Q&A: Understanding the Connection Between Memory and Math

UVA professor Tanya Evans is studying the connections between memory and math – and, in the process, illuminating important building blocks of how children learn.  

An associate professor at the School of Education and Human Development, Evans is a developmental cognitive neuroscientist, meaning she studies what happens in the brain as children learn new skills. Her work aims to help us better understand the most basic, essential skills that create a foundation for learning.

One recent study, published this summer, found a link between simple memory tasks and later math performance for elementary schoolers.

We caught up with Evans to learn more about her research and what the latest findings mean for educators and parents.

"Math learning goes beyond the things we typically think of as math skills – like numbers, patterns and problem-solving. Problem-solving skills are important too, of course. But without a solid foundation, there's nothing to build upon."

Q: In this research, you distinguish between two types of memory: declarative and procedural. Can you explain what the difference is?

Declarative memory is fact-based learning, with quick retrieval – like remembering that the capital of the United States is Washington, DC, or that two plus two equals four. In the brain, this type of memory is stored in the hippocampus, which is located in the temporal lobe.

Procedural memory is remembering how to perform a task, like riding a bike, or using different procedures to solve a math problem. With practice, these tasks can become more automatic, but that generally takes time. It’s rooted in the basal ganglia, so it’s a different brain structure than declarative memory.

Q: How was the study designed?

For this study, I partnered with Daniel Lipscomb here at UVA EHD as well as Laurie Cutting (Vanderbilt University) and Michael Ullman (Georgetown University), amongst other collaborators.  

This was a longitudinal study, meaning we studied the same group of children over several years. The data set included 109 second through fourth graders.

Children were asked to perform simple memory tasks, like showing them pictures of objects and later asking them to recall which objects they saw. We then compared their performance on the memory tasks to their later performance on math assessments.

By looking at these simple memory skills, which we know from prior research are associated with certain regions of the brain, we're able to really pinpoint whether or not declarative memory or procedural memory – or both – are supporting math skills.

It's a way for us to tap into the cognitive behavior without math getting in the way – to really know that we’re measuring memory, and not some other aspect of numeracy or format.

Q: What did you find?

We found that declarative memory is correlated with math skills in second through fourth grade, and also predictive of future math skills. So better performance on declarative memory tasks in second grade predicted better math skills in fourth grade.

This is evidence that basic memory tasks support future math learning, which tells us that math learning goes beyond the things we typically think of as math skills – like numbers, patterns and problem-solving.

Problem-solving skills are important too, of course. But without a solid foundation, there's nothing to build upon.

How might this work impact how schools or parents approach teaching math to young students?

This is still a pretty new area of research, but ultimately, if we find that one or both of these memory systems is very involved in the development of math skills, then we can leverage that knowledge to design curriculum, as well as interventions for those that are struggling with math.  

These findings suggest that fact-based memorization strategies, like flashcards, are an important foundational skill for math learning – particularly early in life. Something like playing simple matching memory games with young kids could contribute to better math skills down the line.

Q: In the study, you highlight the role of “domain-general learning.” Can you explain what that means, and why it’s important?

Declarative memory is what we call a domain-general aspect of cognition – a skill that contributes to learning across areas, not specific to one subject. Another example is executive function. We tend to think of certain domains like math and reading to be so separate, but this work highlights underlying skills that are important for multiple areas of learning.  

I think that’s important because math, as a domain, comes with a lot of social and affective baggage. When people struggle to read, we readily accept that they need intervention because reading is such a critical skill to have in life. But if people struggle in math, we almost seem to accept that they might be inherently “bad” at math.  

I would love to see us move away from “oh, some people are just bad at math” toward “what can we do to help every child gain these skills?” I think this research highlights that we can leverage these domain-general skills to support all types of learning.

Q: What comes next?

There’s still so much to learn in this field. I have an ongoing study as a collaboration with Ian Lyons (Georgetown) using fMRI brain scans, which also focuses on memory and math skill acquisition in early school years. The results will help us understand which memory brain system most closely overlaps with math processing, and how these brain systems and skills change over time in individual children. This will increase our understanding of the best interventions for children who struggle with math.

But I’m also really interested in this from a lifespan perspective – what happens as people age? Math skills are important throughout life.  

It's thought that the declarative memory system is the first to deteriorate as we age. If that's the case, how do we support math skills into later ages, or in people managing things like Alzheimer's disease, when these systems are really hit hard?

Better understanding these underlying systems could illuminate a lot throughout the entire lifespan.