Why We Sleep

participants had learned the sequence in the morning and were tested later that
evening  after  remaining  awake  across  the  day.  The  other  half  of  the  subjects
learned the sequence in the evening and we retested them the next morning after
a similar twelve-hour delay, but one that contained a full eight-hour night of sleep.
Those who remained awake across the day showed no evidence of a significant
improvement  in  performance.  However,  fitting  with  the  pianist’s  original
description, those who were tested after the very same time delay of twelve hours,
but  that  spanned  a  night  of  sleep,  showed  a  striking  20  percent  jump  in
performance speed and a near 35 percent improvement in accuracy. Importantly,
those participants who learned the motor skill in the morning—and who showed
no  improvement  that  evening—did  go  on  to  show  an  identical  bump  up  in
performance when retested after a further twelve hours, now after they, too, had
had a full night’s sleep.
In  other  words,  your  brain  will  continue  to  improve  skill  memories  in  the
absence  of  any  further  practice.  It  is  really  quite  magical.  Yet,  that  delayed,
“offline”  learning  occurs  exclusively  across  a  period  of  sleep,  and  not  across
equivalent  time  periods  spent  awake,  regardless  of  whether  the  time  awake  or
time asleep comes first. Practice does not make perfect. It is practice, followed by
a night of sleep, that leads to perfection. We went on to show that these memory-
boosting benefits occur no matter whether you learn a short or a very long motor
sequence  (e.g.,  4-3-1-2  versus  4-2-3-4-2-3-1-4-3-4-1-4),  or  when  using  one  hand
(unimanual) or both (bimanual, similar to a pianist).

---

Analyzing  the  individual  elements  of  the  motor  sequence,  such  as  4-1-3-2-4,
allowed me to discover how, precisely, sleep was perfecting skill. Even after a long
period of initial training, participants would consistently struggle with particular
transitions  within  the  sequence.  These  problem  points  stuck  out  like  a  sore
thumb when I looked at the speed of the keystrokes. There would be a far longer
pause,  or  consistent  error,  at  specific  transitions.  For  example,  rather  than
seamlessly  typing  4-1-3-2-4,  4-1-3-2-4,  a  participant  would  instead  type:  4-1-3
[pause] 2-4, 4-1-3 [pause] 2-4. They were chunking the motor routine into pieces,
as if attempting the sequences all in one go was just too much. Different people
had  different  pause  problems  at  different  points  in  the  routine,  but  almost  all
people had one or two of these difficulties. I assessed so many participants that I
could actually tell where their unique difficulties in the motor routine were just by
listening to their typing during training.
When  I  tested  participants  after  a  night  of  sleep,  however,  my  ears  heard
something very different. I knew what was happening even before I analyzed the
data:  mastery.  Their  typing,  post-sleep,  was  now  fluid  and  unbroken.  Gone  was
the  staccato  performance,  replaced  by  seamless  automaticity,  which  is  the
ultimate  goal  of  motor  learning:  4-1-3-2-4,  4-1-3-2-4,  4-1-3-2-4,  rapid  and  nearly
perfect. Sleep had systematically identified where the difficult transitions were in
the motor memory and smoothed them out. This finding rekindled the words of
the pianist I’d met: “but when I wake up the next morning and sit back down at
the piano, I can just play, perfectly.”
I went on to test participants inside a brain scanner after they had slept, and
could  see  how  this  delightful  skill  benefit  had  been  achieved.  Sleep  had  again
transferred the memories, but the results were different from that for textbook-
like memory. Rather than a transfer from short- to long-term memory required for
saving  facts,  the  motor  memories  had  been  shifted  over  to  brain  circuits  that
operate below the level of consciousness. As a result, those skill actions were now
instinctual  habits.  They  flowed  out  of  the  body  with  ease,  rather  than  feeling
effortful and deliberate. Which is to say that sleep helped the brain automate the
movement routines, making them second nature—effortless—precisely the goal
of many an Olympic coach when perfecting the skills of their elite athletes.
My final discovery, in what spanned almost a decade of research, identified the
type  of  sleep  responsible  for  the  overnight  motor-skill  enhancement,  carrying
with  it  societal  and  medical  lessons.  The  increases  in  speed  and  accuracy,
underpinned  by  efficient  automaticity,  were  directly  related  to  the  amount  of
stage 2 NREM, especially in the last two hours of an eight-hour night of sleep (e.g.,

