Art exploration is a complex process conditioned by factors at different levels, such as basic visual principles and more complex cognitive factors (Quiroga and Pedreira, 2011). Some studies have emphasized the role of “bottom–up” processes, proposing that gaze behavior during picture viewing is affected by physically salient visual features (Berlyne, 1974; Solso, 1994; Ramachandran and Hirstein, 1999; Zeki, 1999; Graham and Field, 2007, 2008; Graham and Redies, 2010). Other studies have recognized the role of “top–down processes,” demonstrating that the goal of visual exploration (task) (Buswell, 1935; Yarbus, 1967; DeAngelus and Pelz, 2009) and the person’s cultural background, art expertise, and familiarity with and interest in a specific work of art (Kristjanson and Antes, 1989; Nodine et al., 1993; Kapoula and Lestocart, 2006; Illes, 2008) are critical elements in influencing art exploration.

Beyond the different factors discussed above, a consensus exists about the role of the human figure as a critical factor in attracting the attention in art painting. In a controlled experimental study with art paintings, Massaro et al. (2012) have recently observed that when represented content included human subjects, content-related top–down processes prevailed over low-level, visually driven, bottom-up processes in guiding the observers’ explorative pattern. This finding indicated that when a human being was portrayed in a painting, the gazing behavior focused mostly on the human figure. The human body has a strong power to orient and attract visual attention at different ages and it plays a fundamental role in esthetic experience (Savazzi et al., 2014).

This behavior is possible because faces and body parts are stimuli of great biological and social significance; they can be rapidly and efficiently processed, and they can engage the attention system (Ro et al., 2007; Crouzet et al., 2010). Several behavioral studies have demonstrated that faces and body parts may have an attentional advantage over other objects (Downing et al., 2001; Ro et al., 2001; Rensink, 2002; Reed et al., 2004).

Why do people instinctively look at a human figure? Up to now, different approaches have been used to answer this question.

In real-life contexts, the ability to derive socially relevant information from faces is a fundamental requirement for normal reciprocal social interaction and interpersonal communication (Pelphrey et al., 2002). This process has been recognized even in the earliest stages of postnatal development (Goren et al., 1975; Johnson et al., 1991), and gaze contact expresses interest or a desire to collaborate (Emery, 2000). Moreover, this instinctive orientation to focus on faces has recently been attributed to both the recognition of emotion (Baron-Cohen et al., 1997) and the similarity of the subjects to the observers themselves (Hari and Kujala, 2009). The human brain has evolved a highly specialized mechanism for facial perception and recognition (Tsao et al., 2006) based on the analysis of its invariant structural features (e.g., the eyes, nose, and mouth). Due to their informative importance, these components remain the preferred attentional targets in healthy adults (Hernandez et al., 2009). One fundamental function of facial processing is to identify individuals, while the other function is to recognize the intentional state of others via the changes in facial features. The second function plays a crucial role in social interactions and constitutes the kernel of “social cognition” (Brothers et al., 1990). Understanding social intention is essentially linked to the brain capacity to recognize the unique morphology of the eye in primates and to provide information regarding the viewing direction of others. According to Langton (2000), not only the face, but also the orientation of head, the posture of the body, and other gestures influence the direction of social attention, and observers process all these cues automatically (Langton and Bruce, 2000).

The importance of the face has been confirmed also in art painting exploration. In fact, several studies have demonstrated that the face is generally the first part of the body that is scanned in portraits and that face-detection process should be particularly sensitive to global cues, mainly the presence of a face contour that activates a configural processing of the stimulus (Young et al., 1987; Abbas and Duchaine, 2008).

Not only the face, but also other body parts attract the viewer’s attention, as pointed out by Massaro et al. (2012) in the observation of art paintings characterized by image dynamism. In static images, the strong attractor was the face, while in dynamic images the attention was equally spread out across various body parts. The authors suggested an embodied explanation of these results: in the first case, embodied simulation of sensations and emotions guided the exploration pattern, while in the second case it was greatly affected by the simulation of actions.

According to the neuroscientific perspective, there is actual substantial evidence of an automatic conversion from vision to action, which occurs even when a person has no intentions to act on the viewed object (Tipper, 2004, 2010; Tipper and Bach, 2008). The automatic conversion from vision to action occurs not only when viewing inanimate objects that can be acted upon, but also when viewing other people’s actions (di Pellegrino et al., 1992). In this sense, the observation of body parts involved in another person’s actions, such as reaching, grasping, and gaze shifts (Grosbras et al., 2005), evokes similar processes in the viewer.

