Hi, I'm Kerem Keptiğ

AI and Healthcare Revolution: Using Machine Learning to Diagnose Autism Spectrum Disorder

Posted by Kerem Keptiğ on June 26, 2024

Today, artificial intelligence (AI) and data science hold significant potential for revolutionizing healthcare. Lengthy and costly diagnostic processes, such as for Autism Spectrum Disorder (ASD), can be made more efficient using AI techniques. In this post, I want to share my experience with a project aimed at diagnosing ASD through machine learning (ML), which I undertook during my final year at Middle East Technical University.

Project Overview

During my final year, I worked on a project that utilized eye-tracking data to detect ASD. Our goal was to record individuals' eye movements via webcams, process this data using ML models, and provide accurate predictions to doctors. This approach offers a cost-effective and time-saving alternative to traditional diagnostic methods.

Technical Details

We recorded eye movements while individuals performed specific tasks. The collected data was then transformed into metrics and fed into an ML model. The model analyzed these metrics to predict ASD presence, with results being shared with doctors for validation. Throughout this project, I gained in-depth knowledge about fixation patterns and learned how to integrate and optimize ML models effectively.

Results and Impact

This project demonstrated that using eye-tracking data can significantly streamline the ASD diagnostic process, making it faster and more cost-efficient. The potential to enhance healthcare services and improve patient quality of life through AI and data science is immense, as shown by our successful project outcomes.

Future Plans

Based on the promising results of this project, we will aim to write and publish a comprehensive article detailing our methods, findings, and the potential implications for the healthcare industry. This will not only contribute to academic knowledge but also help in advancing practical applications of AI in diagnosing ASD.