How the Brain Learns

Researchers study timing, sensory systems, how regions connect.


By Marlene Cimons, National Science Foundation

If a teacher puts out too much new information in the final minutes of a class, students might have trouble “getting” it. If you have a final exam in six weeks, it might be better to study for it now and then again next week, rather than tonight and tomorrow.

“It depends, to some extent, on when you have to remember it,” says Gary Cottrell, a professor in the computer science and engineering department at the University of California at San Diego.  “Cramming the night before a test is okay, but if it’s something you need to know a month from now, spacing between study sessions makes a big difference.”

In other words, timing is everything.

This is the underpinning for the work of the Temporal Dynamics of Learning Center, whose goal is to understand the impact of timing in learning across brain and social systems.

“Time is an understudied variable,” says Cottrell, who directs the center.  “Timing is crucial in learning from the synaptic levels--connections between neurons--to long-time scales, like months and years.”

Learning occurs at any number of levels, among them, synapses and neurons, brain systems, motor behaviors, and in social interactions between teachers and pupils. Every time someone learns a new fact or interacts with another person, timing is a part of how the neurons function, in how sensory systems communicate, and how different regions of the brain connect with each other.

“We think timing can really be exploited to help us understand some of the basic ways in which information is integrated in our brains across a variety of timescales,” Cottrell says. “By understanding how the brain learns, we hope to improve education.”

For example, research shows that the underlying problem for at least some poor readers is their inability to perceive fast acoustic changes in speech sounds (phonemes). This “slow shutter speed” leads to a poor representation of the sound structure of the language, making it difficult for students to acquire the letter-sound correspondence rules for reading, according to Cottrell. 

“We believe that by investigating the temporal dynamics of learning we can change the capacity of children to learn, as well as change the environment to aid in learning,” Cottrell says. “Unfortunately, the study of the role of time and timing has been piecemeal at best; we aim to change that.”

The center, a National Science Foundation Science of Learning Center, began in 2006 and involves 40 researchers from across the United States, Canada and Australia. The scientists cross multiple disciplines, including machine learning, psychology, cognitive science, neuroscience, molecular genetics, biophysics, mathematics and education.

The center is based at the University of California at San Diego, with primary research partners at Rutgers University, University of California at Berkeley, and Vanderbilt University. NSF supports the center with about $3.5 million a year.

Center scientists are studying numerous areas, among them, the activity of synapses, which are structures that allow neurons (nerve cells) to pass electrical or chemical signals to other neurons; and brain systems, including brain waves and regions of the brain involved in forming memories.

In one experiment, for example, center scientists predicted that neurogenesis, or the addition of newly born neurons in the hippocampus, an area of the brain involved in forming and organizing memories, binds new memories together in time.  “They are the cells that group together things that you learned over a few days or weeks,” Cottrell says. 

Furthermore, they found that the hippocampus has “place cells” that are active for specific locations in an environment. In their experiment, the researchers trained rats to explore three distinct environments in the same room, introducing each new environment one of two ways: either spaced over the course of two to three weeks, or presented all at once in a single day. Rats that received training spaced over long periods of time had place cells that were active during exposure only to one of the three contexts, whereas rats trained to all three environments at once had more cells that were active in all three contexts. Reducing the number of newly born cells in the hippocampus also resulted in more cells active in all three contexts.