Why Do We Sleep?
Sleep is an activity that occupies a whopping
one-third of our time on the planet. That
is why it is so extraordinary to contemplate
that until about 10 years ago, absolutely no
one had any idea why we needed to sleep.
Even now, while there still is not complete
agreement about sleep’s role in the human
experience, there is growing evidence that the
reason we need to sleep is not just so that the
body can physically recuperate from the day’s
activity, but so that the brain can learn.
One of the first things you have to overcome
if you ever get a chance to listen
in on a living brain while it is slumbering
is disbelief. The reason is that the brain does not appear to be
asleep at all. Instead, the organ is almost unbelievably active during
“rest,” with legions of neurons crackling electrical commands
to each other in constantly shifting patterns, actually displaying
greater rhythmical activity during sleep than when the brain is
The reason for this activity has to do with the tug-of-war
nature of the sleep cycle itself. Your sleep/wake tendencies can be
thought of as the byproduct of a constant battle between two
opposing drives in your body. One of these drives, the Circadian
Arousal System, is trying to keep you awake and alert all the time.
The other, the Homeostatic Sleep Drive, is trying to make you
drowsy so that you stay unconscious all the time. These two
drives are locked in a daily warfare with a cycle of victory and
surrender so predictable you can
actually graph it.
An interesting consequence
of this sleep/wake conflict occurs
in the early afternoon and is
sometimes called the “nap zone.”
Usually experienced sometime
between 2 and 4 p.m., it is a
form of extreme, though temporary,
drowsiness. It can be nearly
impossible to get anything
complex done at this time;
if you choose to attempt to
push through (which is what
most of us try), you can spend
much of your afternoon fighting
a gnawing tiredness.
The whole concept of “siesta” — a phenomenon institutionalized
in many other cultures — likely came as an explicit
reaction to the nap zone. Most scientists believe that a long
sleep in the night and a short nap during the midday represents
human sleep behavior in its most naturally genetic, highest quality
Though sleep cycles are burned deep into the genetic fabric of
the human brain, they are subject to regulatory forces, some of
which are also genetic in origin. For instance, approximately one
in 10 humans is what is formally designated in the scientific
literature as an “early chronotype” or “lark.” Larks report being
most alert around noon and feel most productive a few hours
before they eat lunch. They often wake at 6 a.m., get drowsy in the
early evening, and go to bed (or feel like they want to go to bed)
around 9 p.m.
About two in 10 people exist at the other end of the sleep
spectrum. They are called “late chronotypes” or “owls.” In general,
owls report being most alert around 6 p.m., rarely want to go
to bed before 3 a.m., and rarely want to wake before 10 a.m.
The behaviors of larks and owls are very specific and potentially
genetic. The rest of us have chronotypic behavior that falls somewhere
between the patterns of larks and owls.
So what, then, do sleep and its regulatory functions
have to do with learning? Although the
earliest data establishing a link between sleep
and learning focused on the effects of sleep deprivation, the
opposite is now abundantly clear: Healthy sleep can produce a
significant learning boost in certain tasks.
One study stands out in particular. The experiment involved
giving students a series of math problems and providing them with
a method to solve the problems. Unbeknownst to these students,
there was also an easier, “shortcut” way to solve the same problems,
which potentially could be discovered while doing the exercise.
During the initial training, however, students never discovered
the easier method. So the question for researchers was:
“Is there any way to jumpstart, even speed up, their insight?”
The answer is “yes,” but only if you allow them to sleep on it.
When you let 12 hours pass after the initial training, and then
ask the students to do more problems, approximately 20 percent
actually discover the shortcut. But, if in that 12 hours you also
allow eight or so hours of regular sleep, the number of students
discovering the shortcut triples to approximately 60 percent. No
matter how many times the experiment is run, the sleep group
consistently outperforms the non-sleep group to the tune of
approximately three to one.
This is just one example of the phenomenon. Sleep enhancement
has been shown for tasks that involve visual texture discrimination,
motor adaptations testing, motor sequencing, and
other cognitive skills. The type of learning most sensitive to sleep
improvement is a category called procedural learning (not surprisingly,
this involves learning a procedure). The effect can also
be shown in the negative. Simply disrupt the night’s sleep at specific
stages and then retest in the morning. No overnight learning
improvement will be observed.
Even the simple act of taking a nap can show cognitive benefit.
One study conducted by researchers at NASA showed that
a 26-minute nap improved a pilot’s performance by more than
34 percent. Another study showed that a 45-minute nap produced
a similar boost in cognitive performance, a boost that
lasted more than six hours.
