Taste-induced facial responses in black-handed spider monkeys (Ateles geoffroyi)

Taste-induced facial expressions are thought to reflect the hedonic valence of an animal's gustatory experience. We therefore assessed taste-induced facial responses in six black-handed spider monkeys (Ateles geoffroyi) to water, sucrose, caffeine, citric acid and aspartame, representing the taste qualities sweet, bitter, and sour, respectively. We decided not to include salty-tasting substances as the concentrations of such tastants found in the fruits consumed by spider monkeys are below their taste preference threshold. We found that the monkeys displayed significant differences in their facial responses between substances, with significantly higher frequencies of licking, sucking, closed eyes, tongue protruding, mouth gaping and lip smacking in response to sucrose, a presumably pleasant stimulus. The response to caffeine and citric acid, in contrast, yielded the lowest frequencies of these behaviors, but the highest frequency of withdrawals from the stimulus, suggesting these substances are perceived as unpleasant. Lip stretching, a newly described behavior, was performed significantly more often in response to caffeine than to any other substance, suggesting an association with the response to bitter taste. The facial response to the artificial sweetener aspartame was generally similar to the response to water, corroborating the notion that Platyrrhines may be unable to detect its sweetness. Overall, the present study supports the idea of similarity of taste-induced facial responses in non-hominoid primates and humans, suggesting these displays to be evolutionarily conserved across the primate order.


Introduction
The hedonic valence of a taste stimulus is known to elicit characteristic facial expressions in humans (Steiner and Glaser, 1995). Accordingly, we are instinctively able to appropriately associate certain facial expressions with the pleasantness or unpleasantness and the palatability of food. Even if social convention may repress the intensity with which human adults express their affective reactions to taste stimuli (Berridge, 2000), human newborns are known to exhibit these since very early ages, responding particularly to sweet and bitter tastes (Steiner, 1974).
Facial expressions are also important when attempting to interpret how animals perceive a given taste stimulus (Steiner et al., 2001). A number of mammal species ranging from non-human primates to horses, cats, mice and rats have been reported to show different facial responses towards certain taste stimuli (Grill and Norgren, 1978;Steiner and Glaser, 1995;Steiner et al., 2001;Ueno et al., 2004, Jankunins and Whishaw 2013, Hanson et al., 2016Dolensek et al., 2020). In non-human primates, for example, previous studies have revealed distinct facial responses to at least two taste qualities (Steiner et al., 2001), with rhythmic tongue protrusions being associated with sweet taste and mouth gapes being associated with bitter taste. Not surprisingly, the majority of these studies so far focused on our closest relatives, the great apes. However, in order to better understand the evolutionary origin and the adaptive value of taste-induced facial expressions, it is important to also study non-hominoid primates. Current evidence suggests that taste-induced facial responses may serve a communicatory function, providing honest signals about the palatability of food to conspecifics (Steiner et al., 2001;Jankunis and Whishaw, 2013). However, considering that primate taxa differ markedly in the musculature that forms the basis and is the prerequisite for unequivocal facial expressions (Burrows, 2008), it is unclear when taste-induced facial responses evolved along the primate lineage and how universal, i.e. evolutionarily conserved they are across this mammalian order. Further, we cannot exclude the possibility that taste-induced facial expressions may be purely motivational rather than communicative (Andrew, 1963).
Additionally, previous studies showed that facial responses in nonhuman mammals may be a more sensitive indicator of the palatability and hedonic experience, i.e. the perceived pleasantness or unpleasantness of food compared to the measurement of consumption (Hanson et al., 2016). This, in turn, may be useful for the development of novel foods for companion animals and for assessing the causes underlying food-related welfare problems in captive animals. Furthermore, as basically all food consumed by primates comprises a mixture of taste qualities (sweet, sour, bitter, salty, and umami), the study of facial responses may provide further insight into the trade-offs that e.g. frugivorous primates have to make with regard to the attractive and aversive properties of different taste qualities and the ripeness of fruits.
Spider monkeys are a particularly useful primate species to study in this respect as they are highly frugivorus (Gonzalez-Zamora et al., 2009) and their responsiveness to all five taste qualities has been studied in detail. Using two-bottle preference tests, they have been reported to display a clear preference for substances tasting sweet (Laska et al., 1996;Nicklasson et al., 2018) or umami (Laska and Hernandez Salazar, 2004;Laska et al., 2008), a clear aversion to substances tasting bitter (Laska et al., 2009), and concentration-dependent preferences or aversions to substances tasting sour (Laska et al., 2000(Laska et al., , 2003 or salty (Laska and Hernandez Salazar, 2004;Laska et al., 2008).
Therefore, the aim of the present study was to investigate tasteinduced facial responses in black-handed spider monkeys (Ateles geoffroyi). In addition to monomolecular substances representing the classical taste qualities of sweet, bitter and sour, we also included aspartame, a substance which humans describe as having a sweet taste and, at high concentrations, a bitter side taste. Thus, we additionally assessed the suitability of taste-induced facial expressions to elucidate the hedonic valence of this artificial sweetener in spider monkeys. We decided not to include sodium chloride, representing the classical taste quality of salty, into the present study as the concentrations of this tastant found in the fruits consumed by spider monkeys are below their taste preference threshold (Laska and Hernandez Salazar, 2004). Accordingly, it was unlikely that physiological concentrations of this tastant would elicit facial responses in this primate species.

