A recent study used special eye-tracking technology to investigate how people look at each other’s eyes and faces during conversations. The researchers, who published their results in Scientific Reports, found that people who exhibited more direct eye-to-eye contact during their conversation tended to also be better at following the direction of another’s gaze (they were better at understanding where the other person was looking). The research provides unique insights into non-verbal communication.
Much of human social communication occurs nonverbally, and eye contact plays a crucial role in allowing individuals to convey and interpret information such as attention, mental states, intentions, and emotions. Eye contact is not only passively received but also reciprocated through mutual looks.
The researchers wanted to examine the frequency and types of mutual looking behaviors, such as direct eye-to-eye contact and other gaze interactions involving different parts of the face. They were also interested in understanding how the mutual looking behaviors observed during interactions might influence subsequent gaze-following behavior.
“We are all familiar with the saying ‘The eyes are the window to the soul.’ However, there is little scientific knowledge about the information that human eyes convey in natural interactions,” explained study author Florence Mayrand, a doctoral student in Experimental Psychology at McGill University.
“Past work has shown that humans preferentially attend to eyes and faces when passively viewing images of faces. In this study, we examined if such preference for attending to the eyes of others was also observed during real life interactions and how often it would occur in when two persons interacted. Additionally, as eye-to-eye contact in natural interactions is often interpreted to indicate communicative intent, here we also investigated how such real-life looking behaviors connected with the interpretation of gaze information later on.”
Thirty participants, forming 15 pairs, were selected for the study. These participants consisted of 25 females and 5 males, with an average age of 20.3 years, ranging between 18 and 24. The analysis eventually centered on data from 14 participants (12 females, 2 males) in 7 pairs due to the exclusion of data with insufficient eye tracking quality.
Initially, participants engaged in face-to-face interactions with their respective partners, in which specialized eye-tracking eyeglasses were employed. These glasses facilitated real-time monitoring of participants’ eye movements through both front-facing cameras and inward-facing sensors, ensuring comprehensive assessment of eye movement patterns and orientation.
The participants then undertook an individual gaze cuing task, where they were presented with images of their partner’s face displaying various gaze directions (left or right). These images were followed by target letters that required prompt identification, allowing the researchers to gauge participants’ tendencies to follow gaze cues.
“We used dual mobile eye trackers to simultaneously record gaze from two participants while they engaged in a real-life social interaction of approximately 5 minutes in length. After the interaction, each participant completed an individual computerized task in which we measured their response time to targets looked at by images of their interactive partner.”
“We extracted their patterns of mutual looking during the interaction and divided them within different categories: eye-to-eye looking (when both participants looked at each other’s eye region), eye-to-mouth looking (when one participant looked at the eye region and another at the mouth region), and mouth-to-mouth looking (when both participants looked at the mouth regions), and then related the time spent in each of those behaviors to individual task performance.”
The researchers found that dyads predominantly engaged in non-mutual looking behaviors, accounting for approximately 88% of interaction time. In contrast, mutual looks, wherein both participants directed their gazes at each other, were observed in only about 12% of the interaction time. Eye-to-mouth mutual interactions were the most common, surpassing both eye-to-eye and mouth-to-mouth interactions.
“They engaged the least in eye-to-eye mutual looking (3.5% of interaction),” Mayrand noted. “We were surprised to find such a low prevalence of eye-to-eye mutual contact considering the large body of work indicating a preference for attending towards the face and eyes of others. However, because dyadic interactions offer no social breaks from direct gaze, it is possible that our participants opted to look more towards the background of the interaction room to offer themselves relief from the gaze of their partner.”
The researchers also found that specific gaze behaviors during the interactions were associated with the degree of gaze following exhibited during the individual computerized task. Specifically, participants who engaged in more eye-to-eye contact during the interaction tended to have greater gaze following tendencies later on. Conversely, participants who allocated more time to non-interactive looking behaviors during interactions exhibited weaker gaze following tendencies.
“This suggests that while humans may spend proportionately little time in direct eye-to-eye contact during social interactions, this social contact time may be very informative with regard to the messages conveyed by eyes and influence how we attend to social cues in the environment,” Mayrand told PsyPost.
In regard to future avenues of research, Mayrand said that it “would be interesting to study the impact of conversational context on looking prevalence. In the current study, participants completed a cooperative task, which may not have afforded a lot of mutual looks. Different interactive context such as a task in which participants are asked questions to get to know one another may change the prevalence of different types of looking behaviors. Alternatively, interactions among friends rather than strangers may also change the context of gaze dynamics. It would be important to understand how the amount of speech and the content of speech affects interactive gaze behavior. Finally, it would be important to replicate these results using a larger sample.”
The study, “A dual mobile eye tracking study on natural eye contact during live interactions“, was authored by Florence Mayrand, Francesca Capozzi, and Jelena Ristic.