Why Working Memory May Fade


by Carl Sherman

August 17, 2011

Forgetting a name, plot details of a movie, what groceries we intend to buy. As we age we accept (if lament) such “senior moments.”    

But memory is more than brief (short-term) and lasting (long-term) information storage and retrieval. We process bits of data while we keep them in mind—adding a column of figures, thinking of what we’ll say next as we take in the other side of a conversation—with a more fluid faculty called working memory.

This sometimes overlooked capacity is beginning to receive the attention it deserves. The role of working memory in the classroom has attracted the interest of educational researchers [see “Improving Memory to Improve Academic Performance”], and there’s a growing appreciation of its importance throughout life.

“Working memory is such a basic process, it pervades much of what we do cognitively,” says Elizabeth Zelinski, professor of gerontology and psychology at University of Southern California. “When people complain about memory problems, a lot of their difficulty probably involves working memory more than simply being unable to remember things.”

Working memory declines as part of normal aging, although not dramatically, Zelinski says. “Although individuals vary, in general the difference between 18-year-olds and 70-year-olds is about one third of a standard deviation. It’s not tremendous, but it’s there.”

Anything that interferes with working memory makes the small losses of aging loom larger. “A momentary distraction makes it difficult to manage things,” Zelinski says. The thing we’re holding in mind can be readily dislodged when our attention turns elsewhere.

“Working memory is very fragile,” says Adam Gazzaley, director of the Neuroscience Imaging Center at University of California in San Francisco. “There’s a lot more public awareness of this, driven by the increasingly sophisticated ways that technology provides interference throughout the day,” he says.

“People use [the term] ‘distraction’ to cover irrelevant information that surrounds us and that we need to suppress in pursuit of an immediate goal,” Gazzaley says. “But it’s clear to me there’s another kind of interference—we refer to it as ‘interruption’—that is engaged as a secondary task.” The resulting fugue of intentions, information streams, and goals is what most people call “multitasking,” as embodied in such common activities as texting while watching TV.

It has been suggested that older people are worse at multitasking then their juniors, and research by Gazzaley and his colleague Wesley Clapp may explain why.

The problem with multitasking

In a study reported in Neurobiology of Aging in 2010 (PDF), the researchers showed young (mean age: 23) and older (mean age: 69) participants a face, told them to hold it in memory and then asked them to decide which of three faces matched the original—a test of working memory.

In a second session, the volunteers took the same test but were distracted by another face that they were told to ignore. The third time, they were asked to make a judgment about the other face (was it a male over 40 years?) an “interruption” that is considered to correspond to multitasking.

As expected, the older group performed more poorly in the test of working memory. Distraction made their performance worse, and interruption worst of all.

Electrical activity patterns measured via EEG suggested that the brains of older participants devoted extra attention to the distracting face, despite being told it was irrelevant. “The young subjects were able to suppress [their response to the distraction], and the older subjects were not,” says Clapp. This fits in with earlier studies suggesting a general weakening of the ability to inhibit brain responses as we grow older.

The researchers thought a similar failure of inhibition might explain poor performance in the “interruption” task. But here, older and younger groups allocated equal attention to the interrupting face. “It was an unsatisfying conclusion,” says Gazzaley. “We could see the mechanism was different, but we didn’t know what it was.”

To answer that question, the researchers enlisted fMRI for a more detailed picture of brain activity while similar participants dealt with a similar series of situations. In this study, which was reported in the April 10, 2011, Proceedings of the National Academy of Science (PDF), the image that participants viewed first and were asked to match later was a landscape scene, while the interfering image was a face.    

 The researchers analyzed functional connectivity—coordinated activity between the prefrontal cortex (PFC) and the fusiform face area, which processes information about faces, and between the PFC and the parahippocampal place area, which processes places.

Once more, the older group performed particularly badly in the multitasking “interruption” trial, although they had paid no more attention to the interrupting stimulus than the younger group. “What was interesting was what happened after the interrupter had gone away and they were trying to reactivate the scene they were holding in memory,” says Clapp.

In young participants, the PFC promptly disconnected from the face-processing area and re-engaged with the place-processing area. In older people, it didn’t work that way. “It was very hard for them to reactivate the original image, and they were holding on too long to the image that was distracting them,” Clapp says. Behind the age-linked difficulty in multitasking, it appeared, was a reduced ability to switch between networks.

The problem, he speculates, is in the frontal lobe, whose functions include directing attention. “What’s most likely happening with normal aging is that the frontal lobe is not as crisp and clear in its commands,” he says.

A study   reported in the July 27, 2011 issue of Nature supported this conjecture and suggested a molecular explanation for  working memory decline. Analyzing brain activity in young, middle aged, and old monkeys, Yale researchers identified a group of neurons in the PFC that continued firing during a working memory task. These “delay” neurons were significantly less  active  in the older animals, and the researchers found evidence that an accumulation of the signaling chemical cAMP in aged brains was responsible. 

This neurochemical change could well explain working memory’s increasing vulnerability to interference and the “switching” problems observed in the Gazzaley study, said lead researcher Amy Arnsten, professor of neurobiology at Yale. “For all these cognitive tasks, you need persistent network firing.”

Can anything be done to redress the working memory declines of age? Arnsten’s research suggests that  restoring a more “youthful” neurochemical environment might  help. When her team infused a drug, guanfacine, which is known to block cAMP signaling,  into the PFC of the aged monkeys,  “delay” neuron firing rose to youthful levels. Earlier research had found that guanfacine improved working memory in aged animals.

Guanfacine is well-known in medicine—it’s approved for high blood pressure and ADHD—and a clinical trial is underway to see whether it can improve working memory and other executive functions in older people. 

Non-drug interventions are promising as well. Gazzaley is testing a computerized program that involves intensive practice in rapidly disengaging and re-engaging with stimuli. “We want to see if we can push these skills in older adults, to improve both working memory and the ability to resist the negative impact of interference,” he says. “We’re months away from completing a large pilot study, but preliminary data are very encouraging.”

A variety of similar approaches have also had a measure of success.  

A 2010 study in the Gazzaley lab trained older adults to make progressively finer discriminations between visual images, on the theory that sharper mental representations could be more securely held in memory. Elsewhere, Elizabeth Zelinski was part of a research team that used cognitive training to improve a range of memory skills.

Other researchers have had participants—in one study, aged 80 and above— practice standard working memory tasks, like recalling digits several back in a series, increasing the difficulty as they mastered each level. “I believe there is something to working memory training; it seems to work—at least a bit,” Zelinski says.

To Gazzaley, the suggestion that this fundamental faculty can improve through practice is highly promising. “I think there’s a lot to be excited about in using the fact that the brain retains plasticity throughout the lifespan as a target[ for intervention].” But considerable further research, to validate the ability to overcome deficits, remains to be done, he says.