Med thumb 23776517421 98b012458b o

Let’s say you and a few friends (or coworkers or foreign dignitaries) are embroiled in debate about, say, super delegates or March Madness or the state of artisanal ice cream. Whatever the topic, you want to chime in, but someone else interjects before you can. And then another person does. No one’s making your specific point, so you keep your comment top of mind, and process other people’s contributions as you wait to speak. This mental juggling act — keeping some fact or idea on deck in your short-term memory while simultaneously processing other information — is called working memory. And it would be hard to get through a day of work or play without relying on it. 

Since the ‘70s, neuroscientists have understood the brain activity underlying working memory as a stream of continuous neural activity associated with the information that’s being temporarily stored (i.e., that comment you’re waiting to make). But, in a study published this month in the journal Neuron, an MIT-lead research team offers a new model of brain activity underlying the demanding mental process. To enable temporary storage of information, the research suggests, corresponding neurons actually fire in sporadic bursts separated by activity gaps, rather than in an ever-flowing stream. 

“By having these different bursts coming at different moments in time, you can keep different items in memory separate from one another,” said senior study author Earl Miller in a press release.

“By having these different bursts coming at different moments in time, you can keep different items in memory separate from one another.”

Working memory falls under the umbrella of executive function, a group of mental skills we use to assert cognitive control, which happens when we will ourselves to complete some goal-oriented task when auto-pilot behavior wouldn't get the job done. Executive function also includes reasoning, paying attention, task-switching, planning and displaying inhibition. The impact of sleep loss on cognitive performance, research suggests, varies considerably depending on the specific skill being measured, the severity of the sleep loss and the nature of the task used to assess its impact. Regardless of the task, however, sleep-starved people perform consistently worse as time goes on – it's the fatigue effect. But, working memoy appears to be particularly sensitive to sleep deprivation. In studies, exhausted subjects tend to struggle with working memory during even brief cognitive tasks. 

Neuroscientists use an arsenal of tools and techniques to determine what's happening inside the brain. In this case, they use electrodes implanted in brain tissue to measure the intensity of electric currents flowing through small groups of neurons. The signal generated by currents is known as local field potential. The old theory of continuous activity underlying working memory relied on a flawed method of analyzing local field potential. Researchers would record seconds or minutes of neural activity as study subjects (monkeys) performed a working memory task and then they'd calculate the average. But, measuring average activity across time won't catch fleeting, subtle activity changes that could be critical. So, the MIT team tried a new approach, in which they measured and compared moment-to-moment changes in local field potential. 

To do this, trained monkeys (literally) completed a working-memory task in which researchers showed them a sequence of three colored squares. Then, they re-showed them the sequence, but changed the color of one square. The monkeys had been trained to respond upon noticing any color changes. So, the experiment involved a quick, simple working-memory task: Monkeys had to learn something (colors of squares) and temporarily store that knowledge in their short-term memory. Then, they had to process other new information (second set of squares) and analyze it in the context of the stored knowledge. Overall, the task required two seconds of working memory.

During those two seconds, researchers found, electrical activity in prefrontal cortex neurons flared in separate, brief bursts. And those bursts occurred most frequently during the beginning of the task (when monkeys initially learned, or encoded, new information about the colored squares) and at the end of the task, when the monkeys retrieved that information from their short-term memory.

In the press release, Miller used a metaphor to explain the difference between the old and new methods of recording brain activity: “It’s like for years you’ve been listening to music from your neighbor’s apartment and all you can hear is the thumping bass part. You’re missing all the details, but if you get close enough to it you see there’s a lot more going on.”

Based on these results, according to study authors, it would make sense to use the same method to analyze brain activity underlying other cognitive functions, such as attention.