Ekman and Friesen (1975) stated that when wanting to assess the truthfulness of someone’s expression, the feature that should be primarily paid attention to is the eyes. Baron-Cohen et al. (2001) suggested that a person’s state of mind can be decoded by looking at the widening and narrowing of the eyes. Lee and Anderson (2017) endorsed the idea that the eyes’ area reveals emotional states without the need to observe other parts of the face; and that no other facial feature reveals as much information about emotional processing as the eyes. Current studies have also highlighted the fact that facial expressions, including the perception of eye size, can serve as salient nonverbal signals to capture individuals’ attention (Carlson and Aday, 2018; Aday et al., 2023).

Conversely, Eisenbarth and Alpers (2011) observed that, when initially looking at a face, the first gaze tends to be directed toward either the eyes or the mouth. Blais et al. (2012) advocated that the most distinct facial gestures across expressions are found in the mouth area, and that this region provides the most relevant cues to recognize facial expressions. Salience is a key factor in faces with a smile, independently of the eyes (whether or not congruent) that accompany the mouth (Fernández-Martín et al., 2013; Calvo et al., 2016). Carbon (2020) proposed that not seeing the lower half of the face (the mouth and surrounding area) impairs emotional recognition to the point of misinterpreting happy, sad, or angry expressions as neutral.

Eisenbarth and Alpers (2011) addressed the question by subdividing it into specific emotions. According to these authors, sadness and anger are mostly noticed in the eyes. Attention is mainly directed toward the mouth in happy faces; and in fearful and neutral expressions, the eyes and mouth are equally important. Paradoxically, Barrett (2018) indicated that a clear indicator of fear are widely open eyes. Ekman and Friesen (1975) argued that, although fear is initially shown in the eyebrows, its intensity is expressed by the openness of the eyes. Notably, when contradictory stimuli are shown (i.e., happy mouth and sad eyes simultaneously, or vice versa), people choose to rely on the mouth (Dunlap, 1927). Recent studies have compared emotions in faces with and without surgical masks. Langbehn et al. (2022) showed that, when the mouth is hidden, the information about the emotion on the face shrinks. Tsantani et al. (2022) supported the idea that the intensity of facial expressions is reduced in all basic emotions—joy, fear, anger, sadness, disgust, and surprise (Ekman et al., 1969; Jack et al., 2012)—except anger, when the mouth is hidden. Kim et al. (2022) observed that, when compared with the eyes, the mouth was most salient for happiness, sadness, and anger; however, in the case of fear, the eyes achieved higher relevance.

Succinctly, empirical research suggests that both eyes and mouth are important for emotional decoding; that the information drawn from the face varies depending on the emotion; and that attention is paid to one or another part of the face depending on the emotion on display. Furthermore, it has been hypothesized that it is not the components of the face, but rather the areas (i.e., the center), that attract our attention. However, the more familiar the face, the higher the chance that attention will be drawn to the eyes (Royer et al., 2016). Other components of the face also claim visual attention [e.g., eyebrows are fundamental in anger (Ekman and Friesen, 1975)]. Therefore, it is not a simple eye vs. mouth question for all emotions.

Rather than analyzing which components or facial areas provide more information, some studies have focused on how those features work together in the face (Bartlett and Searcy, 1993; Xiao et al., 2014; Matsushita et al., 2015a). It has been suggested that faces could be recognized in a holistic way (Maurer et al., 2002) (as a unit rather than as a collection of components). However, emotions could be interpreted in different ways. For example, Tanaka et al. (2012) suggested that emotion recognition should be holistic when information is contradictory (such as a smiley face with angry eyebrows), but divided into different elements when all components point in the same direction (i.e., a happy face). Contrarily, Omigbodun and Cottrell (2013) argued that facial expression processing is essentially holistic, independent of the components. Curby et al. (2012) proposed that differences in the processing type are associated with the emotion types, negative emotions being less globally processed. Regarding the rating of attractiveness, facial proportions have been suggested to play a fundamental role in the holistic assessment of beauty in women (Ulrich et al., 2019).

