A research team from Stanford University in the United States accurately reproduces the brain activity of children with mathematical learning disability (MLD) using an AI-based model, revealing its fundamental cause./Courtesy of pixabay

It has been suggested through artificial intelligence (AI) research that a child who struggles with math may not have studied less, but rather has a different brain functioning method.

A research team from Stanford University in the United States announced on the 7th that they have identified the fundamental cause of learning difficulties experienced by children with mathematics learning disabilities (MLD) by precisely reproducing their brain activity using an AI-based model. The research results were published that day in the international academic journal 'Science Advances.'

Mathematics learning disabilities occur in 5 to 20% of children worldwide. These children have poorer arithmetic problem-solving abilities than their peers and struggle with basic addition and subtraction. Until now, abnormality in some brain areas has been revealed through brain imaging analysis, but the specific neural functioning issues have not been clearly identified.

The research team went a step further in brain imaging analysis by creating a brain model using digital twin technology. A digital twin replicates a physical entity in the virtual world identically. They developed an AI neural network trained on data regarding the actual mathematics problem-solving abilities and brain activity of children. The brain model learned to solve math problems like children in the virtual world, precisely mimicking the brain's responses and outcomes in the process.

As a result, the model mimicking children with mathematics learning disabilities showed a slower learning pace compared to typical children and struggled to differentiate between types of problems. It exhibited similar handling of different types of problems, such as addition and subtraction. Additionally, the virtual brain of the model with learning disabilities was excessively activated compared to the typical model.

The research team believes that excessive brain activation is the core cause of mathematics learning disabilities. When the brain processes information, the balance between excitation and inhibition is crucial; if this balance is disrupted and the brain becomes overly activated, there is a tendency to struggle with learning.

To verify this, the research team created a separate AI model reflecting a state of excessive brain excitation and observed it. As expected, it exhibited a slower learning pace and lower accuracy similar to children with mathematics learning disabilities. Although it was difficult to accurately determine cause and effect through observational brain imaging studies, this time, the cause was established using AI, and the results were examined to verify the causal relationship.

Even with a brain model of a child with mathematics learning disabilities, the accuracy of problem-solving improved to the level of typical children with repeated learning. However, the way information was organized within the brain differed from some typical models. Externally, performance improved, but the way the brain processed information did not completely change.

The research team noted, "This is a case of individually analyzing the causes of learning disabilities by simulating each child's brain characteristics using AI," adding, "It could help in the development of personalized teaching methods or brain stimulation techniques."

References

Science Advances (2025), DOI: https://doi.org/10.1126/sciadv.adq9990