Sunday, December 19, 2010

Sleep Research Part 1: What is a normal sleep?

For some months now I have been looking for and reading research about babies and sleep.  Not folklore repeated by some self-proclaimed 'sleep expert', not what happened to your best friend's great aunt's baby, and not something someone read on the back of a cornflakes packet - actual scientific studies from pediatric and psychology journals.

What I found out I will be writing about here over the next couple of posts.  There is far too much information to fit into one post.  In future posts, I will be looking at controlled crying (including short term and long term effectiveness, risks and benefits), and other non-crying methods of dealing with children's sleep problems including co-sleeping (and some other methods you may not have thought of).

I hope this information gives new parents greater confidence when sorting through the conflicting information about sleep, and helps them to find the method that's right for them.  All information presented here is referenced so you can check it for yourself if you wish (although you may have to attend a university library in order to do so).  The source of my information is presented in brackets, and all the sources are listed at the end of the post.

Understanding Sleep Study Methods

(This is just some background information for those who are interested.  If you want more practical parenting information, skip down to the next section.)

In considering the meaning and reliability of the results of sleep studies, you need to consider several things:
  • reliability of methods used;
  • who was studied - the age of the children, whether they were a group of 'sleep problem' kids or a general sample of the population etc.;
  • how many people were in the study (enough to draw a meaningful conclusion?);
  • whether the study looks at long term as well as short term results
  • whether the study controlled for other factors that might influence sleep - such as cultural practices, discord in the family home, maternal depression etc.
Bearing these factors in mind helps explain why some studies have different or even conflicting results.

Many sleep studies gather data using parents responses to questionnaires or interviews. These methods can be very unreliable methods of gathering sleep data – as they have been shown to mis-estimate actual sleep times by around 2 hours, and actual nocturnal time awake by around an hour.  Far more reliable methods are actual observation observation of the subject's sleep by the researcher, either using video-recording, brain wave monitoring (EEG) or use of an actigraph (Werner et al).

This is an actigraph - it detects movement a bit like a pedometer:

And leads to print outs like this, which allow you to see when someone was asleep and when they were awake based on the amount of movement:

EEG monitoring is where you have wires stuck to your head that detect electrical activity in the brain, like this:

EEGs give more comprehensive information about exactly what kind of sleep you are having, but they typically require you to sleep in a 'sleep lab' whereas actigraphs can be used in your own home.

In addition to direct observation, daily diaries filled in each morning by a parent also provide highly reliable accounts of sleep for younger babies (Werner et al).  For older children (eg. preschoolers), it has been found that even parental diaries significantly underestimate the amount a child is waking, and overestimate the amount a child is sleeping.

In a study of 59 kindergarten-aged children, where sleep was monitored by use of an actigraph on 4-5 consecutive nights, parents reported the children woke on average 0.5 times a night whereas the actigraphs revealed the children woke an average of 2.7 times a night. 29% of the children slept for less than 90% of the time the parents believed they were sleeping, and 41% had significantly fragmented sleep. Nearly all children woke up at least once a night (Tikotzky and Sadeh). These children had successfully learned not to wake their parents at night for the most part, but that did not mean they were getting a good night's sleep.

How we sleep - the basics:

Normal sleep for adults are controlled by circadian rhythms, ultradian rhythms, and homeostasis. In an adult, the circadian rhythm works like an internal 24-hour clock.

Over 24 hour period our bodies cycle through temperature changes, and particular chemicals (melatonin and cortisol) rise and fall.  Our circadian rhythm helps us go to sleep and wake up at regular times, and it is also responsible for us regularly feeling things like a mid-afternoon slump and later 'second wind'.  Ultradian rhythms make us sleep in sleep cycles that are about 90-110 minutes long. Each sleep cycle contains REM (often dreaming) and non-REM sleep.  Homeostasis is the effect that occurs as time passes since you last slept, and pressure builds up in your body until you feel tired and sleepy again (Jenni and Carskadon).

Newborn babies don't have a 24 hour clock:

Babies are not born with circadian rhythms - these develop over the first six months of life.  They are usually well on their way to being established by the end of months (when melatonin and cortisol start appearing in a 24 hour rhythm), but not fully mature and constant until around 6 months (Jenni and Carskadon).