---

from  five  to  seven  a.m.,  should  you  have  fallen  asleep  at  eleven  p.m.).  Indeed,  it
was the number of those wonderful sleep spindles in the last two hours of the late
morning—the time of night with the richest spindle bursts of brainwave activity
—that were linked with the offline memory boost.
More  striking  was  the  fact  that  the  increase  of  these  spindles  after  learning
was detected only in regions of the scalp that sit above the motor cortex (just in
front  of  the  crown  of  your  head),  and  not  in  other  areas.  The  greater  the  local
increase  in  sleep  spindles  over  the  part  of  the  brain  we  had  forced  to  learn  the
motor skill exhaustively, the better the performance upon awakening. Many other
groups  have  found  a  similar  “local-sleep”-and-learning  effect.  When  it  comes  to
motor-skill memories, the brainwaves of sleep were acting like a good masseuse
—you still get a full body massage, but they will place special focus on areas of the
body that need the most help. In the same way, sleep spindles bathe all parts of
the  brain,  but  a  disproportionate  emphasis  will  be  placed  on  those  parts  of  the
brain that have been worked hardest with learning during the day.
Perhaps  more  relevant  to  the  modern  world  is  the  time-of-night  effect  we
discovered. Those last two hours of sleep are precisely the window that many of
us feel it is okay to cut short to get a jump start on the day. As a result, we miss
out  on  this  feast  of  late-morning  sleep  spindles.  It  also  brings  to  mind  the
prototypical Olympic coach who stoically has her athletes practicing late into the
day,  only  to  have  them  wake  in  the  early  hours  of  the  morning  and  return  to
practice.  In  doing  so,  coaches  may  be  innocently  but  effectively  denying  an
important phase of motor memory development within the brain—one that fine-
tunes  skilled  athletic  performance.  When  you  consider  that  very  small
performance  differences  often  separate  winning  a  gold  medal  from  a  last-place
finish  in  professional  athletics,  then  any  competitive  advantage  you  can  gain,
such as that naturally offered by sleep, can help determine whether or not you will
hear your national anthem echo around the stadium. Not without putting too fine
a point on it, if you don’t snooze, you lose.
The 100-meter sprint superstar Usain Bolt has, on many occasions, taken naps
in the hours before breaking the world record, and before Olympic finals in which
he  won  gold.  Our  own  studies  support  his  wisdom:  daytime  naps  that  contain
sufficient  numbers  of  sleep  spindles  also  offer  significant  motor  skill  memory
improvement, together with a restoring benefit on perceived energy and reduced
muscle fatigue.
In  the  years  since  our  discovery,  numerous  studies  have  shown  that  sleep
improves the motor skills of junior, amateur, and elite athletes across sports as

---

diverse  as  tennis,  basketball,  football,  soccer,  and  rowing.  So  much  so  that,  in
2015,  the  International  Olympic  Committee  published  a  consensus  statement
highlighting  the  critical  importance  of,  and  essential  need  for,  sleep  in  athletic
development across all sports for men and women.
X
Professional sports teams are taking note, and for good reason. I have recently
given presentations to a number of national basketball and football teams in the
United  States,  and  for  the  latter,  the  United  Kingdom.  Standing  in  front  of  the
manager,  staff,  and  players,  I  tell  them  about  one  of  the  most  sophisticated,
potent,  and  powerful—not  to  mention  legal—performance  enhancers  that  has
real game-winning potential: sleep.
I back up these claims with examples from the more than 750 scientific studies
that  have  investigated  the  relationship  between  sleep  and  human  performance,
many  of  which  have  studied  professional  and  elite  athletes  specifically.  Obtain
anything less than eight hours of sleep a night, and especially less than six hours a
night, and the following happens: time to physical exhaustion drops by 10 to 30
percent,  and  aerobic  output  is  significantly  reduced.  Similar  impairments  are
observed  in  limb  extension  force  and  vertical  jump  height,  together  with
decreases  in  peak  and  sustained  muscle  strength.  Add  to  this  marked
impairments  in  cardiovascular,  metabolic,  and  respiratory  capabilities  that
hamper  an  underslept  body,  including  faster  rates  of  lactic  acid  buildup,
reductions  in  blood  oxygen  saturation,  and  converse  increases  in  blood  carbon
dioxide, due in part to a reduction in the amount of air that the lungs can expire.
Even  the  ability  of  the  body  to  cool  itself  during  physical  exertion  through
sweating—a critical part of peak performance—is impaired by sleep loss.
And then there is injury risk. It is the greatest fear of all competitive athletes
and their coaches. Concern also comes from the general managers of professional
teams, who consider their players as prized financial investments. In the context
of injury, there is no better risk-mitigating insurance policy for these investments
than sleep. Described in a research study of competitive young athletes in 2014,
XI
you can see that a chronic lack of sleep across the season predicted a massively
higher risk of injury (figure 10).