Animals and housing
The present study included six adult black-handed spider monkeys (Ateles geoffroyi) maintained at the research station UMA Doña Hilda Á vila de O'Farrill of the Universidad Veracruzana near Catemaco, Mexico. All individuals were group-housed in a series of roofed outdoor enclosures connected to each other via sliding doors. Accordingly, they were exposed to natural environmental conditions with regard to ambient temperature, light, and relative humidity. Testing of each animal was performed in an empty enclosure separately from the remaining individuals to prevent distraction and competition. All animals were already used to being temporarily separated and voluntarily entered the empty enclosure upon opening of the corresponding sliding door and calling them by name. Further, all animals had participated in previous studies on taste perception and were thus familiar with the presentation of sweet, sour, and bitter taste stimuli.
The monkeys were fed a wide variety of fresh seasonal fruits and vegetables once a day. As spider monkeys fulfill their water requirements by consuming juicy fruits rather than drinking from open water sources, no water deprivation schedule was adopted. The group of subjects was composed of three females and three males aged between twelve and fifteen years.
The experiments reported here comply with the American Society of Primatologists' Principles for the Ethical Treatment of Primates, as well as with current Swedish and Mexican animal welfare laws. The study was performed according to a protocol approved by the Ethical Board of the Federal Government of Mexico's Secretariat of Environmental and Natural Resources (official permits no. 09/GS-2132/05/10).

Taste stimuli
Four different taste substances were presented to the spider monkeys. Three of the substances represent different taste qualities. Sucrose (CAS# 57-50-1) was used as a prototypical sweet taste stimulus, caffeine (CAS# 58-08-2) was used as a bitter taste stimulus and citric acid (CAS# 77-92-9) was used as a sour taste stimulus. The fourth stimulus, aspartame (CAS# 22839-47-0), was included to better understand the spider monkey's perception of this artificial sweetener which has been reported to be perceptible for only some, but not all primate species tested so far  and which is known to elicit a bitter side taste in humans when presented at high concentrations (Schiffman et al., 1995). Tap water was used as a neutral stimulus and thus, a control. Sucrose, caffeine and citric acid were presented at a concentration of 200 mM, 100 mM, and 500 mM, respectively, in order to provide taste stimuli that are clearly detectable for the animals, but not overwhelmingly sweet, bitter, and sour, respectively (Laska et al., 1996(Laska et al., , 2009Laska et al., 2000). Aspartame was used at a concentration of 100 mM as a previous study reported that already a concentration of 20 mM elicited either a preference or a rejection in spider monkeys in a two-bottle test when water was the alternative (Pereira, 2020).

