New Evidence Refutes Longstanding Theories About Working Memory

ARTICLE IN BRIEF

A new study challenges the idea that short-term memory depends on an elevated and sustained activation of neurons. Instead, it finds that working memory may be encoded in a silent latent manner that may be reactivated.

Scientists at the University of Wisconsin–Madison presented evidence that information sitting in short-term memory does not depend on an elevated and sustained activation of neurons, as has been the dogma for decades. Instead, they proposed that a growing body of evidence suggests that there is no one special place to call working memory, and that short-term memories are spread throughout the brain, depending on what someone is remembering.

The researchers measured working memory by recording neural activity using encephalography (EEG) and transcranial magnetic stimulation (TMS), and whether different or other more sensitive technologies would find anything different remains to be seen, experts said. But they agreed that if the current finding, reported in the December 2 edition of Science, is replicated, it may lead scientists to rewrite or rethink ideas about working memory.

“Although there is a lot of activity in the prefrontal cortex, there is little evidence that this activation corresponds to the short-term storage of information,” said Bradley R. Postle, PhD, professor of psychology and psychiatry and senior author on the study. “We are taught that this is the way that short-term memory works, that information is shuttled into the prefrontal cortex and it stays in an active mode until it is shuttled into long-term memory storage in the hippocampus. Our work suggests a different scenario.”

STUDY METHODS, FINDINGS

In the current study, the scientists used an encephalogram (EEG) to record brain activity while volunteers were shown two items from a set of three — a face, a word, and a pattern of dot motion — on a computer screen. They would first be asked to remember both items (say, a word and a face); then, after the items has disappeared from the screen, a cue indicated which of the two would be the first to be tested.

This experiment allowed the researchers to see what happens when the second (uncued) item was waiting in working memory. With the subject's attention focused on the first item, the scientists thought that they would see a weaker signal for the uncued item in the EEG. Instead, they saw a complete vanishing of the signal. However, study participants could still recall it later, so it seemed as if the uncued item was still somehow in working memory, but “silent” rather than represented by continuous neural activity.

Then, they used transcranial magnetic stimulation (TMS) to stimulate the area that was activated during the memory test. The TMS produced a temporary reactivation of working memory for the uncued item. It was a weaker signal and short-lived, but it influenced participants' behavior and affected their memory performance only when the item was potentially relevant later in the trial, said Dr. Postle, as if the TMS-triggered memory was interfering with the cued memory. If participants knew that the item would not be relevant, the signal disappeared and could not be re-activated, even with TMS. This suggests that “the representation is dynamic and modifiable via cognitive control” because the participants' knowledge of whether the item might need to be remembered or could be completely forgotten influenced whether it remained in a silent, yet available, state, or whether it would be lost completely.

“This finding suggested that the information in working memory (but outside of focal attention) can be maintained in a latent state via mechanisms other than sustained, elevated activity,” the researchers wrote in the paper. There was activity in the prefrontal cortex but it was not specific to the memory, as Dr. Goldman-Rakic and others had predicted, said Dr. Postle.

It may be that working memory resides in patterns of connectivity between neurons, he added, and it may not require sustained, elevated brain activity.

In another study, they used TMS to provoke an uncued item that was no longer relevant in the memory trial. They argued that “if the TMS reactivation effect is a consequence of an item being maintained in a privileged state it should only be observed when that item is still potentially relevant for the trial.” And that is what they found.

EXPERTS COMMENT

“This is an important development,” said Bettina Buhring, PhD, program chief of the substrates of learning and memory program at the National Institute of Mental Health. “It had been thought that neurons in the prefrontal cortex had to be persistently active when information is held there in working memory. However, this finding shows that working memory might be encoded in a ‘silent’ way, such as via the strength of connections between neurons.”

“The study is important because it is the first to show that TMS can trigger the reactivation of an item, or more precisely, the category of an item in working memory,” said Clayton Curtis, PhD, associate professor of psychology at New York University. “The theoretical implications are that the brain may employ multiple mechanisms to support working memory. Although the results do not call into question the more traditional forms of persistent neural activity that have been known for 40 years to be critical for sustaining working memory representations, they indicate that yet undiscovered alternative mechanisms, such as fast synaptic plasticity, may play a role.”