Working memory is an important cognitive activity.
The “hidden code” that the brain employs to generate a specific sort of memory has now been deciphered.
Working memory is the sort of memory that allows people to temporarily hold on to and manipulate information for brief periods of time. Working memory  is used when you search up a phone number and then quickly recall the sequence of digits to dial, or when you ask a friend for directions to a restaurant and then keep track of the turns as you drive there.
According to Derek Nee, an assistant professor of psychology and neuroscience at Florida State University, the new research offers a “major step forward” in the study of working memory.
A crucial procedure
Scientists have been wondering for decades how and where the brain encodes fleeting memories.
According to one theory, working memory is stored in separate “storehouses” in the brain from where the brain handles incoming sensory information from the eyes or nose, or where long-term memories — such as who you went to prom with or foundational knowledge you learned in school — are stored, according to Nee, who was not involved in the new study.
Another alternative theory contends that “no such unique storehouses exist.” Working memory, according to this alternative view, is fundamentally an emergent process that appears “when sensory and motor representations are maintained around as we link the past to the future,” according to Nee. According to this theory, the same brain cells light up when you initially read a phone number as they do when you run that number through working memory again and again.
The latest study, published on April 7 in the journal Neuron, calls both of these views into question. Working memory appears to operate one step above sensory information gathering, extracting only the most relevant sensory information from the environment and then summarising that information in a relatively simple code, rather than reflecting what happens during perception or relying on special memory storehouses.
“There have been hints for decades that what we store in [working memory] may be different from what we experience,” says research senior author Clayton Curtis, a psychology and neuro science professor at New York University (NYU).
Curtis and co-author Yuna Kwak, a doctoral student at NYU, utilised a brain scanning method called functional magnetic resonance imaging (fMRI) to answer the puzzles of working memory. fMRI detects changes in blood flow to different areas of the brain. Because active brain cells require more energy and oxygen, fMRI measures brain cell activity indirectly.
The researchers utilised this technology to scan the brains of nine volunteers while they completed a task that required working memory; the two study authors also completed the task and provided brain scans to the study.
In one experiment, participants were shown a circle made of gratings, or slashes, on a screen for around four seconds; the visual subsequently vanished, and the participants were asked to recall the angle of the slashes 12 seconds later. In other trials, participants were shown a cloud of moving dots that all shifted in the same direction, and they were asked to recall the exact angle at which the dot cloud moved.
“We hypothesised that participants would recode the complicated stimulus” — the angled grating or moving dots — “into something more simple and relevant to the task at hand,” the researchers wrote. The researchers hypothesised that because participants were only required to pay attention to the orientation of the slashes or the angle of the dot cloud’s motion, their brain activity would reflect only those specific aspects of the visuals.
That’s exactly what the scientists discovered after analysing the brain scan data.
The researchers visualised the intricate brain activity using computer modelling, creating a kind of topographical map reflecting peaks and valleys of activity in distinct groupings of brain cells.
Brain cells that process visual data have a distinct “receptive field,” which means they respond to stimuli that appear in a certain zone of a person’s visual field.
The scientists incorporated these receptive fields into their models, which allowed them to better understand how the participants’ brain activity connected to what they saw on-screen during the memory exercise.
This investigation demonstrated that, rather of encoding all of the fine features of each visual, the brain merely stored the information required for the job at hand.
The brain activity used to represent this information seemed to be a simple, straight line when seen on topographical maps. Depending on which picture the participants were shown, the angle of the line would correspond to the orientation of the gratings or the angle of the dot cloud’s motion.
These line-like brain activity patterns were observed in the visual cortex, which receives and processes visual information, as well as the parietal cortex, which is important for memory processing and storage.
What’s important isn’t that the brain chose to depict the images with lines. “It’s because the representation has been abstracted from gratings [or] motion to something else,” Nee explained.
According to Nee, one limitation of the study is that the team employed very rudimentary images that may not always reflect the visual complexity of the real world. This constraint is shared by many studies of working memory, and Nee stated that he employs comparable simplistic images in his own study.
“In order to progress from the laboratory to practical utility, the discipline will need to move toward richer stimuli that better reflect our normal visual experiences,” he said. Despite this, he believes the current study “provides a novel perspective into what it means to hold anything online in mind for the future.”
Working memory serves as a link between perception (such as when we read a phone number) and action (when we dial that number). “This study offers an unparalleled glimpse into this mysterious intermediary zone between perception and action by defining a representational format that matches neither what was perceived nor what would be done but can be clearly read out from visual data,” Nee said.