Machine learning in the classroom: AI can assess learning disabilities, says study
For educators seeking to identify learning difficulties of children struggling at school, machine learning could offer a powerful, new way to assess their cognitive skills, according to researchers from the Medical Research Council (MRC) Cognition and Brain Sciences Unit at the University of Cambridge.
The team used machine learning to identify clusters of learning difficulties which did not match the previous diagnosis the children had been given.
"Our study is the first of its kind to apply machine learning to a broad spectrum of hundreds of struggling learners."
Published in Developmental Science, the study involved 550 children who were struggling at school and referred to the Centre for Attention Learning and Memory. What made this study noteworthy was the inclusion of children with all difficulties regardless of diagnosis — earlier research had dealt with children who had already been given a particular diagnosis (attention deficit hyperactivity disorder (ADHD), an autism spectrum disorder, dyslexia, etc). With this comprehensive approach, scientists could observe a more complete picture of the variation of difficulties within, as well as the overlap between the diagnostic categories.
"Receiving a diagnosis is an important landmark for parents and children with learning difficulties, which recognizes the child's difficulties and helps them to access support," Dr. Duncan Astle of the MRC Cognition and Brain Sciences Unit at the University of Cambridge, and lead on the study, said. "But parents and professionals working with these children every day see that neat labels don't capture their individual difficulties—for example one child's ADHD is often not like another child's ADHD. Our study is the first of its kind to apply machine learning to a broad spectrum of hundreds of struggling learners."
Their computer algorithm processed massive amounts of cognitive testing data from each child, including measures of listening skills, spatial reasoning, problem solving, vocabulary, and memory. From these data, four clusters of difficulties were identified:
- Difficulties with working memory skills (commonly linked with math struggles);
- Difficulties with processing sounds in words (associated with reading struggles);
- Children with broad cognitive difficulties in many areas;
- Children with typical cognitive test results for their age.
Among other findings, machine learning revealed how diagnostic labels can obfuscate other learning disabilities that may be present, emphasizing the need for better interventions that address cognitive needs on an individual level.
"Past research that's selected children with poor reading skills has shown a tight link between struggling with reading and problems with processing sounds in words," Dr. Astle adds. "But by looking at children with a broad range of difficulties we found unexpectedly that many children with difficulties with processing sounds in words don't just have problems with reading—they also have problems with maths.”
Dr. Joni Holmes, senior author on the study, emphasizes the need for more precise interventions: "Our work suggests that children who are finding the same subjects difficult could be struggling for very different reasons, which has important implications for selecting appropriate interventions."
The application of machine learning to decipher the neuroscience behind learning difficulties is showing early signs of promise, according to Dr. Joanna Latimer, head of neurosciences and mental health at the MRC.
"These are interesting, early-stage findings which begin to investigate how we can apply new technologies, such as machine learning, to better understand brain function. The MRC funds research into the role of complex networks in the brain to help develop better ways to support children with learning difficulties."