Studies have addressed the relationship between eye size, perceived level of attractiveness and perceived intensity of emotions (Reis et al., 1990; Oosterhof and Todorov, 2009; Schmidt et al., 2012; Golle et al., 2013; Przylipiak et al., 2018). Regarding the three possible analyses (association between eye size and attractiveness, eye size and intensity, and attractiveness and intensity), research has mainly focused on the impact of eye size on attractiveness, based on the potential evolutionary advantage of larger eyes (increased eye size has been associated with improved visual perceptiveness, a key capacity for finding food and detecting predators in animals). Berry and McArthur (1985) proposed that large eyes, high eyebrows, and narrow chins generate baby-like appearances. Glocker et al. (2009) suggested that babies’ eye size is attractive due to evolution and that cuteness motivates caretaking behavior, which increases the child’s chances of survival. Larger eyes have also been linked to attractiveness in adulthood (Berry and McArthur, 1985). Although averageness has repeatedly been linked to attractiveness (Thornhill and Gangestad, 1993; Grammer and Thornhill, 1994; Scheib et al., 1999), larger eyes might signify health, which is an influential factor when choosing a sexual mate (Thornhill and Gangestad, 1999). When other parts of the face (e.g., nose, lips) are smaller—making the eyes more dominant—attractiveness rates increase (Przylipiak et al., 2018). On the flip side, effects that make the eyes more attractive have been shown to make them seem larger (Matsushita et al., 2015b). These studies suggest that eye size also has an influence on attraction in adulthood. Noticeably, the relationship between attractiveness and eye size has been analyzed using essentially positive emotions, mostly happiness (Wagner, 1990), and it has been observed that faces characterized by positive emotions are perceived as more attractive than faces with neutral or negative expressions (Reis et al., 1990; Oosterhof and Todorov, 2009; Schmidt et al., 2012; Golle et al., 2013; Ueda et al., 2016).

The link between attractiveness and intensity has only been found to be significant when positive emotions are considered. The study by Golle et al. (2013) observed that happier faces were perceived as more attractive than faces reflecting other emotion types. Another study aiming to compare the intensity and attractiveness levels of faces with happy, sad, and neutral expressions observed that more intense happy faces were considered more attractive than less intense happy faces (Ueda et al., 2016). This work also observed that sad faces evaluated as more intense were not perceived as less attractive. This evidence suggests that positive and negative emotions do not have opposite effects on the perception of attractiveness. Other studies also obtained higher intensity levels associated with angry and happy faces, as compared with neutral faces (Garrido and Prada, 2017).

Studies have also explored the role of the participants’ gender when interpreting emotion intensity assessments of facial expressions. Hama and Koeda (2023) suggested that female participants rated happy and sad expressions as more intense than male individuals. Happiness has also been identified as more intensely perceived among women rating female images, and less intensely when men are rating male images. Gong et al. (2018) have also suggested that females perceive both positive and negative emotions as more intense than males. However, these results also seem to be associated with age and culture. The aforementioned studies were carried out in Japan and China. A study carried out with German participants (Gong et al., 2018) suggested that women perceive negative emotions more intensely and positive emotions less intensely, when compared to men. Regarding attractiveness levels, studies suggest that this measure may be similar regardless of participants’ gender (Kranz and Ishai, 2006; Levy et al., 2008).

In summary, empirical research provides inconsistent results about the role of the eyes in the recognition of emotions. A number of studies support the theory that the eyes provide cues to the emotional state of people, thus claiming the main role in the recognition of emotions in the face (Ekman and Friesen, 1975; Baron-Cohen et al., 2001; Lee and Anderson, 2017). Other findings suggest that the eyes are not always the main values of emotion, but that this role is instead claimed by the mouth (Dunlap, 1927; Blais et al., 2012; Carbon, 2020; Langbehn et al., 2022). Either way, this approach sustains that eyes could be a central element (primary or secondary, in addition to the mouth) for decoding and assessing emotional states, and that the specific role of the eye could vary depending on the emotion type (Eisenbarth and Alpers, 2011; Kim et al., 2022; Tsantani et al., 2022). An alternative theory proposes that emotions, like faces, are processed in a holistic way, rather than by separate components (Omigbodun and Cottrell, 2013). However, results in this research area are contradictory (Curby et al., 2012; Tanaka et al., 2012), suggesting that although some features of the face could be relevant in understanding specific emotions, the face as a whole could also have effects on emotional decoding.

Previous studies have also established a relationship between perceived attractiveness and intensity of emotions, as well as between attractiveness and eye size. It must be stressed that the link between intensity and eye size remains mostly unexplored, and some results seem inconsistent and related to the emotions type. Overall, eye size ratings have been shown to increase alongside attractiveness, and apace with the intensity of happiness (Ueda et al., 2016). Fewer studies have addressed negative emotions. Regarding attractiveness, negative emotions are perceived as less attractive than happy and neutral ones (Ueda et al., 2016; Garrido and Prada, 2017). Regarding intensity, Eisenbarth and Alpers (2011) argued that eyes achieve a more central role than the mouth for the assessment of angry and sad faces, which in turn might make bigger eyes distort the perception of intensity in those emotions. Moreover, fearful and neutral faces are understood equally from the eyes and the mouth, which reduces the role of the eyes in comparison to the other emotions. Paradoxically, fearful faces are characteristically recognized by widened eyes (Barrett, 2018), which suggests the relevance of the eyes when it comes to this emotion. Gender and age might also have an effect on intensity (Gong et al., 2018; Hama and Koeda, 2023). Participant’s gender and sexual preference should also be considered when interpreting attractiveness ratings (Kranz and Ishai, 2006; Levy et al., 2008).

