"AI Diagnoses Autism Symptoms... Accuracy 98.5%"

Development of AI for Early Autism Diagnosis
Diagnoses with 98.5% Accuracy Using DT-MRI

An artificial intelligence (AI) technology capable of early diagnosis of Autism Spectrum Disorder (ASD) in children aged 2 to 4 has been developed, attracting attention.

Image unrelated to the article content. [Photo by Getty Images Bank]

According to reports from major foreign media on the 25th (local time), a research team led by Dr. Ayman El-Baz from the Department of Biotechnology at the University of Louisville in the United States announced that they have developed an AI capable of diagnosing autism by analyzing special magnetic resonance imaging (MRI) of the brain.

The DT-MRI presented by the research team is a special imaging technology that uses AI to detect how water molecules move along the brain's white matter pathways.

Autism Spectrum Disorder mainly occurs due to improper connections in the brain. The research team stated that DT-MRI identifies abnormal neural connections that may lead to autism.

The AI separates brain tissue images from DT-MRI scans and extracts images representing the level of connectivity between brain regions. Then, it compares these with the brain neural patterns of autistic children using machine learning algorithms.

The research team conducted DT-MRI on 226 children aged between 24 and 48 months from the 'Autism Brain Imaging Data.' As a result, 126 autistic children and 100 neurotypical children were classified. The AI used diagnosed autism with an accuracy of 98.5%.

Dr. El-Baz said, "Current methods for diagnosing autism are very subjective," adding, "There is an urgent need to develop new and objective technologies for early diagnosis of autism."

Dr. Leandra Berry, head of the Autism Program at Texas Children's Hospital in Houston, USA, stated, "I support all diagnostic technologies that help accurately diagnose autism early," but added, "Much research must back this technology before we can actually use it."

Meanwhile, according to the 2023 Autism Report by the U.S. Centers for Disease Control and Prevention (CDC), less than half of autistic children were diagnosed before age 3, and 30% of children meeting autism criteria were not formally diagnosed until age 8.

The research team claims that this AI analysis technology can reduce psychologists' workload by up to 30%. However, some scientists say, "There is insufficient data on how much this technology can reduce workload," and "It is premature for public use."

○ Characteristics of Autism Spectrum Disorder

1. Lack of social communication skills
- Difficulty in reciprocal interactions such as conversation and emotions
- Difficulty with nonverbal communication, making it hard to understand implied meanings
- Difficulty forming and maintaining relationships

2. Narrow and repetitive interests
- Engages in repetitive behaviors
- Dislikes changes
- Has narrow, deep, and intense interests
- Is either sensitive or insensitive to sensory stimuli

3. Poor eye contact
4. Does not respond well when called by name
5. Rarely attempts to share interests with others
6. Does not understand how to communicate with facial expressions
7. Narrow and repetitive interests

This content was produced with the assistance of AI translation services.