Experimental procedure
Each taste stimulus was presented to the animals using a glass dropper (Fig. 1). Prior to any critical tests, the animals were made familiar with this device in a series of 20 trials across five days using sucrose as an attractive and thus motivational stimulus. Once the animals were familiar with the glass dropper, each taste stimulus was presented for a total of ten times to each individual. During the critical tests, a maximum of four trials were performed per day with each individual and care was taken to present a given stimulus not more than once per day and to have inter-trial intervals of at least ten minutes. Accordingly, a total of 13 testing days were needed to complete all trials. Testing took place in the morning, prior to the feeding of the animals. The amount of food offered daily was such that leftovers were still present in the morning. Thus, it was unlikely that ravenous appetite affected the animals' ingestive behavior in the tests.
The sampling of the taste stimuli was recorded with a video camera (Sony Handycam HDR-CX405), starting with the presentation of the dropper to the animal until the dropper (2.5 mL) was emptied or until the animal refused to sample more substance for more than five seconds. The four taste stimuli and the water control were tested in a pseudorandomized order to prevent a potential result bias due to order effect. This was true both within and between testing days. A maximum of four trials were performed per day with each individual and care was taken to present a given stimulus not more than once per day and to have inter- trial intervals of at least ten minutes. Once all trials were completed, two human coders experienced in working with spider monkeys were asked to analyze the footage and record the frequency of selected facial expressions and facial motor patterns based on previous studies. They were, however, not involved in the statistical analysis of these data.
The following behaviors were considered: Licking (ingestive): The monkey is ingesting the liquid from the dropper by licking the tip of the device.
Sucking (ingestive): The monkey is ingesting the liquid from the dropper by putting its lips around the device's tip and sucking upon it.
Sniffing dropper: The monkey is sniffing the dropper from a distance < 2.5 cm.
Eyes open > 50 %: The eyes of the monkey have a rounded shape and > 50 % their area is visible.
Eyes open < 50 %: The eyes of the monkey are not fully open and < 50 % of their area is visible.
Eyes closed: The eyes of the monkey are closed and thus not visible.

Flat tongue protrusion:
The tongue is directed out of the mouth on the horizontal plane, with no upward or downward bending, but not in contact with any object.
Tongue protrusion angled downward: The tongue is directed out of the mouth and angled in a downward direction, but not in contact with any object.
Tongue protrusion angled upward: The tongue is directed out of the mouth and angled in an upward direction, but not in contact with any object.
Repetitive tongue protrusion: The tongue is repeatedly directed out of the mouth at a high frequency, but not in contact with any object.
Tongue protrusion gape: The tongue is directed forward out of the mouth while the monkey has its mouth wide open in a gape.
Mouth gape: The monkey has its mouth wide open without protruding its tongue.
Lip stretching: The monkey stretches its lips backwards, showing its teeth. The mouth can be slightly open.
Lip smacking: The monkey brings its lips together repetitively in a savoring-like motion.
Wrinkles nose: The monkey wrinkles its nose, possibly elevating its upper lip.
Withdrawal from dropper: After having sampled the taste stimulus, the monkey abruptly moves away from the dropper.
To exclude the possibility of experimenter bias, both coders were unaware of which taste stimulus was being presented to the animal in each video. The selected facial expressions had previously been reported to be indicative of the pleasantness or unpleasantness of taste stimuli as perceived by other non-human primate species (Steiner et al., 2001).

Data analysis
As a preliminary validation of the data and to verify a sufficient level of inter-rater agreement between the two coders' independent analysis of the video footage, an Intraclass Correlation Analysis (ICC) was run. For the data to be considered reliable and included in the study, a correlation coefficient (α) of at least 0.7 was required. The recordings for each substance which achieved sufficient inter-rater agreement were further analyzed for significant differences in facial responses between the different stimuli. The Related-Samples Wilcoxon Signed Rank test for pairwise comparisons was used to assess differences in the frequency of occurrence of the selected behaviors between the different stimuli. All statistics were performed using IBM SPSS Statistics (version 26) and, if not otherwise stated, an alpha level of 0.05 was used.

