'Secret code' behind key type of memory revealed in new brain scans

The "secret code" the brain uses to create a key type of memory has finally been cracked. 

This type of memory, called working memory, is what allows people to temporarily hold on to and manipulate information for short periods of time. You use working memory, for example, when you look up a phone number and then briefly remember the sequence of digits in order to dial, or when you ask a friend for directions to a restaurant and then keep track of the turns as you drive there.

The new work represents a "fundamental step forward" in the study of working memory, Derek Nee, an assistant professor of psychology and neuroscience at Florida State University, told Live Science in an email.

A critical process 

For decades, scientists have wondered how and where the brain encodes transient memories. 

One theory suggests that working memory relies on special "storehouses" in the brain, separate from where the brain handles incoming sensory information from the eyes or nose, for instance, or where long-term memories — like memories of who you attended prom with, or foundational knowledge you learned in school — are stored, said Nee, who was not involved in the new study. 

Another, opposing theory suggests that "there are no such special storehouses," Nee told Live Science. In this alternate theory, working memory is essentially an emergent phenomenon — one that shows up "when sensory and motor representations are kept around as we link the past to the future," Nee said. According to this theory, the same brain cells light up when you first read through a phone number as do when you run through that number again and again in working memory.

Related: Your brain exaggerates memories to remember them better 

The new study, published April 7 in the journal Neuron, challenges both of these theories. Rather than reflecting what happens during perception or relying on special memory storehouses, working memory seems to operate one step up from sensory information gathering; it extracts only the most relevant sensory information from the environment and then sums up that information in a relatively simple code. 

"There have been clues for decades that what we store in [working memory] might be different from what we perceive," study senior author Clayton Curtis, a professor of psychology and neural science at New York University (NYU), told Live Science in an email. 

To solve the mysteries of working memory, Curtis and co-author Yuna Kwak, a doctoral student at NYU, used a brain scanning technique called functional magnetic resonance imaging (fMRI), which measures changes in blood flow to different parts of the brain. Active brain cells require more energy and oxygen, so fMRI provides an indirect measure of brain cell activity.

The team used this technique to scan the brains of nine volunteers while they performed a task that engaged their working memory; the two study authors also completed the task and contributed brain scans to the study. 

In one of the trials, the participants viewed a circle composed of gratings, or slashes, on a screen for roughly four seconds; the graphic then disappeared, and 12 seconds later, the participants were asked to recall the angle of the slashes. In other trials, the participants viewed a cloud of moving dots that all shifted in the same direction, and they were asked to recall the exact angle of the dot cloud's motion.  

"We predicted that participants would recode the complex stimulus" — the angled grating or moving dots — "into something more simple and relevant to the task at hand," Curtis told Live Science. Participants were only asked to pay attention to the orientation of the slashes or angle of the dot cloud's motion, so the researchers theorized that their brain activity would reflect only those specific attributes of the graphics. 

And when the team analyzed the brain scan data, that's just what they found.

Related: Sherlock Holmes' famous memory trick really works 

The researchers used computer modeling to visualize the complex brain activity, creating a kind of topographical map representing peaks and valleys of activity in different groups of brain cells. Brain cells that process visual data have a specific "receptive field," meaning they activate in response to stimuli that appear in a particular zone of a person's visual field. The team took these receptive fields into account in their models, which helped them understand how the participants' brain activity related to what they'd observed on-screen during the memory task.  

This analysis revealed that, instead of encoding all of the fine details of each graphic, the brain stored only the relevant information needed for the task at hand. When viewed on the topographical maps, the brain activity used to encode this information looked like a simple, straight line. The angle of the line would match the orientation of the gratings or the angle of the dot cloud's motion, depending on which graphic the participants had been shown.

These line-like brain activity patterns appeared in the visual cortex, where the brain receives and processes visual information, and the parietal cortex, a key region for memory processing and storage. 

What's crucial isn't that the brain settled on using lines to represent the images. "It is the fact that the representation has been abstracted from gratings [or] motion to something different," Nee said. 

One limitation of the study is that the team used very simplistic graphics, which don't necessarily reflect the visual complexity of the real world, Nee noted. This limitation extends to many studies of working memory, and Nee said he uses similar simple graphics in his own research.

"The field will need to move towards richer stimuli that better match our natural visual experiences to bring us from the laboratory to practical utility," he said. But with that in mind, the new study still "provides a novel insight into what it means to hold something online in mind for the future," he said. 

Working memory essentially acts as a bridge between perception (when we read a phone number) and action (when we dial that number). "This study, in identifying a representational format that resembles neither what was perceived nor what will be done but can be clearly read out from visual signals, offers an unprecedented look into this mysterious intermediate zone between perception and action," Nee said.

Originally published on Live Science.