Targeted Memory Reactivation

In 2007, researcher Jan Born and his team at the University of Lubeck had volunteers learn the locations of picture cards on a grid while a rose scent filled the room. That night, during deep sleep, they quietly piped the same rose scent back into the air. The next morning, those volunteers remembered significantly more card locations than a group who slept without the scent cue. Their brains had been nudged, without waking them up, to practice the exact memory they needed.

By the end of this lesson you will be able to: (1) define targeted memory reactivation (TMR) and explain the sensory-cue mechanism behind it, (2) describe the key findings of the Rasch (2007) odor study and the Rudoy (2009) sound study, (3) explain why TMR works specifically during slow-wave sleep and not during REM sleep, and (4) evaluate one real-world application and one limitation of TMR research.

Targeted memory reactivation, or TMR, is a technique where a sensory cue, such as a sound, smell, or image, is paired with new information during learning. Later, while the person is asleep, researchers quietly re-present that same cue. The brain partially reactivates the neural pattern tied to the original memory, strengthening it, without the sleeper consciously waking up or realizing anything happened. TMR builds on a well-established idea called sleep-dependent memory consolidation: newly learned information is not fully 'locked in' the moment you learn it. Instead, the hippocampus replays recent experiences to the neocortex during sleep, gradually transferring memories into more stable, long-term storage. TMR essentially hijacks that replay process and points it at a specific memory the researcher chooses.

Rasch, Buchel, Gais, and Born published the odor-cueing study in Science in 2007. Participants learned card-pair locations while smelling rose odor, then slept either with or without the odor re-presented during slow-wave sleep. The odor-cued group recalled more pairs the next day, and fMRI scans showed the odor triggered activity in the hippocampus during sleep, exactly the structure responsible for spatial memory. Two years later, Rudoy, Voss, Westerberg, and Paller ran a similar experiment using sounds instead of smells. Participants learned the locations of 50 objects, each paired with a matching sound, like a cat's meow for a cat picture. During slow-wave sleep, half of the sounds were replayed at a volume too quiet to cause waking. Objects whose sounds were replayed were recalled significantly more accurately than objects whose sounds were withheld, published in Science in 2009.

TMR cueing only works reliably during slow-wave sleep (also called deep or N3 sleep), the stage where large, synchronized brain waves reflect the hippocampus actively replaying the day's experiences to the cortex. Cueing during REM sleep, the stage associated with dreaming, generally fails to strengthen memories the same way and can even disrupt ongoing consolidation. This is why sleep researchers monitor brain waves with EEG during TMR experiments: cues have to be timed precisely into slow-wave sleep windows, and delivered quietly enough to avoid triggering an arousal or awakening, which would interrupt the very process being studied.