Results
The correlation coefficients obtained from the ICC analysis ranged from 0.7 to 0.92. Accordingly, all data met the criterion of sufficient inter-rater agreement and were included in the statistical analyses.
For the frequency of licking, significant differences were found between all five taste stimuli (Wilcoxon, p < 0.01 with all ten pairwise comparisons, 0 < T < 771, N = 6). Licking occurred significantly more often with sucrose, followed by the water control, aspartame, caffeine and citric acid (Fig. 2, Panel a).
Regarding sniffing frequency, all taste stimuli differed significantly from each other (Wilcoxon, p < 0.01, 29 < T < 324, N = 6) with the exception of the pairwise comparisons between sucrose and citric acid (Wilcoxon, p > 0.05, T = 52.5, N = 6), and aspartame and caffeine (Wilcoxon, p > 0.05, T = 87, N = 6) (Fig. 2, Panel b). Sniffing frequency was highest with the water control and lowest with sucrose. A trend rather than a significant difference was found for the comparison between aspartame and citric acid (Wilcoxon, p = 0.056, T = 37, N = 6).
The frequency of closed eyes differed significantly between aspartame and caffeine, aspartame and citric acid, aspartame and sucrose, sucrose and caffeine, sucrose and citric acid, and sucrose and tap water (Wilcoxon, p < 0.05 with all six comparisons, 0 < T < 252, N = 6) ( Fig. 3, Panel b). No other significant differences were found (Wilcoxon, p > 0.05, 30.5 < T < 104, N = 6). The highest frequency of this behavior was found with sucrose, followed by aspartame. Eyes were closed the least with citric acid.
The frequency of flat tongue protrusions differed significantly between all five taste stimuli (Wilcoxon, p < 0.01, 1 < T < 799, N = 6) with the exception of the pairwise comparison between caffeine and citric acid, for which a trend rather than a significant difference was found (Wilcoxon, p = 0.064, T = 70, N = 6) (Fig. 4, Panel a). Flat tongue protrusions occurred most often with sucrose, followed by the water control and aspartame. This behavior was least frequent with caffeine and citric acid.
The frequency of tongue protrusion gapes differed significantly between all five taste stimuli (Wilcoxon, p < 0.01 with all ten comparisons, 0 < T < 728.5, N = 6) (Fig. 4, Panel b). Tongue protrusion gapes occurred most often with sucrose, followed by the water control and then aspartame. Tongue protrusion gapes occurred the least with caffeine and citric acid.
The frequency of lip stretching differed significantly between all five taste stimuli (Wilcoxon, p < 0.05, 0 < T < 228, N = 6) with the exception of the pairwise comparison between aspartame and citric acid (Wilcoxon, p > 0.05, T = 52, N = 6). Trends rather than significant differences were found for the pairwise comparisons between aspartame and tap water, citric acid and tap water, and sucrose and tap water (Wilcoxon, 0.05 < p < 0.1, 6 < T < 45, N = 6) (Fig. 5, Panel a). Lip stretching occurred the most with caffeine, followed by citric acid.
The frequency of lip smacking differed significantly between aspartame and caffeine, aspartame and sucrose, sucrose and caffeine, and sucrose and citric acid (Wilcoxon, p < 0.05 with all four comparisons, 112.5 < T < 347.5, N = 6) (Fig. 5, Panel b). Trends rather than significant differences were found for the pairwise comparisons between aspartame and citric acid, caffeine and tap water, and citric acid and tap water (Wilcoxon, 0.05 < p < 0.1 with all three comparisons, 189 < T < 385.5, N = 6). No significant differences were found for the remaining pairwise comparisons (Wilcoxon, p > 0.05, 188 < T < 340.5, N = 6). Lip smacking occurred most often with sucrose, followed by the water control and aspartame. Lip smacking occurred least often with citric acid and caffeine.
Regarding withdrawals from the dropper, frequencies differed significantly between all five taste stimuli (Wilcoxon, p < 0.01, 0 < T < 465, N = 6), with the exception of the pairwise comparisons between caffeine and citric acid (Wilcoxon, p > 0.05, T = 223.5, N = 6), and aspartame and tap water (Wilcoxon, p > 0.05, T = 85.5, N = 6). The pairwise    comparison between aspartame and caffeine revealed a trend rather than a significant difference (Wilcoxon, p = 0.057, T = 168, N = 6). The spider monkeys withdrew from the dropper most often with citric acid, caffeine and aspartame, and withdrew the least often with sucrose and the water control (Fig. 6).
The following behaviors yielded only few significant differences in frequency between the different stimuli and are therefore not displayed as graphs: Sucking frequency was significantly higher with sucrose than with caffeine (Wilcoxon, p < 0.01, T = 0, N = 6). No other pairwise comparison between any other of the taste stimuli revealed a significant difference (Wilcoxon, p > 0.05, 0 < T < 12, N = 6).
The frequency of eyes open more than 50 % differed only in the pairwise comparisons between aspartame and citric acid, caffeine and water, and citric acid and water (Wilcoxon, p < 0.01 with all three comparisons, 125 < T < 379, N = 6). No significant differences were found for the remaining pairwise comparisons (Wilcoxon, p > 0.05, 90 < T < 297.5, N = 6). The highest frequencies of this behavior were recorded for water, and the lowest for caffeine and citric acid.
Regarding mouth gapes, only sucrose and caffeine differed significantly in the frequency of this behavior (Wilcoxon, p < 0.05, T = 0, N = 6). No other significant differences were found in the frequency of mouth gapes between taste stimuli (Wilcoxon, p > 0.05, 2 < T < 31.5, N = 6).
No significant difference was found between the frequencies of nose wrinkles across the five taste stimui (Wilcoxon, p > 0.05 0 < T < 4.5, N = 6, with all ten comparisons). Fig. 7 shows two typical facial expressions in response to a sweet-and to a bitter-tasting stimulus, respectively.

