Written by Elin Williams, Nick Thompson, Gareth McCray, and Bhismadev Chakrabarti 

Imagine this: you’re passionately narrating the latest plot twists in your favourite TV show to your friend, but you can’t help but wonder if they’ve stopped listening. You realise their gaze has shifted; surprisingly, they’re no longer focused on your animated storytelling, but focused instead on the cute dog passing by. This seemingly simple act of judging where others are looking, and consequently, where their attention is directed, plays a pivotal role in our daily interactions. 

We make this judgement with remarkable speed. Yet, contrary to our initial assumptions, perceiving gaze extends beyond reading information from others’ eyes. Gaze perception is shaped by an array of factors. Emotions, the context, the angle of the nose, the orientation of the head, and even the position of the body, collectively influence how we interpret other people’s gaze. When an individual’s eyes, head, and body are all turned towards us, it’s an unmistakable signal that we’re the focus of their attention. However, when these cues are misaligned, uncertainty can cloud our perception of gaze. 

The recent COVID-19 pandemic served as a real-world social experiment, potentially altering the way in which gaze is perceived due to the widespread use of face masks. These masks obscure nearly half of the face and all cues therein, forcing people to rely primarily on the eyes to judge gaze. This may make gaze perception more challenging.  

To explore whether face masks influence how people perceive gaze, we conducted an online study with 157 participants. We presented participants with photographs of faces, half wearing face masks and the other half unmasked, with varying gaze directions. Some faces gazed directly at the participants, some clearly averted their gaze away from them, and others had more ambiguous eye directions (Figure 1). The task for participants was simple: decide whether each face was looking at them or not.

Our findings revealed that when participants viewed masked faces, they interpreted a broader range of gaze angles as directed towards them. More simply, face masks led participants to overestimate direct gaze. This overestimation in the face of increased uncertainty may be something of a ‘safety net’ which ensures we don’t miss social opportunities or overlook potential threats lurking nearby. 

Previous studies have highlighted the existence of individual differences in gaze perception. Individuals with social anxiety, for example, are more likely to overestimate direct gaze, while those possessing more traits associated with Autism Spectrum Condition are less likely to do so. Additionally, research suggests that some autistic individuals prioritise cues from the head or body over eye-region cues when judging gaze. It is therefore possible that obscuring the lower face with a face mask poses a greater challenge for autistic individuals, who may place greater reliance on lower face cues during gaze perception. To examine this, all our participants, including 33 with a clinical diagnosis of autism, completed a questionnaire measure of autistic traits. We found that participants reporting higher levels of autistic traits exhibited more pronounced changes in gaze perception when face masks were introduced. While some became more uncertain (i.e., overestimation of direct gaze), others actually became more precise in their judgements. 

This result highlights the individual differences in how we process gaze signals. Those who prioritise non-eye-region cues during gaze perception may experience less precise gaze judgements, or overestimate direct gaze in the absence of lower face cues. On the other hand, those who are indifferent to others’ gaze, might be more likely to shift their focus to the most reliable cue available (i.e., the eyes) when lower face cues are unavailable, resulting in more precise gaze judgments when face masks are present.

This study not only sheds light on the complex interplay of factors that shape our perception of gaze, but also underscores the need to consider broader implications of interventions, such as mask mandates, on our communication methods. Our research demonstrates that the seemingly simple act of covering one’s lower face significantly influences the way we perceive gaze, and that this effect is heightened for individuals reporting more autistic traits. Understanding the factors that might explain why autistic individuals exhibit more pronounced changes in gaze perception as a result of obscuring the lower face is a key question to be addressed by future work in this area.