Additionally, when human representations are artificially altered to diverge from reality, it is essential to consider the degree of these changes. The uncanny valley is the phenomenon in which artificial human representations are perceived as uncomfortable as they become more similar to reality (Mori et al., 2012). This phenomenon is commonly seen in video games, robotics, and human-like dolls (Cheetham, 2011).

To summarize, there is debate concerning to what extent emotions are perceived in the eyes. Empirical results suggest that eye size could be a key indicator of some emotions, and that this characteristic might modify the intensity of the emotional cues provided by the face as a whole. Concretely, larger eyes would make happy faces look happier, angry faces angrier, and so on. According to evolution-based theories, neutral faces with bigger eyes may also be rated as more intense than those with smaller eyes (Berry and McArthur, 1985; Glocker et al., 2009). The hypothesis that eye size intensifies facial emotional clues also supports the argument that emotions are processed holistically, and that emotional expressions are deciphered by understanding the face as a compendium of pieces that must be looked at together, as the emotions in which the eyes do not have as much of an effect (e.g., happiness) would still be affected by the face in its entirety. However, the results are uncertain and new empirical studies are needed to consolidate (or refute) this line of research.

The main goal of the present study was to examine the relationship between eye size and facial emotional expression with the perceived intensity and attractiveness ratings. In addition, the study explored the potential moderator role of gender and age within the aforementioned relationships. Based on the existing empirical studies, we hypothesized that larger eyes should be associated with higher scores in attractiveness and intensity, particularly for fear and happiness. Due to the lack of previous empirical evidence, no hypothesis was formulated regarding the potential interaction of gender and age in the relationships.

The aim of this study was to obtain empirical evidence regarding the association between eye size and the perceived intensity and attractiveness of different emotions, in an experimental task based on the visualization of female faces. The comparison of the intensity levels showed an interaction with participants’ gender: within the male participants, differences were found across all eye size comparisons. Within the females group, differences were only found when comparing the unchanged and larger eyes. No interaction was found between gender and emotion type. The emotion perceived as the most intense was fear, and neutral faces were rated with the lowest intensity. No interaction was found when comparing age by emotion and age by eye size. Regarding the perception of attractiveness, no significant interactions were found with either age or gender. The lowest mean was associated with the smaller eyes, with no differences between the unchanged and larger eyes. The most attractive emotion was happiness.

The results of this study are based solely on the visualization of images of female faces. Therefore, they should not be considered representative of or generalizable to both men and women.

Neutral expressions were expected to be perceived as the least intense (Garrido and Prada, 2017), and fearful ones the most intense (Ekman and Friesen, 1975; Becker, 2012; Żurowska et al., 2018). Happy and angry faces were expected to be equally intense, and sadness more intense than neutral, but less than the other three (Ueda et al., 2016; Garrido and Prada, 2017). The results of this study support these findings, and outline the premise that codification of face expressions is related to eye size. Moreover, participants’ gender reached a moderator role in the relationships.

The results of our study also suggest that attractiveness ratings are not dependent on the participants’ age or gender. However, differences have been found associated with the emotion type. We hypothesized that happy expressions should be rated as more attractive than all the others, and neutral more attractive than fearful, sad and angry (Golle et al., 2013; Ueda et al., 2016; Garrido and Prada, 2017). Some of these differences have been observed. Happy faces were more attractive than all the rest. Neutral, fear, sadness and anger were not rated as different from each other except for the comparison neutral-fear with large eyes, and neutral-anger with small eyes. In regard to attractiveness and eye size, bigger eyes were expected to be perceived as more attractive than smaller eyes (Glocker et al., 2009; Przylipiak et al., 2018). In particular, it was expected for large eyes to be rated as more attractive than unchanged and small. A difference was also anticipated between unchanged and smaller eyes, as previously found by Glocker et al. (2009). A main effect of eye size on attractiveness was replicated, but the difference was not found where it was expected: participants found smaller eyes to be less attractive than unchanged and larger; but there was no difference between unchanged and larger.

