Predicting Dyslexia in Adolescents from Eye Movements during Free Painting Viewing

Area of Science:

• Neuroscience
• Ophthalmology
• Cognitive Science

Background:

• Dyslexia is associated with abnormal eye movements.
• Previous research demonstrated eye movement abnormalities predict dyslexia during reading and specific tests.
• This study investigates eye movements during free visual exploration.

Purpose of the Study:

• To examine eye movement patterns during free exploration of paintings to identify dyslexia.
• To introduce and evaluate a novel feature for quantifying eye movement disconjugacy.
• To assess the efficacy of machine learning classifiers using these eye movement features.

Main Methods:

• Eye movement data collected during free exploration of paintings.
• Analysis using machine learning (ML) classifiers.
• Extraction of features using AIDEAL software (saccade amplitude, speed, disconjugacy).
• Introduction and calculation of a Disconjugacy Global Index (DGI) for continuous disconjugacy evaluation.

Main Results:

• Both AIDEAL features and the Disconjugacy Global Index (DGI) successfully categorized dyslexics from non-dyslexics.
• The DGI captures continuous disconjugacy, differing from saccade amplitude inequality.
• High predictability was observed for the specific paintings used.

Conclusions:

• Eye movement analysis during free visual exploration, particularly using DGI, shows potential for dyslexia detection.
• The findings highlight the physiological relevance of extracted eye movement features.
• The content of visual stimuli may influence the effectiveness of this diagnostic approach.