By Erin Digitale
On a simple math task — indicating which of two amounts is greater — kids with math learning disability get the right answer as often as their good-at-math peers, but behind the scenes, their brains are working differently, a new Stanford Medicine study has found. The differences shed light on what causes their math struggles.
The findings, which will be published online Feb. 9 in the Journal of Neuroscience, show that children with a math learning disability are less likely to solve problems at the right speed, or to slow down after they make mistakes, particularly while working with number symbols. Their brain scans show different levels of activity in centers involved in executive function — which is similar to air traffic control, enabling focused attention — and checking for errors.
The discoveries point toward new ways to help kids with math learning disability, the researchers said.
"Our findings suggest that interventions should target not only basic number sense, but also metacognitive processes, like performance monitoring — how do you adjust when you notice an error?" said senior author Vinod Menon , PhD, the Rachael L. and Walter F. Nichols, MD, Professor and a professor of psychiatry and behavioral sciences. "We need to provide these children with feedback and training to build those cognitive skills."
While having their brains scanned via functional magnetic resonance imaging, kids in the study answered problems that were posed in two ways: They compared different-sized groups of dots, and compared Arabic numerals.
"What we found was that children with math learning disability show specific difficulties with symbolic numbers, and particularly with updating their strategy as they work with number symbols," said co-lead author Hyesang Chang, PhD, a former research scientist at Stanford Medicine. In other words, kids with math learning disability are less likely to adjust after making mistakes, she said.
Research scientists Percy Mistry , PhD, and Yuan Zhang , PhD, share lead authorship of the paper with Chang.
Early help keeps math learners on track
The study focused on children in second and third grades. Of the 87 children whose data was analyzed, 34 had math learning disability, meaning they scored at or below the 25th percentile on a standard test of math fluency; the other 53 children had higher scores indicating typical math-learning ability.
Math learning disability can be defined in different ways. The criteria used in this study — placing in the bottom quarter of scores on a standard test — is fairly broad, which was intentional, Menon said, as it helps make the findings applicable to many children who struggle with math. Some people, about 3% to 7% of the population, have a more strictly defined form of math learning disability called dyscalculia that includes difficulties understanding and comparing quantities, learning to count, understanding number symbols and learning math-based skills such as telling time.
Identifying students with math learning disabilities early and offering prompt, effective help is important for keeping their education on track, Menon said.
"If you're not doing well, you lose interest and motivation, and you may get more anxious during problem solving because you feel you're not good at it," he said. "It's a cascading set of problems; it becomes a bottleneck to further learning."
For the study, the researchers chose the simple task of comparing two quantities because they wanted to assess differences in the children's brain processes as they worked with numbers, independent of differences in their math performance. While in the MRI scanner, children saw pairs of quantities and pressed a button to indicate which of each pair was larger. Some problems used dots to show quantities; others used Arabic numerals. Problems were classified as easy (those with a large gap between the numbers, such as 7 versus 2) or hard (separated by only one number, such as 6 versus 7).
Scans show differences
Although children with and without math learning disability had similar rates of correct answers, data from their responses and brain scans revealed differences between the two groups.
The researchers developed a computational model to assess factors such as how cautious participants were about deciding which number was larger, on both easy and hard problems, and how readily participants detected their own mistakes and slowed down or changed strategies for the next problem after making an error. (Participants were not told if they had made a mistake.)
On problems with numeric symbols, children with typical math abilities slowed down more for harder comparisons than easy ones, while children with math learning disability didn't modify their strategy as much. Those with math learning disability were also less likely to slow down or change their behavior after they made an error than children with typical math ability.
However, for problems presented as groups of dots, children with math learning disability were actually more cautious after making an error.
"Many of these kids — unless their disability is severe — have normal representation of non-symbolic quantities, so they can tell five from 10 dots quite easily, but when you ask them to reason with and manipulate number symbols, they become deficient," Menon said.
Brain scans revealed patterns that lined up with these behaviors. While doing problems with number symbols, children with math learning disability had less neural activity in the middle frontal gyrus, which has roles in executive function and in sustaining and directing attention, and in the anterior cingulate cortex, which detects errors and helps with decision making and impulse control.
"Seeing weaker function in kids with math learning disability in those brain regions — which are known to be involved in executive function and error monitoring — suggests that they may be using those brain resources less than is necessary to solve these problems efficiently," Chang said.
The findings could help educators craft ways to help struggling students become better at math as well as other types of problem-solving.
"One of our surprising findings is that even in the absence of an overt difference in the kids' behavior, we could pick up strong signals about what their brains are doing behind the scenes," Menon said. "It gives us insight that how we reason — how we think about problems and adjust our problem-solving behavior — is just as important as having a core domain of knowledge."
The research was supported by grants from the National Institutes of Health (HD094623, HD059205 and MH084164), the National Science Foundation (DRL-2024856), and the Stanford Maternal and Child Health Research Institute Postdoctoral Support Awards.