Thus, face and body parts can be seen as elements of the human figure activating different processes: the first related to the understanding of affective states and social intentions and the second related to the automatic activation of a motor-based representation associated with parts of the body.

The situation appears still more complex depending on the social content of the scene. According to Birmingham and colleagues (Birmingham et al., 2008), the social content appears to affect the scanning patterns in complex real-world action scenes. In fact, the level of activity in the scene influences the attention toward the faces when social content is high because face and eye information are critical to understand the social meaning of the action.

To date, this is the first study aiming to investigate how complex social scenes affect the exploration of different parts of the human figure. Specifically, the goal of this study was to investigate the exploration of different elements of the human figure performing an action (face vs. body parts in action) in complex social scenes characterized by the different levels of activity between agents depicted in scenes (individual vs. social) by using an eye-tracking methodology.

We hypothesized (Hp 1) that the orienting of attention toward human figure elements (face and other parts of the body involved in the actions) is different in scenes containing individual actions (several agents performing something separately) compared to scenes containing social actions (several agents performing something together). Specifically, we expected individuals to pay a greater attention to faces rather than arms in social action condition compared to individual action condition. In individual action condition individuals are expected to look at agents’ parts of the body involved in the action to understand the nature of the action, while in social action condition individuals are expected to look at agents’ faces to understand the social nature of the interaction between the agents.

To verify this hypothesis we used an eye-tracking methodology to explore human figure elements in complex social scenes depicted in art paintings. Eye tracking is a common methodology used to trace the inner operations of attention and cognition (Just and Carpenter, 1976; Jacob and Karn, 2003; Rayner, 2009) and to understand art exploration. Tracking eye movements is advantageous because they are fast and natural (Jacob and Karn, 2003), and it is possible to identify both specific areas of interest (salient regions of an image), as well as specific viewer’s explorative patterns.

Furthermore, we considered that the observation of body in action and motor resonance derived from simulation processes has been identified as a foundation of empathy (Jola et al., 2012). Freedberg and Gallese (2007) indicated that when contemplating artistic works by virtue of their visual content, it is possible to live embodied experiences and feel an empathetic engagement with the work of art. These feelings might consist of the viewer’s perceived understanding of the emotions of the agents represented in the pictures or, most strikingly, a sense of inward imitation of their observed actions, resulting in a seemingly mimetic engagement with the human figures depicted in the paintings.

Thus, we wondered if the observation of the body in action of the depicted agents might vary among people according to their empathic dispositions.

Starting from these premises, we aimed to examine (Hp 2) whether the observer’s individual characteristics in terms of emphatic abilities were associated with a specific scanning of the human figure (face and other parts of the body involved in actions) in scenes containing individual or social actions. For this purpose, we introduced a self-report instrument to assess the empathic responsiveness of participants and to understand whether this ability could affect the exploration of the individual and social scenes within an art painting.

The data analysis related to eye movement indicators and the self-report questionnaire was performed using SPSS analysis software (version 19.0; Statistical Package for the Social Sciences, Chicago, IL, USA).

As the human figure attracts gazing behavior in paintings, the goal of the study was to use eye-tracking methodology to investigate, with untrained participants, which part of a human figure orients visual exploration. For this reason, we analyzed selected paintings containing complex social scenes (Kuchincke et al., 2009; Graham and Redies, 2010), which are characterized by body dynamism and the presence/absence of the agents’ social interaction.

The first hypothesis, which stated that the human figure in art paintings characterized by action representation is scanned differently in scenes containing individual actions compared to scenes with social actions, was confirmed. The primary purpose of the analysis was to understand whether there was an interaction between the two factors (body parts and activity) on visual exploration patterns. In our analysis, we found significant interaction effects of scene characteristics in terms of social interaction among agents and human figure elements on the participant’s fixations. In paintings representing social actions, faces gained the traditional attractiveness, probably associated with the attempt to recognize intentions and emotions from the face (Baron-Cohen et al., 1997; Wallraven et al., 2009), and participants looked first at the ROIs of faces and for a longer period. The opposite pattern was found in paintings representing individual actions, where the participants looked first at the ROIs of arms and for a longer period.

This result was similar to that found by Birmingham et al. (2008). When the social content of a scene was high, that is, when several agents are performing something together, the social action drew the participants’ attention toward the face because facial information is critical to understand the social meaning of the action. However, when the social content of a scene is low, that is, when several agents are performing something separately, the individual action draws attention away from the face because face information is not critical to understand the action.

