Brainwave signature shows how the sleeping brain decides which memories to keep

Scientists have identified a specific brainwave signature that appears to mark which memories the brain selects for long-term storage during sleep, according to a study published Thursday in the journal PLOS Biology. The finding offers the first direct human evidence for a decades-old theory that some experiences are "tagged" for overnight consolidation the moment they happen, while others are quietly left to fade.

The research, led by Dan Denis of the University of York and colleagues from King's College London, Baylor University and Ludwig-Maximilians-Universität München in Germany, used scalp electroencephalography (EEG) to watch 31 healthy young adults learn 160 word-object pairs. Roughly two hours later, after some participants had napped and others had stayed awake, the same items were tested again. People who had slept remembered the associations significantly better than those who had not — and the researchers could tell, purely from brain activity recorded during the original learning session, which specific memories would go on to benefit from that sleep.

A tag written in brainwaves

The marker turned out to be theta oscillations, rhythmic waves of electrical activity that pulse through the brain three to eight times per second. Using a machine-learning classifier applied to the EEG recordings, the team found they could distinguish, item by item, which memories had been earmarked for sleep-dependent consolidation versus which had not — but only for memories later tested after the delay, not for immediate recall. Items that showed a stronger theta response during learning were the ones more likely to survive intact after a nap; items encoded without that theta boost were more likely to be forgotten.

Critically, theta activity during learning did not predict memory performance directly. Instead, it predicted something that happened later, during sleep itself: how tightly two other brain rhythms — slow oscillations and sleep spindles — locked together. That coupling, in turn, was the strongest predictor of which memories survived the two-hour break. The chain of events suggested a relay: theta tags an experience at encoding, and that tag later recruits the spindle-coupling machinery that does the actual work of consolidation while the brain sleeps.

• 31 healthy young adults (mean age 20) completed the two-visit sleep and wake sessions
• Sleep boosted memory for word-object pairs over wakefulness (p = .019)
• Theta power rose 5.68% at learning for items later remembered after sleep, versus a 2.73% decline for items tested after wakefulness
• About 16% of sleep spindles were coupled to slow oscillations, coupling that independently predicted post-sleep recall (adjusted R² = 0.19)

Decades of theory, a missing piece of evidence

Researchers have long known that sleep is not a passive backdrop to memory but an active editor of it. Rodent studies going back years have shown that the hippocampus "replays" recent experiences during sleep, and human neuroimaging has linked slow oscillations and sleep spindles to overnight memory gains. What has been missing is a mechanism for how the brain decides, in the moment an experience happens, which of the thousands of things a person encounters each day are worth flagging for that nighttime replay. The York-led study is among the first to trace that selection process in humans from the instant of learning through to the sleep physiology that follows.

The data and analysis code behind the paper have been posted publicly on the Open Science Framework, allowing other labs to test the theta-tagging model against their own EEG datasets — a step the authors say will be necessary given the study's modest sample size of 31 participants, all of them young and healthy.

"How the brain decides what is important to remember and what can be forgotten is still a mystery. These new findings help to answer that question by uncovering, for the first time, a signature of neural activity that instructs the brain which experiences to process during sleep and form into long-term memories," the authors said in a statement accompanying the paper.

Why it matters beyond the sleep lab

Denis and his co-authors point to a potential clinical thread running through the finding. If theta activity governs which experiences get tagged for consolidation, disruptions to that tagging process could help explain why memory works differently in conditions such as depression, where patients often show a bias toward dwelling on negative experiences while positive ones fail to stick. The authors suggest overactive or misdirected tagging of negative material — and under-tagging of positive material — could be one mechanism behind that pattern, though the current study did not test clinical populations directly.

More broadly, the finding fits into an expanding body of work on sleep's role in mental health, learning disorders and aging-related memory decline, all areas where understanding the biological "sorting" of memories could eventually inform new interventions, from sleep-timed learning strategies to therapies aimed at correcting faulty tagging in psychiatric illness.

What happens next

The immediate next step, according to the authors, is replication in larger and more diverse samples, since the current cohort was limited to 31 young, healthy adults tested in a single laboratory setting. Researchers will also want to know whether the same theta-tagging signature appears for other kinds of memory beyond simple word-object pairs, such as emotional experiences, procedural skills or real-world autobiographical events, and whether it can be detected — or even manipulated — outside a controlled EEG session. Denis's group has previously studied how sleep can selectively preserve negative over positive memories, and the new tagging mechanism gives that line of research a candidate biological handle to test directly, including in people with depression or anxiety disorders where memory bias is a defining clinical feature.

For now, the result stands as a proof of concept rather than a clinical tool: a measurable signal, present in ordinary brainwaves during learning, that appears to determine which of the day's experiences the sleeping brain will keep.