Fixation Sequences as Time Series: A Topological Approach to Dyslexia Detection

2026-04-23 · Source: Computation and Language · Field: Science & Research — Mathematics & Computational Sciences, Artificial Intelligence & Machine Learning, Health & Medical Research · Depth: Expert, quick

Summary

A new study introduces a topological data analysis method for dyslexia detection using eye-tracking data, interpreting fixation sequences as time series. The research develops novel filtrations for time series and constructs "hybrid models" that integrate topological features derived from persistent homology with traditional statistical features. Applied to the Copenhagen Corpus, which includes scanpaths from dyslexic and non-dyslexic L1 and L2 readers, these hybrid models demonstrate superior performance compared to existing approaches relying solely on traditional features. The topological features alone achieve performance comparable to established baseline methods, highlighting their ability to capture complementary information within fixation sequences. The proposed filtrations specifically outperform previously existing methods.

Key takeaway

For AI scientists developing diagnostic tools from time-series data, consider incorporating persistent homology. This method captures complementary information, as demonstrated in dyslexia detection from eye-tracking, and can significantly improve model performance when combined with traditional statistical features. Explore novel filtration techniques to maximize the utility of topological data analysis in your applications.

Key insights

Topological data analysis, via persistent homology, enhances dyslexia detection from eye-tracking data.

Principles

• Persistent homology extracts robust, multi-scale features.
• Hybrid models combine topological and statistical features.

Method

Interpreting fixation sequences as time series, novel filtrations are applied to extract persistent homology features, which are then combined with traditional statistical features for classification.

In practice

• Apply novel filtrations to time series data.
• Integrate topological features with statistical models.

Topics

• Persistent Homology
• Dyslexia Detection
• Eye-Tracking Data
• Fixation Sequences
• Topological Data Analysis

Best for: AI Scientist, Research Scientist

Related on AIssential

• Tracking Representation Dynamics in Large Language Models with Persistent Homology — Persistent homology uncovers internal representation dynamics in LLMs during fine-tuning, showing early topological reorganization and objective-specific trajectories.
• PHINN: Persistent Homology Inspired Neural Network for Rare-Event Time Series Generation — Rare events in time series possess stable topological fingerprints, enabling more effective generative modeling than statistical moments.
• TRACE: A Temporal Conditional Estimation for Multimodal Time Series Foundation Models — TRACE systematically infers missing time series modalities from available data, enhancing robustness and cross-modal representation learning.
• A Zipf-preserving, long-range correlated surrogate for written language and other symbolic sequences — A new model generates symbolic sequence surrogates preserving both symbol frequencies and long-range correlations.
• A Parameter-free Adaptive Resonance Theory-based Topological Clustering Algorithm Capable of Continual Learning — A new ART-based clustering algorithm autonomously estimates critical parameters for superior, adaptive performance.
• Beyond Means: Topological Causal Effects under Persistent-Homology Ignorability — Persistent homology enables identifying causal effects on outcome distribution shapes, addressing limitations of mean-based estimands.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computation and Language.