Moreover, this result was consistent with the embodied perspective proposed by Freedberg and Gallese (2007), despite some slight differences. In the observation of still images of actions in works of art, the authors referred to a motor simulation process in which the physical response can be located in the part of the body engaged in purposeful physical actions. From our results, we found that in experiencing an artwork, the observers were not generally focused on the entire human figure; instead, they appeared to focus on different human figure elements depending on the feature of the scene in terms of social interaction. Moreover, the body parts involved in the action attracted the viewers’ attention when the represented action was individual rather than social. Thus, it was likely that the viewers looked at arms to convert information from visual perception into corresponding physical movements and to understand the effort required to perform that action. In contrast, when the represented action was social, the body parts involved in the action were not highly attractive. In this case, the viewers focused on faces of agents involved in the action most likely to understand the social nature of the interaction between the agents.

According to Tessari et al. (2012), the ability to detect the intentions and the emotions of others during an interaction requires individuals to anticipate the consequences of their behavior. This was primarily possible via the observation of the agents’ body postures because intention could be automatically extrapolated from the body movement based on a shared motor representation. From our study, the face appeared to constitute socially relevant information in terms of valence and potential action when the intention has a social nature while body movement appears to be a determining element only when the movement presents a non-social nature.

This perspective appears to be critical to understand the visual exploration of figurative art paintings, particularly when the viewers are laypersons and the representations include several agents engaged in bodily actions.

This study also aimed to examine (hypothesis 2) whether the observers’ individual characteristics in terms of emphatic abilities resulted in the differential scanning of the human figure (face and other body parts involved in the actions) in scenes containing individual actions compared to scenes with social actions. Other studies have recently showed that among the individual characteristics of the viewers, gender and personality emerged as critical aspects that differentially oriented attention in visual exploration based on the fact that what we look helps construct what we see (Mercer et al., 2012). In this study, the empathic level of the participants affected the initial orientation and screening, but not further evaluation (Russo and Leclerc, 1994). Our results suggested two different roles of affective and cognitive empathic abilities in scenes with high level of social interaction. On the one hand, participants with higher levels of empathic concern (i.e., participants’ emotional identification with agents depicted in the scene) showed fewer fixations before focusing on faces. This finding is consistent with the interpretation that participants were paying attention to the emotional content of the scene – agents’ faces – in order to understand others and simulate their emotional state (Vuilleumier and Driver, 2007). It further suggests that emotional information processing is related to self-report of affective empathy (Hooker et al., 2010). On the other hand, participants’ ability to cognitively adopt another’s perspective and understand the meaning of the represented action resulted in the attraction toward the body parts involved in the action. This result is consistent with the notion that cognitive empathy skills, such as perspective-taking, relies on the use of internal models of observed action in order to understand another person’s situation (Gallese, 2007). These processes should be further investigated to confirm the different role of affective and cognitive empathy in orienting of attention toward human figure elements.

Taken together, the results of the present study revealed a critical role of social content in the exploration of the human figure elements in action. As demonstrated by Bayliss and Tipper (2005), it is important to investigate the role of the context in which orienting of attention occurs. This aspect is particularly important in investigating the attentional control of social orienting (Jordan and Tipper, 1998).

The face is confirmed to be as one of the most important elements of the human figure in attracting the viewers’ attention. Through facial observation, the participants understand others and simulate the emotional state via the generation of representations of the associated body state both in case of individual and social actions (Adolphs, 2002; Gallese et al., 2004). This process was confirmed specifically in participants with high empathic abilities that immediately orient their fixations on faces.

We must consider that the non-homogeneity in the visibility of the faces is a weakness of the study, and future studies are encouraged to include stimuli in which the faces are visible and detailed. Nevertheless, we want to emphasize that we used a configural or holistic approach to face processing that does not consider agent facial features (Penton-Voak et al., 2006). Specifically, we focused on orienting of attention toward different parts of human figure (Frischen et al., 2007). Emotional signals may still be perceived under difficult circumstances, such as when faces are behind a veil (Fischer et al., 2012). In this sense, the visibility or orientation of the faces did not compromise the results of the study, which can be generalized to other situations where expressive cues are not visible. Furthermore, future studies are encouraged to include motor actions where other body parts are involved in the bodily action movement.

Finally, as this study included a single experiment, this limitation calls for viewing the obtained results with caution. Thus, additional studies are required to investigate affective and cognitive processes by integrating other measures, such as psycho-physiological ones, which are more suitable to understand the activation of the mimetic engagement with other agents represented in the pictures and their implication for embodiment processes.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.