Learning while you sleep

By Brooke Napier

A paper came out this week in Nature Neuroscience with the conclusion: Humans can learn new information during sleep. 

This is not the plot of a Saved By the Bell episode where Zach Morris listens to books on tape before his big test the next day and remembers everything; scoring an “A” and the hottest babe in class – but, these new data show that like Pavlov’s dog, we too can learn without being conscious.

In fact, the experiments used to show this phenomenon are a little too close to Pavlov’s methods for comfort.

Since sleep is defined as “a loss of consciousness and reduced responsiveness to external stimuli”, these researchers had to be sly about how they went about testing learning.

Anat Arzi, et al. knew that the inherent nature of the olfactory response provided an easy read out for learning while unconscious - or sleeping, eureka!

How exactly did they measure learning using odors while people slept?

Simply, bad odors drive weak sniffs and pleasant odors drive stronger sniffs. So they paired pleasant and unpleasant odors with different tones during sleep. Next, they measured the sniff response to the tones alone during the same nights’ sleep and while they were next awake.

In the paper they call this, “partial-reinforcement trace conditioning”.

This sounds trivial, but there is never an experiment on humans that is simple.

Their first hurdle was to make sure that the odors they chose did not wake any of the subjects. A sleep technician used standard polysomnography (diagnostic tool used to study sleep by measuring biophysical changes that occur during sleep) standards for “awake” and “asleep”. Additionally, they characterized the changes in the brain when subjected to the odors using an electroencephalogram (EEG).

I'd like to think my dogs are this smart.

Next, they needed to know that their “pleasant” odor was actually pleasant to everyone (because I know some people who like the smell of Windex, and that’s just wrong) and that it is processed as pleasant while the subject is sleeping. You’ll be happy to know that they chose deodorant and shampoo as pleasant odors and that these were perceived so by the subjects while awake (by verbally agreeing to pleasantness) and while asleep (a measured increase in sniffing while asleep after the pleasant odor).

Ok, now they needed to pair the odor and tone. Since there was a notable increase in concentration of sniffing after exposure to the pleasant odor while sleeping, this was an easy read-out for learning while sleeping.

The studied followed as listed below:

1. Pair tones with odors during sleep.

2. Measure sniffs following tones alone on the same night or during their next arousal, or wake, state.

3. Analyze data. 

3 1/2 . Repeat experiments.

4. Get Nature paper.

Hey, look at this sweet data:

The sniff response revealed learning during sleep. (a) The averaged normalized sniff trace and (inset) sniff volume during sleep following a pleasant (blue) or unpleasant (brown) odor (n = 28). (b) The averaged normalized sniff volume during sleep following a tone (alone) previously paired during sleep with a pleasant odor (blue outline) and a tone (alone) previously paired during sleep with an unpleasant odor (brown outline) (n = 20). (c) The average learning curve across five continuous repetitions of tones (alone) previously paired with a pleasant odor (blue outline), or five continuous repetitions of tones (alone) previously paired with an unpleasant (brown outline) odor (total = 10 trials). (d) The averaged normalized sniff volume awake following a tone (alone) previously paired during sleep with a pleasant odor (blue outline) and a tone (alone) previously paired during sleep with an unpleasant odor (brown outline) (n = 6). (e) The averaged wake nasal inhalation volume following tones (400 and 12,00 Hz) in a control group that was not conditioned during sleep (n = 10). Statistical analysis was conducted using two-tailed t test (a,b,d,e) and one-tailed t test (c).

They also showed that conditioning (or learning) was possible during rapid eye movement (REM) sleep – but was lost when the subject was woken up. They mention that the stronger transfer from non-REM learning to wake may be linked to the previously observed notion that there is increased functional connectivity between olfactory and neocortical areas during slow-wave activity. Interestingly, they then mention that stronger transfer from non-REM may be consistent with previous data showing that humans have rapid loss of memory for REM-related memories (called dream amnesia).

One final interesting tid-bit, they asked the subjects if they remember any of the smells or tones they heard/smelled while sleeping and they have no recollection.

Freaky-deaky.  

In all good Nature paper ediquette they end with a very provocative statement, “[Our study] implies that, beyond the general health advantages associated with good sleep, humans may be able to utilize toward learning new information a state in which they spend about a third of their lives.”

Arzi A, Shedlesky L, Ben-Shaul M, Nasser K, Oksenberg A, Hairston IS, & Sobel N (2012). Humans can learn new information during sleep. Nature neuroscience, 15 (10), 1460-5 PMID: 22922782

Comments

[Sarah]

I like Step 4 and also the smell of Windex.