A Blog by Ed Yong

Genetic snooze button shows that broken sleep impairs memories

Many mental disorders can disrupt the sweet embrace of a long, continuous sleep, including alcoholism, depression, Alzheimer’s and parenthood. And that’s bad news. We know that a good night’s sleep helps to solidify our memories of the previous day’s experiences. And according to a new study, we need a certain amount of continuous sleep for those benefits to kick in.

From an evolutionary point of view, it seems strange that we sleep for hours on end. Rather than leaving ourselves unresponsive and vulnerable for large chunks of time, why not simply sleep over several shorter fragments?

This is not an easy question to answer. Until recently, it has been all but impossible to break up the continuity of sleep without also affecting its quality, or stressing out the animals in question.  But Luis de Lecea from Stanford University has found a way. He has engineered mice with in-built silent alarm clocks. These animals can be woken up at will with a pulse of light delivered directly to their brains.

De Lecea focused on a group of neurons that exude a hormone called hypocretin (Hcrt), which helps to rouse a sleeping animal. In 2007, he loaded the Hcrt neurons of mice with a light-sensitive protein called Chr2. De Lecea used an optic fibre to shine blue light onto Chr2, causing the Hcrt neurons to fire, and the sleeping mice to wake up. This is just one of the many applications for optogenetics, the exciting new field where scientists can control the behaviour of animals and cells using bursts of light.

Asya Rolls, who is part of de Lecea’s team, used this technique to gently wake her mice once a minute, without actually disrupting the overall quality of their sleep. The rodents’ brain waves revealed that they did indeed start to wake up, before falling asleep again in less than two seconds – a so-called “microarousal”.

Their sleep became fragmented, but they slept for the same amount as their undisturbed peers. Their brainwaves showed that they were sleeping just as deeply, and they spent the same amount of time in the two major phases of sleep – REM and non-REM.  The levels of stress hormones in their blood didn’t go up, and they showed no signs of stress or groggy behaviour when they finally woke up. Rolls had essentially invented the world’s least annoying snooze button. For all intents and purposes, the mice had had a good night’s sleep, albeit one that was broken up into several small chunks.

But that made a lot of difference. Before the mice had their naps, Rolls placed them in an arena with two similar objects, which they explored equally.  Half an hour later, they were sent to sleep and half of them were briefly woken up with blue light every minute. After four hours of this, they were placed back into the arena, where they saw one of the two objects from before, along with a completely new one.

Mice are drawn to new things and the undisturbed individuals spent around 70% of their time looking at the new object.  But those whose sleep had been broken up spent just as much time with each object; they didn’t remember which one they had seen before. However, if Rolls woke the mice up every two minutes, they had no problems remembering which of the two objects was familiar.

This suggests that mice need to sleep for between one and two minutes straight, if they are to consolidate newly formed memories. It’s hard to say what the equivalent amount of time is for humans. Mice tend to have short sleep cycles of around 9 minutes, while the equivalent for a human is 90 minutes. De Lecea says, “I prefer not to speculate about the minimal sleep unit in humans, but patients with sleep apnea with severe sleep fragmentation (more than 30 waking moments per hour) tend to have memory deficits.”

This link between sleep and memory is now becoming very clear. Sleep allows us to prune away any new connections between our brain cells, leaving behind the most useful ones. It also gives us time to replay the events of the day before, fitting new information in among the old, and solidifying our latest experiences into long-term memories. By briefly disturbing her mice, Rolls thinks that she interfered with these replays, preventing the animals’ brains from consolidating what they learned.

Slowly, science is waking up to the purpose of sleep, and new genetic techniques are speeding up the pace of such discoveries. Just a month ago, I wrote about the work of Jeffrey Donlea, who has genetically engineered flies that he can send to sleep by raising the temperature. He found that the insects developed stronger memories after they’d had a nap. Meanwhile, a Japanese team have developed mice that they can similarly send to sleep on demand, by using bursts of light to switch off their hypocretin neurons. And Rolls has just showed how important it is to get unbroken bouts of sleep, regardless of the total amount.

Reference: Rolls, Colas, Adamantidis, Carter, Lanre-Amos, Heller & de Lecea. 2011. Optogenetic disruption of sleep continuity impairs memory consolidation. PNAS http://dx.doi.org/10.1073/pnas.1015633108

More on optogenetics:

More on sleep:

8 thoughts on “Genetic snooze button shows that broken sleep impairs memories

  1. Ed wrote:

    This suggests that mice need to sleep for between one and two minutes straight, if they are to consolidate newly formed memories. It’s hard to say what the equivalent amount of time is for humans. Mice tend to have short sleep cycles of around 9 minutes, while the equivalent for a human is 90 minutes.

    I suspect the experiment to test the reactiveness and retentiveness of humans on the scale done for mice borders on what might be considered cruel and inhuman … I’m sure it’s lots of fun, much like Venkman’s and Spengler’s experiments in the Ghostbusters movies.

  2. Sleep is really fascinating. Many people take it for granted that we are supposed to sleep all through the night and stay up all day but it sounds like that was more a product of the industrial revolution. I am very interested in studies that probe into discovering the most effective sleep cycles, not that they would be the same for everyone!
    I mean, I sleep well when I don’t stay up all night doing any of various things, but I’m sure I could be doing it better!

  3. I have absolutely no restrictions on my time, no where I must be, no one to see…etc., and I also keep a written record of my activivities, diet, etc. I sleep when I feel like it, I wake up when I feel like it.

    The majority of my sleeps are about 90 minutes with an average of a bit less than once a day at four to six hours, the usual total for 24 hours being about 7-8 hours. About once every two weeks, I may not sleep any for up to 36 hours and then drop off for 8-12 hours straight.

    I am now retired, but had been self-employed in business for 40 years where I made my own rules and own schedule. Work for the week was usually done in 2-3 days each week by working ’til tired, then sleeping for 2-4 hours, then working ’til tired again. This allowed me to have 4-5 days of free time each week, a real benefit of a loose sleep schedule triggered by actual need or tiredness, and not by routine, habit, or the necessity of accommodating oneself to the schedules of others.

    Additionally another benefit accrued through my atypical behaviour. During those years, I was able at times to work for three days straight without a break for anything but food (this work was generally heavy physical labor necessitated by the exigencies of business, paperwork was always postponed until it was convenient). When the work session was ended, I would characteristically sleep in an apparent comatose condition for 24-36 hours.

    In my opinion, this is probably close to the natural sleep pattern of ancient paleolithic humans who might have been “at the edge” both in procurement of foodstuffs and in avoidance of predators. Sleep would have been a necessity (for brain maintenance) but also something that could be done at almost any time especially with a full belly and a safe place to nod off.

    Many other animals also use the same type of sleep patterns showing perhaps that this is necessary to survival in and to thriving in hostile environments.

  4. @7. I enjoyed your comments. Perhaps someone should study you?! I have also found that “napping” for 20 – 30 minutes at a time reduces my need to sleep at night. I used to sleep about 9 hours straight and never felt energetic during the long day. Now I have changed my working hours to 4-midnight with a 30 min nap in-between and I feel much better! Also, I only sleep about 7 – 7.5 hours during the night and wake without an alarm clock. I am obviously not retired yet, but it will be interesting to see how my sleep pattern/requirements change as I move from 45 onward. Thanks for your post.

Leave a Reply

Your email address will not be published. Required fields are marked *