Eye Movements During Everyday Behavior Predict Personality Traits

Sabrina Hoppe, Tobias Loetscher, Stephanie A. Morey, Andreas BullingFrontiers in Human Neuroscience 2018 · DOI

Abstract

Besides allowing us to perceive our surroundings, eye movements are also a window into our mind and a rich source of information on who we are, how we feel, and what we do. Here we show that eye movements during an everyday task predict aspects of our personality. We tracked eye movements of 42 participants while they ran an errand on a university campus and subsequently assessed their personality traits using well-established questionnaires. Using a state-of-the-art machine learning method and a rich set of features encoding different eye movement characteristics, we were able to reliably predict four of the Big Five personality traits (neuroticism, extraversion, agreeableness, conscientiousness) as well as perceptual curiosity only from eye movements. Further analysis revealed new relations between previously neglected eye movement characteristics and personality. Our findings demonstrate a considerable influence of personality on everyday eye movement control, thereby complementing earlier studies in laboratory settings. Improving automatic recognition and interpretation of human social signals is an important endeavor, enabling innovative design of human-computer systems capable of sensing spontaneous natural user behavior to facilitate efficient interaction and personalization.

Introduction

For decades, psychologists have posited that eye movements serve as a "window to the mind." However, the vast majority of research in this field has been restricted to controlled laboratory environments where participants view static images while seated. This study bridges the gap between the lab and the real world by investigating whether eye movements during spontaneous, everyday behavior can serve as reliable predictors of personality.

Beyond simple correlations, this research employs predictive modeling to determine if a machine learning system can categorize an individual's personality based solely on gaze data collected while they navigate a natural environment. This has significant implications for the development of "socially aware" computing systems that can adapt to user traits in real-time.

Methods and Feature Engineering

The researchers monitored 42 participants using head-mounted eye trackers as they performed a 10-minute errand on a university campus. Following the task, participants completed validated personality assessments, including the NEO Five-Factor Inventory (NEO-FFI) and curiosity scales.

The technical innovation of this work lies in its high-dimensional feature extraction pipeline. Rather than relying solely on traditional metrics like fixation duration, the team extracted 207 distinct features, including:

  • Gaze Heatmaps: 8x8 grids representing where participants looked relative to their head position.
  • N-grams: A novel application that encodes sequences of eye movements (e.g., a specific saccade followed by a fixation), capturing the temporal rhythm of visual behavior.
  • Pupillometry and Blinks: Metrics such as variance in pupil diameter and blink rates.
Feature importance across personality traits, highlighting the role of n-grams and gaze heatmaps.

A Random Forest classifier was trained using a nested cross-validation scheme to ensure the model generalized well to unseen data, preventing overfitting on the small sample size.

Results

The model's performance was measured using the F1 score. Against a theoretical chance level of 33.3% (for three-way classification into low, medium, and high trait levels), the system achieved significant predictive accuracy for most traits:

  • Extraversion: 48.6% (The most predictable trait)
  • Agreeableness: 45.9%
  • Conscientiousness: 43.1%
  • Neuroticism: 40.3%
  • Perceptual Curiosity: 37.1%
Classification performance (F1 score) of the proposed model compared to baseline measures across different personality traits.

Notably, the model failed to predict "Openness to Experience" better than chance. The researchers suggest that the routine nature of the errand task may not have elicited the specific exploratory behaviors associated with high openness.

Key Insights and Discussion

The study revealed that the sequence of eye movements (n-grams) is often more informative than isolated gaze events. This suggests that personality is reflected in the rhythmic patterns of how we scan our environment. Furthermore, the study found that previously neglected features, such as pupil diameter variance, were critical for identifying traits like Neuroticism.

Crucially, the personality "signal" remained stable across different contexts (e.g., walking vs. shopping), indicating that these visual signatures are consistent individual traits. While the current accuracy levels are a "proof of concept" rather than a final product, they represent a major step toward pervasive, socially-aware technology.

Limitations and Ethics

As a pilot study with 42 participants, larger datasets are required to refine these models. Additionally, the ability to infer personality from gaze data raises important ethical and privacy considerations, particularly as wearable eye-tracking technology (such as AR glasses) becomes more common in daily life.

Summary updated

Abstract

Besides allowing us to perceive our surroundings, eye movements are also a window into our mind and a rich source of information on who we are, how we feel, and what we do. Here we show that eye movements during an everyday task predict aspects of our personality. We tracked eye movements of 42 participants while they ran an errand on a university campus and subsequently assessed their personality traits using well-established questionnaires. Using a state-of-the-art machine learning method and a rich set of features encoding different eye movement characteristics, we were able to reliably predict four of the Big Five personality traits (neuroticism, extraversion, agreeableness, conscientiousness) as well as perceptual curiosity only from eye movements. Further analysis revealed new relations between previously neglected eye movement characteristics and personality. Our findings demonstrate a considerable influence of personality on everyday eye movement control, thereby complementing earlier studies in laboratory settings. Improving automatic recognition and interpretation of human social signals is an important endeavor, enabling innovative design of human-computer systems capable of sensing spontaneous natural user behavior to facilitate efficient interaction and personalization.
BibTeX
@inproceedings{hoppe2018eye,
  author = {Sabrina Hoppe and Tobias Loetscher and Stephanie A. Morey and Andreas Bulling},
  title = {{Eye Movements During Everyday Behavior Predict Personality Traits}},
  booktitle = {Frontiers in Human Neuroscience},
  year = {2018},
  doi = {10.3389/fnhum.2018.00105},
}
Figure