Researchers at the School of Medicine’s Stanford Brain Development Project have recently discovered that MRI scans—rather than IQ tests or academic achievement measures—can most successfully predict how much a student’s math skills will improve after intensive tutoring.

The team’s results were detailed in a study co-authored by Professor of Psychiatry and Behavioral Sciences Vinod Menon, postdoctoral scholar Kaustubh Supekar Ph.D. ’10 P.D. ’10 and research assistants Anna Swigart and Caitlin Tenison. Postdoctoral scholars Dietsje Jolles and Miriam Rosenberg-Lee also collaborated on the study.

According to Rosenberg-Lee, the researchers found that certain features of a student’s hippocampus—one of the major components of the brain’s limbic system—determined how much the student’s math skills improved after tutoring.

“Children who had a larger hippocampus got more out of our training,” Rosenberg-Lee said. “We found that the connections between this memory area and other parts of the brain also predicted how much these kids would get out of our training.”

The team recruited 24 eight and nine-year-old participants from schools in Palo Alto and around the Bay Area by distributing flyers that encouraged students who struggled with math to sign up for the study.

At the beginning of the study, the participants went through a “pretty thorough” series of standardized tests that included measurements of IQ, reading comprehension and other cognitive skills, according to Rosenberg-Lee. The researchers also assessed the participants’ arithmetic performance, which was later used to measure how much each student improved after tutoring.

For the next four weeks, each student received one-on-one math tutoring for 40 minutes a day, three times a week. The lab’s research assistants and Stanford students served as tutors, focusing on teaching the participants to memorize and recall math facts.

“Everybody got better at this arithmetic task after doing the tutoring, but there was wide variation in how much they improved,” Rosenberg-Lee said. “Some kids improved only eight percent, while others improved 198 percent, and so what we really wanted to understand is what predicted how much kids were going to improve.”

According to Rosenberg-Lee, none of the standardized tests administered at the beginning of the study, such as the IQ test, accurately predicted which students would benefit the most from tutoring, disproving a commonly-held belief that students who have well-developed math skills will benefit more from tutoring than those who struggle with math.

“It wasn’t the case that the kids who were good at math were the ones who learned the most,” Rosenberg-Lee said. “At all different levels of ability, it was how big the brain area was that predicted how much a student was going to learn.”

Swigart noted that the study was a continuation of the lab’s investigation into how math skills are developed and whether or not students who struggle with math have the potential to improve.

Though Swigart said that the team’s findings were significant, she cautioned against assuming that the size of a student’s hippocampus could predict how much a student would benefit from all types of math tutoring, as the team’s research was focused on testing memorization of math skills.

“Right now [the study] won’t have direct implications other than saying that there are individual differences in learning and we can predict from the brain which children will benefit from this particular training more than other children,” Swigart said. “But it could be that there are other types of training that other children would benefit more from, and that’s something that we don’t know.”

The researchers hope to conduct a series of follow-up studies ranging in focus from how students feel about being tutored in math to further investigation into numerosity, which Swigart described as one of the “core principles that math cognition is built on.”

Rosenberg-Lee noted that the team also hopes to investigate what makes a successful tutoring program, and which types of math tutoring help students improve the most.

“We found that the tutoring was very effective, but we didn’t know how much was due to flashcards or understanding the concepts,” Rosenberg-Lee said. “For our follow-up study, we are going to try to pull those apart.”