Overall, the results of this study suggest that emotions are perceived more intensely when the eye size is digitally increased in women’s faces, and that these changes are gender-dependent, men being more susceptible to changes than women. Noticeably, neutral faces with smaller eyes were rated as less intense than larger and unchanged (as the rest of emotions). This phenomenon—“more intense neutrality”—may suggest that larger eyes not only intensify emotions, but also increase the perceived intensity of the face in its entirety, which supports the theory that emotions are, at least partially, holistically perceived (Omigbodun and Cottrell, 2013), as bigger eyes intensify the message projected by the rest of the face.

Glocker et al. (2009) suggested that larger eyes in babies lead to a higher perception of attractiveness, which motivates caretaking behavior, thus increasing the child’s chances of survival. The phenomenon “the bigger the eyes, the more intense the emotion” supports the theory that emotions are perceived as more intense due to evolutive mechanisms: a sad baby with bigger eyes might seem sadder, which might prompt caregivers to take care of them more urgently. Both Glocker et al. (2009)’s attractiveness-caretaking behavior theory, and the theory “the bigger the eyes, the more intense the emotion” support the idea that bigger eyes increase the chances of survival by increasing both the intensity of the emotion on display, and the attractiveness of the face. This phenomenon is also applicable to adulthood age, as bigger eyes also make adult faces gain in attractiveness (Berry and McArthur, 1985; Thornhill and Gangestad, 1999; Matsushita et al., 2015b; Przylipiak et al., 2018) and intensity, as seen in this study.

The results of the present study support the premise that eye size has an impact on attractiveness: faces with larger eyes are generally perceived as more attractive than those with smaller eyes (Glocker et al., 2009; Przylipiak et al., 2018). However, our data also indicated that bigger than normal eyes do not enhance attractiveness any more than unchanged eyes. This is a novel finding, as previous literature has repeatedly suggested that bigger eyes are perceived as more attractive, especially in happy faces (Berry and McArthur, 1985; Glocker et al., 2009; Matsushita et al., 2015b; Ueda et al., 2016; Przylipiak et al., 2018; Hine and Okubo, 2021). In any case, our results should be interpreted considering that our research tested increase and decrease changes of 15%, while other experiments have tested smaller changes in the 5% or 10% range (Przylipiak et al., 2018). The modifications used in our study might have altered the facial expression too much making the images reach the uncanny valley, resulting in faces that were perceived as uncomfortable (missing any chance of gaining in attractiveness).

The results of this work should be interpreted in accordance with certain limitations. First, the asymmetrical distribution of the participants’ gender (the proportion of male individuals was lower than female individuals) should be considered for generalization purposes.

Second, the female faces visualized in the experimental task showed characteristics in addition to the face area, such as hair, and therefore differences in attributes such as hairstyles could have affected attractiveness ratings due to individual preferences. Furthermore, the variety of faces in the experimental task was relatively low, which limited the possibility of including intra-ethnical differences. However, it must be considered that the high number of experimental conditions in our study (three eye size values multiplied by five emotion types) did not allow a greater number of images, otherwise the experiment would have been too long and tiring for the participants.

In relation to the previous point, the study was carried out using images of female faces. Other replication studies including male faces are necessary for generalization purposes, mostly because females are usually perceived as more attractive and/or intense than males (Garrido and Prada, 2017). Lastly, additional research that explores a wider range of emotions (including other primary emotions such as surprise or contempt) is necessary to further understand the role of eye size in the interpretation of facial expressions.

To our knowledge, this is the first study addressing the combined impact of eye size and emotions on the perceived intensity and attractiveness of facial expressions, considering the potential role of the participants’ gender and age. The results indicate that the eyes act as intensifiers of emotions, particularly among male respondents, and that facial emotional expressions are, at least partially, holistically perceived. The results of this work also show that the attractiveness ratings were affected by eye size, with smaller eyes perceived as less attractive, independently of the emotion. This evidence could be interpreted as an evolutionary phenomenon.

The datasets presented in this article are not readily available because restrictions apply to the datasets: limitations to making data publicly available include participant privacy, in accordance with the ethical consent provided by participants on the use of confidential/identifiable human data. Requests to access the datasets should be directed to AEJ, Alanis.Este@autonoma.cat.

The studies involving humans were approved by Ethics board of the Department of Psychological Sciences, Birkbeck College, University of London (approval number: 2122068, date of approval: 09/05/2022). The studies were conducted in accordance with the local legislation and institutional requirements. Participants provided informed consent by ticking a box within the website before starting the experiment. Written informed consent was obtained from the individual(s) for the publication of any identifiable images or data included in this article.

AEJ: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing. RG: Formal analysis, Writing – original draft, Writing – review & editing.