Summary: A new study pinpoints the exact moments when the brain detects the direction of another person’s gaze, improving our understanding of social perception and conditions such as autism and Alzheimer’s disease. Using EEG combined with machine-learning decoding, researchers measured brain responses while participants viewed 3D avatars with different head and eye orientations.
The results show that the brain first registers head orientation and only later processes eye direction, and that attention or task demands can sharpen gaze perception. These findings may help refine early diagnosis and intervention strategies for disorders marked by altered social attention.
Key Facts:
- Hierarchical processing: Head orientation is detected from about 20 ms, while eye direction is decoded from roughly 140 ms.
- Task influence: Gaze-direction decoding is more accurate when participants are explicitly attending to gaze.
- Clinical relevance: Insights could contribute to earlier diagnosis and new approaches for autism spectrum disorders and Alzheimer’s disease.
Source: University of Geneva
The direction of someone’s gaze is a fundamental cue in social interaction. Our ability to communicate instantly often depends on how quickly and accurately our brain detects where others are looking. But what is the timing of this process, and which visual cues are prioritized?
A research team at the University of Geneva (UNIGE) published a study in the journal NeuroImage that narrows down, with high temporal precision, when the brain discriminates gaze direction. By combining high-density electroencephalography (EEG) with machine-learning decoding, the investigators isolated the neural signatures that reflect head orientation and eye gaze and examined how task context changes their timing and accuracy.
Because faces are among the most consistent visual stimuli from birth, the human brain has evolved specialized mechanisms to recognize faces and interpret the social signals they convey. Direct gaze typically signals social engagement, while averted gaze often indicates withdrawal. Understanding how quickly these cues are parsed by the brain—and how head position interacts with eye direction—helps clarify the neural basis of social attention.
Previous research frequently examined the eyes in isolation, but this study examined their processing within the context of realistic faces and head orientations. Participants viewed 3D avatars that combined five head angles and corresponding gaze directions. Two separate tasks were used: in one, participants judged head orientation; in the other, they judged eye direction.
Cerebral analysis of gaze
EEG recordings revealed that information about head orientation and eye direction can be decoded independently from brain activity. The patterns showed a clear temporal hierarchy: global visual cues such as head orientation are detected extremely early—around 20 milliseconds—followed by more local information from the eye region beginning at approximately 140 milliseconds.
“This hierarchical organization enables the brain to integrate head orientation and eye-region cues to form an accurate and efficient judgment of gaze direction,” explains Domilė Tautvydaitė, a postdoctoral fellow at UNIGE’s Faculty of Psychology and Educational Sciences and first author of the study.
The researchers also found that decoding accuracy for gaze direction increased when participants were explicitly instructed to attend to gaze. In other words, task relevance modulates how reliably and how quickly the brain represents gaze direction. “In everyday settings, when people enter a social mode of attention, they become faster and more accurate at inferring others’ intentions,” notes Nicolas Burra, senior lecturer at UNIGE and director of the Experimental Social Cognition Laboratory (ESClab), who led the research.
A cutting-edge method
The team combined electroencephalography with advanced machine-learning techniques—specifically an event-related potential (ERP) decoding approach using support vector machines and error-correcting output codes—to detect neural patterns tied to head and gaze orientation. This approach allowed them to predict the decoding of gaze and head direction before participants consciously reported them, providing far greater temporal precision than conventional ERP analyses.
“The method is a technical advance for the field, offering a much more precise way to track the timing of social perception processes,” says Nicolas Burra.
These insights are relevant for conditions where gaze decoding or face recognition is disrupted. In many individuals with autism spectrum disorders, gaze processing can be atypical and eye contact may be avoided. Similarly, Alzheimer’s disease can erode face recognition and social engagement as memory declines. Better understanding the neural timing and task-dependence of gaze perception could inform early diagnostic markers and guide therapeutic strategies.
The study’s methods and results contribute to potential early diagnosis of autism spectrum disorders in children and to a deeper understanding of social and memory-related impairments in Alzheimer’s disease. Ongoing research from UNIGE’s ESClab and related projects—such as Dr. Tautvydaitė’s continuing work—aims to extend these findings to more naturalistic, real-life social interactions.
About this visual neuroscience research news
Author: Antoine Guenot ([email protected])
Source: University of Geneva
Contact: Antoine Guenot – University of Geneva
Image: The image is credited to Neuroscience News
Original Research: Open access. “The Timing of Gaze Direction Perception: ERP Decoding and Task Modulation” by Nicolas Burra et al., NeuroImage
Abstract
The Timing of Gaze Direction Perception: ERP Decoding and Task Modulation
Recognizing another person’s eye gaze direction provides vital information about their focus of attention and intentions. The temporal dynamics of gaze processing have been studied with event-related potentials (ERPs) measured by EEG, but the precise timing at which the brain distinguishes gaze direction independently of other facial cues has remained unclear.
This study used an ERP decoding approach that combined support vector machines with error-correcting output codes to investigate the time course of gaze processing. EEG was recorded from healthy young adults: 32 participants completed a gaze-direction detection task and 34 completed a face-orientation task. Each task presented realistic 3D faces with five head and gaze orientations (30° and 15° to the left or right, and 0°).
Classical ERP analyses did not reveal clear gaze-direction effects, but ERP decoding showed that gaze-direction discrimination—irrespective of head orientation—emerged at about 140 ms in the gaze-direction task and at about 120 ms in the face-orientation task. Decoding accuracy was higher when gaze was task-relevant and was highest for direct gaze in both tasks. These findings indicate that task relevance modifies the neural decoding of gaze, affecting both latency and accuracy.