Scientists discover a way to erase bad memories

Sleep does more than rest the body – it reshapes memory. Erasing “bad” or painful memories and traumatic flashbacks could transform how we treat a wide range of mental health conditions.

Now, scientists have uncovered a promising new strategy: they can actually entice a sleeping brain can lean toward a newer, positive memory when it competes with an older, negative one.

The work focused on cues and competition. A cue can be a word, place, or smell. When one cue links to two experiences, the brain chooses which memory “wins,” often favoring the newer or more emotional one.

“We found that this procedure weakened the recall of aversive memories and also increased involuntary intrusions of positive memories,” explained the authors in their study, which is published in the peer-reviewed journal PNAS.

Bad memories and sleep

The team tested whether brief reminders played during non-REM sleep could tilt recall toward a newer positive association without broadly weakening memory.

The plan was to help the positive memory win the next time the cue appeared, not to remove the older experience.

The researchers used nonsense words as neutral cues to track exactly which memories competed and when those memories returned.

How the study was done

Healthy young adults visited the lab over several days. On the first main day, participants learned nonsense words paired with unpleasant pictures – negative but not traumatic.

Each word became a cue for one specific aversive image. Participants then slept at home so those new negative memories could stabilize.

The next evening, half the same nonsense words were paired with positive images – animals, infants, and pleasant scenes – creating interference: an older negative association and a newer positive one for those cues.

Bad memories and non-REM sleep

On the second night, during non-REM sleep (monitored by EEG), some nonsense words were played quietly, a stage that supports memory consolidation.

Participants did not wake, but their brains registered the sounds. Each sound reactivated the network tied to that cue.

Not every word was played; some served as controls. Items fell into two types: interference items (both positive and negative links) and negative-only items.

This let the team test whether sleep cueing broadly harmed negative memories or mainly supported the positive competitor.

The next morning, the scientists tested memory for the negative associations. Recall of the old negative image dropped for words that had gained a positive partner and were cued during sleep.

For negative-only words, cueing did not significantly reduce recall. The pattern did not indicate a broad loss; instead, cueing let the positive competitor take the lead more often.

Memory interruptions

The team also measured intrusions – moments when the “wrong” memory intrudes.

When participants tried to retrieve the old negative image, a positive image came to mind more often for cued interference words. In plain terms, the positive memory interrupted the negative one.

This effect did not appear for negative-only words, reinforcing the core claim: cueing favored the competitor; it did not generally suppress negative content in isolation.

Participants saw each cue and quickly judged whether it felt more positive or negative.

After sleep and cueing, snap judgments shifted more positive for interference items, suggesting a change in the cue’s emotional leaning, not just recalled content.

This matters because quick evaluations often guide behavior. If a cue feels less negative at first glance, people may approach it differently the next time they see it.

Sleep signals tied to bad memories

EEG recordings linked behavior to brain activity. During non-REM sleep, responses in the theta range were associated with stronger positive memories.

Larger cue-evoked responses predicted better later recall of the positive images, indicating active replay and reinforcement when the cue was heard.

This link supports targeted memory reactivation and offers a physiological marker for tuning timing and intensity in future work.

Decision model used

To describe the choice process, the team used a drift-diffusion model, which treats decisions as evidence accumulating toward one of two options (“positive” or “negative”).

After cueing in sleep, the estimated drift toward the positive boundary increased for interference items – meaning the brain moved faster toward “this feels positive” once a cued word appeared.

Because the model ties choice speed and direction to evidence, these results align with the shifts in recall and quick judgments; cueing altered decision dynamics, not just outcomes.

The sample included about three dozen participants who completed behavioral and EEG sessions.

Each person served as their own control: some words were cued and others were not, and some words had a positive competitor while others did not.

Learning and sleep were spread over several days to keep the positive associations newer than the negative ones.

Limitations of the study

The study used artificial lab memories under controlled conditions and short time frames. It did not test durability over months or effects on everyday behavior.

The work focused on non-REM sleep and did not measure how REM sleep might contribute.

These limits matter for interpretation: the findings do not erase or rewrite deeply personal experiences; they show a way to bias retrieval when competing links exist.

Sleeping away bad memories

The approach taken in this study aims to shift which memory comes up first during recall, not to erase the past.

By forming a positive association that competes with an older aversive one and then reactivating the cue during sleep, the positive memory moved ahead in recall, intrusions, and split-second feelings.

With careful development, this could inform strategies for people who struggle with intrusive negative memories, pending clinical tests, longer follow-ups, and attention to ethics and safety.

For now, the takeaway is clear: sleep offers a window when the brain reorganizes memory. With precise cueing, it may be possible to guide that process so a healthier association is more likely to be retrieved next time.

Not a cure-all – just a nudge in a better direction.

The full study was published in the journal PNAS.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–