Eye-tracking based classification of Mandarin Chinese readers with and without dyslexia using neural sequence models

TL;DR

This study develops sequence models to classify Mandarin Chinese readers with and without dyslexia based on eye movement data, demonstrating that sequential analysis is effective even for logographic scripts and that linguistic context does not enhance classification.

Contribution

The paper introduces simple sequence models for eye movement classification that consider the entire stimulus and compares the impact of linguistic features, achieving state-of-the-art results.

Findings

Sequence models classify dyslexia effectively in Chinese eye movement data.

Incorporating linguistic stimulus features does not improve classification accuracy.

Sequence analysis outperforms aggregated feature approaches.

Abstract

Eye movements are known to reflect cognitive processes in reading, and psychological reading research has shown that eye gaze patterns differ between readers with and without dyslexia. In recent years, researchers have attempted to classify readers with dyslexia based on their eye movements using Support Vector Machines (SVMs). However, these approaches (i) are based on highly aggregated features averaged over all words read by a participant, thus disregarding the sequential nature of the eye movements, and (ii) do not consider the linguistic stimulus and its interaction with the reader's eye movements. In the present work, we propose two simple sequence models that process eye movements on the entire stimulus without the need of aggregating features across the sentence. Additionally, we incorporate the linguistic stimulus into the model in two ways -- contextualized word embeddings and manually extracted linguistic features. The models are evaluated on a Mandarin Chinese dataset containing eye movements from children with and without dyslexia. Our results show that (i) even for a logographic script such as Chinese, sequence models are able to classify dyslexia on eye gaze sequences, reaching state-of-the-art performance, and (ii) incorporating the linguistic stimulus does not help to improve classification performance.