Related Experiment Video
Updated: Sep 8, 2025
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
1.6K
Data Mining-Based Model for Computer-Aided Diagnosis of Autism and Gelotophobia: Mixed Methods Deep Learning Approach
Mohamed Eldawansy1, Hazem El Bakry1, Samaa M Shohieb1
1Information Systems, Faculty of Computers and Information, Mansoura University, Mansoura, Egypt.
JMIR Formative Research
|August 13, 2025
Summary
A new AI system accurately detects gelotophobia, the fear of being laughed at, by analyzing facial expressions and using questionnaires. This tool shows promise for early diagnosis in individuals with autism spectrum disorder (ASD).
Area of Science:
- Artificial Intelligence
- Psychology
- Computer Science
Background:
- Gelotophobia, the fear of being laughed at, affects 6% of neurotypicals and up to 45% of individuals with autism spectrum disorder (ASD).
- This comorbidity significantly impairs quality of life, especially in adolescents with high-functioning ASD.
- Automated detection tools are crucial for early diagnosis and intervention.
Purpose of the Study:
- Develop a deep learning system integrating facial emotion recognition and questionnaires.
- Detect gelotophobia in individuals with or without ASD.
- Enhance diagnostic accuracy for a socially vulnerable population.
Main Methods:
- Trained a deep learning classifier on 2932 facial images (50% ASD, 50% neurotypical) using the DeepFace library.
- Analyzed facial expressions for gelotophobia signs; used the GELOPH<15> questionnaire for ambiguous cases.
- Implemented the system in Python, utilizing PyTorch, scikit-learn, NumPy, and Pandas.
Main Results:
- Achieved 92% overall prediction accuracy for ASD identification.
- Successfully classified gelotophobia with clear facial expressions.
- Combined facial analysis with the GELOPH<15> questionnaire for improved diagnostic reliability.
Conclusions:
- Deep learning combined with diagnostic tools effectively detects gelotophobia, especially in individuals with ASD.
- High accuracy indicates potential for clinical and research applications.
- Future research can expand the system for broader psychological assessments.
