Predicting the Presence of Autism Spectrum Disorder Based on Eye-Tracking Scan Path Images

This research explores using eye-tracking data and machine learning to predict Autism Spectrum Disorder (ASD), achieving 74.6% accuracy with logistic regression, highlighting eye-tracking's potential for early, non-invasive ASD diagnosis and its applicability to other neurological conditions.

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that affects communication, social interaction, and behavior. With over 5.4 million individuals in the United States diagnosed with ASD, early detection is critical for improving outcomes through timely intervention. Eye-tracking technology, which monitors visual attention patterns, has emerged as a potential tool for early diagnosis. Our research aimed to explore whether eye-tracking scan path images could be used to predict the presence of ASD by applying machine learning techniques. We collected eye-tracking data from individuals with ASD and non-ASD controls, processed the images to enhance clarity, and labeled them accordingly (ASD = 1, non-ASD = 0). Several machine learning models were tested, with logistic regression being the most effective, achieving an accuracy of 74.6% using a regularization parameter (C) of 0.1. This indicates that while eye-tracking can aid in ASD classification, there is still room for improvement in model performance. Our findings suggest that eye-tracking combined with machine learning has the potential to assist in early ASD detection. This non-invasive method may also be useful in diagnosing other neurological conditions, such as Alzheimer’s, offering healthcare professionals a new approach for early intervention.

Explore More!

Source Code

Detecting Racial and Ethnic Bias in Social Media Posts Using Machine Learning

This study investigates the use of machine learning to detect racial and ethnic bias in social media posts using TF-IDF text embeddings and multiple classification models. Results show that Random Forest achieved the highest overall accuracy of 93.45%, demonstrating the potential of machine learning to support automated bias detection despite challenges posed by imbalanced data and the complexity of language.

Mira C.

Mentored by Matan Gans

Fast and Accurate Gamma-Ray/Hadronic Particle Shower Classification Using Machine Learning

This research applies dimensionality reduction techniques, such as Pearson Correlation and Principal Component Analysis, to machine learning models like Random Forests and Support Vector Classifiers to simplify the prediction of atmospheric gamma-ray particle showers, achieving a modest accuracy increase while reducing the number of features used.