Response to sucrose
The facial response to sucrose was characterized by the highest frequencies of licking, sucking, having the eyes closed or open less than 50 %, tongue protruding, mouth gaping and lip smacking compared to the four other taste stimuli included in the present study. High-frequency licking and sucking have been previously associated with the response to sweet taste in several non-human primate species (Steiner and Glaser, 1984). No studies so far have described a fully closed state of the eyes in response to any particular taste quality. Nonetheless, "slightly closed eyes" have been described as being part of the typical response to sweet taste in several non-human primate species (Steiner and Glaser, 1984). Although the behavior "slightly closed eyes" is not unequivocally defined by Steiner and Glaser (1984), it seems plausible to assume that it would be comparable to a state of the eyes in which less than 50 % of the area of the eyeball is visible, and thus comparable to the behavior categorized in the present study as eyes open less than 50 %. However, the fact that water triggered approximately the same frequency of eyes open less than 50 % (comparable to the previously described "slightly closed eyes") as did sucrose allows us only to conclude that this behavior may indicate either that the taste is perceived as pleasant (which can be considered as a positive hedonic response) or that it is perceived as neither pleasant nor unpleasant (which can be considered as a neutral hedonic response). Our results also suggest that higher frequencies of flat tongue protrusions, tongue protrusion gapes and mouth gapes, are characteristic for a facial response to sweet taste in spider monkeys. Previous studies have reported tongue protrusions to be associated with the response to sweet taste in chimpanzees (Steiner et al., 2001;Ueno et al., 2004) and rhesus macaques (Ueno et al., 2004), which is congruent with the results of the present study. However, mouth gaping has sometimes also been associated with the response to bitter taste Glaser, 1984, 1995;Steiner et al., 2001), which is in contrast to our results. The fact that lip smacking occurred most often with sucrose is in line with the notion that this behavior also occurs typically in response to sweet taste, as reported by previous studies in both human and non-human primates (Steiner and Glaser, 1995;Steiner et al., 2001). The response to sucrose was also characterized by the lowest frequencies of withdrawals from the dropper, suggesting that this sweet-tasting substance is indeed perceived as an attractive stimulus by the spider monkeys as it is plausible to assume that an animal would avoid a pleasant stimulus less often compared to an aversive stimulus. This notion is also supported by findings from two-bottle preference tests in which spider monkeys have been reported to clearly prefer sweet-tasting substances, including sucrose, over water as an alternative stimulus (Laska et al., 1996).

Response to caffeine
The facial response to caffeine was characterized by a pattern of facial expressions opposite to that found with sucrose. Accordingly, low frequencies of licking, sucking, tongue protruding, mouth gaping and lip smacking were found in response to bitter taste. Conversely, the highest frequencies of lip stretching were recorded in response to caffeine, suggesting that this behavior may be associated with the response to bitter taste. As this behavior has not been described in previous studies on taste-induced facial responses in non-human primates, it remains to be determined whether this facial response may be specific for spider monkeys or perhaps typical for non-human primates in general. Although it is known that chimpanzees possess the necessary musculature to perform lip stretching, no similar display has been described so far (Vick et al., 2007). In human subjects, a display comparable to lip stretching described as "lips exposing teeth with stretched lip corners" was reported to be typically associated to anger (Kohler et al., 2004). This expression is also associated to painful experiences in elderly human subjects (Hadjistavropoulos et al., 2002). Even though lip stretching has so far never been described as a taste-induced facial response, its resemblance to grinning is quite suggestive. As grinningretracting the lips while baring the teethis known to be sometimes triggered in situations of potential danger in humans and rhesus monkeys (Landis and Hunt, 1936;Hinde and Rowell, 1962), it seems plausible that a similar lip stretching response could be evoked in contexts perceived as an alarm for risk e.g. the sampling of an aversive-tasting food. The fact that the highest frequencies of withdrawals from the dropper were recorded in response to caffeine suggests that this stimulus is indeed perceived as aversive by the spider monkeys. This notion, too, is supported by findings from two-bottle preference tests in which spider monkeys have been reported to reject bitter-tasting substances, including caffeine, when water was presented as an alternative stimulus (Laska et al., 2009).

