Brain reorganizes to make room for math

WASHINGTON — It takes years for children to master the ins and outs of arithmetic. New research indicates that this learning process triggers a large-scale reorganization of brain processes involved in understanding written symbols for various quantities.

The findings support the idea that humans’ ability to match specific quantities with number symbols, a skill required for doing arithmetic, builds on a brain system that is used for estimating approximate quantities. That brain system is seen in many nonhuman animals.

When performing operations with Arabic numerals, young adults, but not school-age children, show pronounced activity in a piece of brain tissue called the left superior temporal gyrus, says Daniel Ansari of the University of Western Ontario in London, Canada. Earlier studies have linked this region to the ability to associate speech sounds with written letters, and musical sounds with written notes. The left superior temporal gyrus is located near the brain’s midpoint, not far from areas linked to speech production and understanding.

In contrast, children solving a numerical task display heightened activity in a frontal-brain area that, in adults, primarily serves other functions.

Ansari presented his findings November 19 at the annual meeting of the Society for Neuroscience.

“Left superior temporal regions may also be responsible for mapping numerical symbols onto quantities,” remarks Filip van Opstal of Ghent University in Belgium, who studies adults’ neural responses to number tasks.

In addition, Ansari and his colleagues find that nearby parts of the brain, in the parietal cortex, contribute far more to both number understanding and the ability to estimate quantities in adults than they do in children. At the same time, both types of numerical knowledge recruit the prefrontal cortex far more in youngsters than in adults, according to the scientists.

“Our results demonstrate that the brain basis of number processing changes as a function of development and experience,” Ansari says.

The new findings support the idea that symbolic number use unique to people builds on an evolutionarily ancient brain system many animals share for estimating approximate quantities. In the past five years, studies of adult people and monkeys have suggested that parts of both the parietal and prefrontal cortex foster quantity estimates and symbolic number knowledge, with a specific parietal region looming especially large in adult humans. But little is known about quantity-related neural activity in kids.

Ansari’s new study consisted of 19 children, ages 6 to 9, and 19 adults, ages 18 to 24. Participants first viewed pairs of Arabic numerals, ranging from 1 to 10, and indicated which number was larger. Volunteers then viewed pairs of images showing arrays of one to 10 squares and indicated which array contained more squares. During these tasks, a functional MRI scanner measured where blood flow changed in the volunteers’ brains, providing a glimpse of rises and falls in neural activity.

Young adults performed the tasks more accurately than children did. But like kids, these adults took increasingly longer to discriminate between two numbers or two arrays as quantities got closer. So, it took longer to tell 2 apart from 1 than 9 apart from 1.

Correspondingly, one part of the parietal cortex in young adults, but not in children, grew increasingly active as pairs of numerals or quantities got closer. This area aids in initial efforts to translate knowledge about approximate quantities into comprehension of symbolic numerals, Ansari hypothesizes. With increasing math experience, the left superior temporal gyrus assumes major responsibility for symbolic number knowledge, he suspects.

Disturbances in that region and in nearby parietal areas may lie at the root of a dyscalculia, a childhood disorder characterized by an inability to conceptualize numbers and understand arithmetic, Ansari adds.

In related research presented at the neuroscience meeting, Ilka Diester of Stanford University reported that monkeys trained to associate Arabic numerals with corresponding quantities in dot arrays show robust prefrontal cortex activity but little parietal activity. Monkeys, like children, may achieve a budding grasp of numerals with the help of the prefrontal cortex, Diester proposes.