience and memory toward the odorant stimulus (e.g., Delplanque et al., 2008), as well as in relation to age, culture, eating habits, medication and personality traits. Intraindividual differences can occur in relation to physiological states (e.g., arousal, weariness, stress, hunger,...), hormonal states (e.g., Trellakis et al., 2011) and psychological states (attention, motivation, mood,...). Besides, intra-and inter-individual differences in odor preferences are not specific to humans and have been reported in animals (e.g., Jagetia et al., 2018) in rigorously homogeneous cohorts. This suggests that odor hedonic perception is a highly complex response that prevents direct and relevant comparisons between men and women in the general population. Besides, current research demonstrates that individual pleasantness of odors is not detectable in fMRI (Ruser et al., 2021), probably because this process is coded in joint networks (Mantel et al., 2019). Thus, future research in this field -especially food studies, toxicological or specific exposure studies, quality of life studies, must cross several factors in a multidimensional approaches among which gender could be included in order to contribute to the knowledge of the human affective neurosciences (Becker et al., 2019).