Learning new information during sleep.

Arzi et al. do an ingenious experiment to show that we can do associative learning during our sleep. We can associate a sound with a pleasant or unpleasant odor and react, both while still asleep and after waking, with a deeper or shallower breath. This does not, however, represent the kind of 'sleep learning' long sought by students who unsuccessfully try to remember scientific or literary facts needed for an exam by playing a tape softly during sleep. Here is the abstract:

During sleep, humans can strengthen previously acquired memories, but whether they can acquire entirely new information remains unknown. The nonverbal nature of the olfactory sniff response, in which pleasant odors drive stronger sniffs and unpleasant odors drive weaker sniffs, allowed us to test learning in humans during sleep. Using partial-reinforcement trace conditioning, we paired pleasant and unpleasant odors with different tones during sleep and then measured the sniff response to tones alone during the same nights' sleep and during ensuing wake. We found that sleeping subjects learned novel associations between tones and odors such that they then sniffed in response to tones alone. Moreover, these newly learned tone-induced sniffs differed according to the odor pleasantness that was previously associated with the tone during sleep. This acquired behavior persisted throughout the night and into ensuing wake, without later awareness of the learning process. Thus, humans learned new information during sleep.