These data and others like them represent a consistent and
increasingly well-characterized finding: Sleep is intimately
involved in learning. It is this consistent finding that caused some
researchers, now almost 10 years ago, to ask a deeper question
about rats and mazes, specifically having to do with leaving a
bunch of electrodes deliberately stuck inside some rats’ brains.
Rats, Mazes, and a
It is possible these days to
put electrodes inside a living
brain and listen in on
the neural chatter while it is thinking. Even in a tiny rat’s brain,
it is not unusual to listen to up to 500 different neurons at once.
And just what are they saying? If you listen in while the rat
is acquiring new information, for instance while it navigates
a new maze, a very discrete “maze-specific” pattern of electrical
stimulation begins to emerge. Working something like the old
Morse Code, these neurons begin to crackle in a specifically timed sequence. Once acquired, the rat will always fire off that
pattern when negotiating the maze.
Interestingly, if the electrodes stay in place when the rat goes to
sleep, something very mysterious can be observed. The rat begins to
replay the maze pattern sequence. Always executed in a very specific
stage of sleep, the rat repeats the pattern over and over again during
the night. If the rat is awakened during that stage, the rat has trouble
remembering how to navigate the maze the next day. The rat is
consolidating the day’s learning the night after that learning
occurred. An interruption of that sleep disrupts this learning cycle.
Does something like that happen in humans? You bet it does,
depending upon the type of learning being measured. Humans
appear to replay certain types of daily learning experiences at
night. Some of the best research has been done with spatial
memory tasks, but it has also been shown for other types of
learning as well. Interestingly, the replay of the day’s learning is
done in a highly compressed format, and even appears to be happening
in the same stage of sleep as the rats’.
These findings represent a bombshell of an idea. The data
seem to suggest that some kind of offline processing is occurring
at night. And that begins to address the question posed in the
title of this article: Why do we sleep? Is it possible that the reason
why we need to sleep is simply to shut off the exterior world
for a while, allowing us to divert more attentional resources to
this processing? The answer to that question may be “yes” and
represents the first real clue as to why sleep is so important.
We may need to sleep so that we can learn.
Lessons for the
As is true with any good scientific discourse,
these data have been met with
some controversy. Increasingly, however,
it is becoming clear that one of the functions of sleep involves the
need for our really big brains to review what they have learned
during the day. This isn’t a trivial need and is demonstrable in the
negative. The effects of sleep deprivation are thought to cost U.S.
businesses more than $100 billion a year.
What if schools and businesses took the sleep needs of their
students and employees seriously? Here are three possibilities:
CHRONOTYPES. What if we began to match chronotypes to
educational experiences? Teachers are just as likely to be late chronotypes
as their students. Why not put them together at the time
of the day when their teaching/learning would be maximized?
A similar case could be made in the work world. What if we began
to match chronotypes to work schedules? Given that 20 percent of
the workforce is already at sub-optimal productivity in the current
8–5 model, what if that model were permanently broken up?
NAPS. Given the data about the powerful effects of short
bursts of sleep on human cognition, is it time to take the nap zone
seriously in schools and offices? Do you recall the NASA
research? “What other management strategy will improve people’s
performance 34 percent in just 26 minutes?” exclaims Mark Rosekind,
the NASA scientist who conducted the research.
SLEEPING ON IT. Lastly, if we took the data about sleep and
the ability of a good night’s rest to upgrade the insight rate, schools
and businesses might tackle the most intractable issues at miniretreats.
After being presented with the problem, students or
employees would not start coming to conclusions, or even begin
sharing ideas, before they had had an intervening eight hours of
sleep. Would they experience the same increase in problem-solving
rates demonstrated in a laboratory setting?
Notice that these ideas are posed as questions, not as prescriptions.
This is done deliberately, because I have no idea whether
anything I just mentioned here would work. Beyond the laboratory,
there is very little research that tells us how we might harness
the wild bucking bronco of the human sleep schedule. Thus,
we don’t know if creating a chronotypically conscious learning
environment or workplace would boost test scores or productivity.
We don’t know if creating a school or workspace sensitive to
the nap zone would give a lift to learning and productivity. We
have no idea if, in a real-world setting, letting a group “sleep on it”
would actually do anything more than waste time.
Why don’t we know? Because, as I say so often, brain scientists,
educators, and business professionals haven’t gotten together to do
such research. And that was exactly why The Brain Center for
Applied Learning Research was created at Seattle Pacific University.
In the meantime, we are left with a simple brain rule — that
sleep is important to the learning process — and an opportunity
to find out how that might apply to the real world of education.
—By John Medina
Director of the
Center for Applied
By John lavin
Back to the top
Back to Home