Rethinking Learning Difficulties

Using artificial intelligence and a data-based approach, a group of researchers found that the cognitive profiles of struggling learners corresponded with specific brain organizations, but not with common diagnoses such as ADHD.

• According to the authors of a new study on the topic, common diagnoses such as ADHD constitute a poor understanding of learning difficulties, because
• They overemphasize within-group homogeneity. “It is widely documented that symptoms vary between children with the same diagnosis” (all quotes are from the original article referenced below)
• They overemphasize between-group differences. “This approach can fail to accommodate the high rates of comorbidity within developmental disorders.”
• They underestimate the scope of the problem. They “[do] not capture the majority of struggling learners” many of whom do not meet official diagnostic criteria. Ineed, prevalence rates for developmental disorders linked with learning difficulties range from 3-8%. But the number of children who struggle at school is far higher.

As a consequence, the team decided to work on an alternative, empirically-based classification system. In their words: “The aim of this approach is to move away from identifying highly selective discrete groups and instead focus on identifying continuous dimensions that distinguish individuals and can be used as potential targets for intervention.”

To do so, the researchers adopted a data-driven approach, using artificial intelligence to “map” the learning difficulties of 530 children who had been referred to the Centre for Attention Learning and Memory (CALM) by health and education professionals for problems in attention, memory, language, or poor school progress.

First, the participants took a battery of cognitive tests assessing their fluid intelligence, phonological processing, working memory, and executive functions. Their spelling, reading, and mathematical ability were also measured, as well as their communication skills.

Then, the team crunched the data using a type of machine-learning that is capable of representing a multidimensional space on a “map” where proximity indicates a similarity in cognitive profile.

Results indicated that the children fell under four categories:

• Children with broad cognitive difficulties and severe reading, spelling, and maths problems.
• Children with working memory problems.
• Children with phonological difficulties.
• Children with age-typical cognitive abilities and learning profiles.

Interestingly, a child’s cognitive profile was not predicted by diagnosis or referral reason.

What is more, the researchers proceeded to scan the brains of 184 of their participants, and found specific differences in white-matter connectivity between each of the four groups (and a control group).

The authors explain: “The brain can be modelled as a network of brain regions connected by white matter…” As such, “white matter maturation is a crucial process of brain development” that is “related closely to cognitive development.”

This study thus “represents a novel move towards identifying data-driven neurocognitive dimensions underlying learning-related difficulties” and could help design innovative and more adapted strategies to support struggling learners.

Reference: Astle, Bathelt, Holmes, and CALM team (2019), “Remapping the cognitive and neural profiles of children who struggle at school”, Developmental Science, 22:1.