Response to citric acid
The facial response of the spider monkeys to citric acid was similar to that to caffeine. Accordingly, we found low frequencies of licking, sucking, tongue protruding, mouth gaping and lip smacking in response to this stimulus. The fact that, along with caffeine, the response to citric acid elicited one of the highest frequencies of withdrawals from the dropper, indicates that this taste stimulus was also perceived as aversive by the spider monkeys. It should be mentioned, however, that spider monkeys have been reported to display an inverted U-shaped function of preference in two-bottle preference tests with sour-tasting substances, including citric acid (Laska et al., 2000). This suggests that the attractive or aversive properties of citric acid are concentration-dependent. Consistent with the facial responses recorded in the present study, the concentration of citric acid used here (500 mM) was clearly rejected by the spider monkeys in the earlier study that employed two-bottle preference tests (Laska et al., 2000). It might therefore be interesting to assess the facial responses of spider monkeys to lower concentrations of citric acid in future studies to elucidate whether their switch from rejection (of high concentrations of citric acid) to preference (with low, but detectable concentrations of the sour tastant) is accompanied by corresponding changes in their facial expressions.

Response to aspartame
The facial response of the spider monkeys to aspartame often ranked intermediately between the response to water and the response to the presumably aversive stimuli, caffeine and citric acid. This was true for having the eyes open less than 50 %, protruding the tongue flat and protruding the tongue while in a gape. Taking water as a neutral stimulus and caffeine and citric acid as presumably aversive ones, the intermediate position of aspartame regarding these behaviors suggests that this substance may be perceived as either neutral or slightly aversive. The response to aspartame yielded similar frequencies of sniffing and lip smacking to those recorded in response to water, pointing towards a neutral hedonic valence of this substance as perceived by the spider monkeys. On the other hand, aspartame elicited significantly more withdrawals from the dropper than water did, suggesting that the former may not be perceived by the animals as being as neutral or tasteless as water. In any case, the facial responses of the spider monkeys to aspartame do not suggest that this substance would be perceived as an attractive, sweet-tasting stimulus. This notion is in line with the finding that spider monkeys did not display a consistent preference for this artificial sweetener in two-bottle preference tests (Pereira, 2020). Rather, some of the animals rejected aspartame when water was presented as an alternative stimulus.

Evolution of taste-induced facial responses
The origin and causation of facial expressions have been of scientific interest since more than a century. Charles Darwin already attempted to identify in primates and other mammals the potentially primordial traces of the facial displays expressed by humans (Darwin, 1872). While it is commonly accepted that facial expressions convey some sort of information about the individual who displays them (Andrew, 1963), no consensus has been established so far with regard to the function of taste-induced facial responses. It is still under debate whether taste-induced facial responses serve as social signals, mediating intraspecific communication (Steiner et al., 2001;Jankunis and Whishaw, 2013). The term signal implies that a receiver, e.g. a conspecific of the animal transmitting information using facial displays, is able to recognize and correctly interpret the information conveyed through the animal's face (Steiner et al., 2001). Alternatively, it is possible that facial expressions do not serve as communicative signals but that facial displays are purely motivational (see also Andrew, 1963). Nonetheless, if facial responses to taste stimuli do serve as social signals, they may be adaptive by allowing the signaling of e.g. the nutritional value or the palatability of a food item or, conversely, the toxicity of an item improper for ingestion in a given species (Ueno 2004). This idea is in line with our findings in the sense that Platyrrhines, which diverged from the hominoids about 40 million years ago, display taste-induced facial responses that are similar to the ones reported for human infants. The notion that the facial responses towards attractive and aversive taste stimuli appear to be evolutionarily conserved across primates suggests that they play an important role for intraspecific communication. However, this does not rule out the possibility that the facial movements typically associated with pleasant and aversive stimuli may have been selected simply as the mechanics involved in the sampling of such stimuli (Steiner et al., 2001), as e.g. tongue protrusions facilitate the ingestion of a pleasant stimulus such as sucrose and lip stretching aids in eliminating a stimulus perceived as aversive and potentially harmful such as caffeine. To corroborate the idea that taste-induced facial displays indeed serve a communicatory function in non-human primates, future studies should investigate whether e.g. an animal's food selection behavior is affected by observing the facial response of a conspecific towards the food in question.

