“The soul, fortunately, has an interpreter — often an unconscious but still a faithful interpreter — in the eye,” Charlotte Bronte wrote. As it turns out, eyes are also pretty reliable indicators of personality.

In a recent study conducted by the University of South Australia, University of Stuttgart, Flinders University, and the Max Planck Institute for Informatics in Germany, researchers describe a machine learning model that can predict traits like sociability, curiosity, and conscientiousness from a person’s eye movements alone.

“One key contribution of our work is to demonstrate, for the first time, that an individual’s level of neuroticism, extraversion, agreeableness, conscientiousness, and perceptual curiosity can be predicted only from eye movements recorded during an everyday task,” the researchers wrote. “This finding is important for bridging between tightly controlled laboratory studies and the study of natural eye movements in unconstrained real-world environments.”

The team notes that previous studies suggest a strong relationship between eye movements and personality traits. Optimists, for example, tend to spend less time looking at negative or depressing images than pessimists, and naturally curious people dwell on locations in abstract animations. But historically, gazes have been largely measured and correlated with personality test results manually, in lab environments. Now researchers are able to automate the process.

The team recruited fifty students from Flinders University for the study: 42 females and eight males. To track their eye movements and gazes, they used a head-mounted system from SensoMotoric Instruments (SMI) and a mobile phone that continuously recorded video footage.

They then assessed the participants’ personality traits using three questionnaires: the NEO Five-Factor Inventory, which consists of 60 questions designed to gauge neuroticism, extraversion, openness, agreeableness, and conscientiousness; Perceptual Curiosity, which assesses interest in “perceptual stimulation” and “visual-sensory inspection”; and the Curiosity and Exploration Inventory (CEI-II). Prior to the personality and curiosity tests, the scientists instructed them to walk around the Flinders campus in Adelaide, Australia for 10 minutes and purchase any item of their choice.

After collecting the subjects’ gaze data and recording their written responses, the team used a trained neural network to predict baselines for each of the aforementioned personality traits. The model performed “well above” average in some cases, they found, predicting neuroticism with 40.3 percent accuracy, extraversion with 48.6 percent accuracy, and agreeableness with 45.9 percent accuracy. Its performance was a bit weaker on openness and curiosity, however.

The researchers expect that larger real-world gaze dataset, along with further research on trait-specific eye movements (i.e., eye movements associated with certain personalities), would improve the model’s prediction accuracy and reliability.

“While predictions are not yet accurate enough for practical applications, they are clearly above chance level and outperform several baselines,” the team wrote.

The work also shed light on the link between personality traits and eye movements. By analyzing eye movement and ranking them by their importance for personality, the researchers discovered that certain characteristics like pupil diameter were strongly linked with specific traits. (Pupil diameter was associated with neuroticism, for example.)

“Going beyond characteristics investigated in earlier works, [our] approach also allowed us to identify new links between previously under-investigated eye movement characteristics and personality traits,” they wrote.

The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here