Limitations of our study
Our use of sucrose as a sweet and thus attractive taste stimulus to familiarize the spider monkeys with the glass dropper and the experimental procedure raises the question whether this pre-exposure with only one of the taste stimuli used later on in the critical tests may have affected the animals' behavior and thus our results. For the following reasons we consider this as highly unlikely: firstly, all of our animals had participated in previous studies on taste perception and were thus familiar with the presentation of sweet, sour, and bitter taste stimuli. Secondly, the diet of spider monkeys includes a wide variety of sweet, sour, and bitter taste stimuli (Gonzalez-Zamora et al., 2009) and thus both animals only pre-exposed to sucrose and animals pre-exposed to sucrose, citric acid and caffeine would have had plenty of experience with all three taste qualities. For the same reason we consider it as highly unlikely that neophobia may have affected our results. Thirdly, and perhaps most importantly, there is both electrophysiological (Plata- Salaman et al., 1993Salaman et al., , 1995Scott et al., 1999) and behavioral (Laska et al., 1998(Laska et al., , 2003 evidence suggesting that human and nonhuman primates are monogeusic for each of these three taste qualities. In other words: if we had used a purely sweet-tasting stimulus other than sucrose for our familiarization trials in order to avoid pre-exposure with one of the stimuli used in the critical tests, then this would have made no difference as primates perceive e.g. all carbohydrates (such as sucrose, fructose, glucose, maltose, and lactose) as purely sweet and are incapable of distinguishing them based on their taste quality. The same is true for purely sour tastants such as citric acid and purely bitter tastants such as caffeine. Finally, behavioral studies on both primate and non-primate mammals suggest that taste-induced facial expressions are extraordinarily robust (Steiner et al., 2001) which is consistent with the notion that such facial expressions may serve as an honest signal of palatability for conspecifics.
Other possible limitations of our study include the low number of animals tested and the low number of taste stimuli used. Future studies should therefore aim at employing a higher number of animals and a wider variety of taste stimuli.

Conclusion
The present results suggest two distinct patterns of facial responses in spider monkeys to taste stimuli, one being associated with presumably pleasant and thus attractive stimuli and the other being associated with presumably unpleasant and thus aversive stimuli. The facial responses to pleasant stimuli include high frequencies of licking, sucking and tongue protruding as well as having the eyes open less than 50 %. The facial responses to aversive stimuli, in contrast, include low frequencies of licking, sucking and tongue protruding, as well as high frequencies of withdrawal from the taste stimuli. Based on this dichotomy of facial expressions, the results suggest that the artificial sweetener aspartame is probably not perceived as sweet by spider monkeys. Rather, it is most likely perceived as either neutral or as a slightly unpleasant taste, such as bitter, as indicated by the mentioned similarities in the recorded frequencies for some of the analyzed behaviors with those found for water and caffeine, respectively. These findings are in line with the notion that spider monkeys may not be able to detect the sweetness of aspartame (Pereira, 2020). This, in turn, supports the usefulness of analyzing taste-induced facial responses as a means to assess an animal's gustatory and hedonic experience of a taste stimulus. Future studies on facial responses to taste stimuli in non-human primates should include the presentation of systematically varying taste mixtures which mimic, for example, the change of taste quality in fruits across the process of ripening. Overall, the present study supports the idea that taste-induced facial expressions have been evolutionarily conserved within the primate order, potentially serving as a communicatory signal.

Author statement
Sofia Pereira: Conceptualization, Formal analysis, Investigation, Visualization, writing of both original draft and revised version.
Laura Teresa Hernandez Salazar: Conceptualization, Methodology, Investigation, Resources, writing of both original draft and revised version.
Matthias Laska: Conceptualization, Methodology, Formal analysis, Resources, writing of both original draft and revised version.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest
The authors report no declarations of interest.