It is believed that the circadian rhythm develops as a result of the baby experiencing night and day, and also variations in social behaviour according to the time of day (Jenni and Carskadon). A very small study of a handful of babies found massive variation in the time it took an individual baby's clock to mature, from 49 days (1.5 months) to 110 days (3.5 months), and the amount of sleep each baby needed daily fluctuated over time (Jenni, Deboer and Achermann).

For parents, this means that it is important to distinguish between night and day to help your baby's circadian rhythm develop.  This is primarily to do with light and dark (Jenni and Carskadon) so make sure bub gets plenty of daylight during the day and that you don't leave bub in bright lights at night (which has been known to happen in some institutional settings, like hospitals). It doesn't mean that you can't put bub in a dark place to nap, so long as awake times are somewhere light.

It also doesn't mean your baby has to be on routine of daily activities.  But do differentiate your daytime and nighttime behaviour.  You don't have to avoid eye-contact or comforting your baby at night, but you want to keep things quiet and mellow.  Babies react to social cues as well as light-dark cues (Jenni and Carskadon).

It also means that most babies will not be able to follow a strict routine during the first few months, and some will not manage it until about six months.  You don't know what kind of baby you have.  It is only by trial and error over time that you will figure out when they are ready.  But don't fall into the trap of thinking that because little Johnny slipped easily into a routine at 2 months that your baby will too, and there's something wrong with him if he doesn't, or that you have to fight him into a routine when he's too young to get the hang of it. A gentle, flexible routine won't hurt, but you may find yourself waking a tired baby and endlessly settling one that's awake without him learning anything.

Different babies need different schedules, adapted to them and adapted over time:

Once your baby has a circadian rhythm reasonably established, you may want to help him or her have a regular daily schedule.  In doing this, it is important to realise that different babies can have very different sleep requirements, and that even for an individual these fluctuate over time.  This means that if you are trying to follow a schedule in a book, you need to adapt it to your baby.  If your baby is getting very cranky and sleepy, they probably need more sleep - on the other hand, if you are fighting your baby to sleep, he or she may simply not be tired.

Why does it vary?  Because homeostasis is different. When an individual does not get enough sleep, this creates what is known as 'homeostatic sleep pressure' to encourage the individual to go back to sleep. Homeostatic pressure is very strong in babies, which is why they can't stay awake for very long at a time, but it varies from baby to baby (Jenni and Carskadon). Studies suggest that the amount of sleep a child needs varies dramatically.  At age 6 months, both 10.5 hours and 18 hours have been found as sleep patterns for individual infants (Jenni and Carskadon).

In one recent study of Swiss children aged 12 months, a normal sleep was found to be anywhere between 11.4 hours and 16.5 hours within a 24 hour period (Jenni, 2007). This study followed the same children over several years and found that shorter sleepers tended to stay shorter sleepers, and longer sleepers tended to stay longer sleepers, indicating strongly that individual children need different amounts of sleep.

There is now also evidence that the chemicals in the brain that trigger sleep timing and duration vary from person to person (Tafti, Aeschbach). It also seems normal for an individual to need less sleep as they get older, but for this not to reduce steadily over time but instead for children to have 'progression' periods, but then 'regression' periods where they need more sleep again (Jenni, 2007).

Sleeping in 45-50 minute cycles:

Ultradian rhythms are noticeable more quickly, within a couple of weeks of birth, but baby sleep cycles are 50 minutes.  They do not develop longer, adult-like sleep cycles until they are about 6 years of age (Jenni and Carskadon).  Much has already been written on this and this information is commonly found in popular parenting books on sleep, so I won't repeat it here.  But it is the reason your baby often wakes after about 45 minutes, or multiples of 45 minutes (eg. 1.5 hours), as at the end of each sleep cycle they come into a lighter sleep.

Sleeping through the night:

Most, but not all, infants have the capacity to sleep an 8 hour stretch by 9 months and many achieve it as early as 6 months (Jenni and Carskadon).  For those that have this capacity, some may not actually sleep 6 hours in a row unless they are in certain circumstances - which is a Western-style solitary sleep arrangement and sometimes bottle feeding.  In one large Canadian study, infants were less likely to sleep for 6 hours in a row at 6 months if they were breastfed (Touchette et al).  Co-sleeping babies are also less likely to sleep through at 6 months - a study of 6 month old Swedish infants found that co-sleeping was associated with waking more than 3 times a night, although it was not clear at this age whether co-sleeping caused the waking, or the waking caused parents to try co-sleeping (Möllborg et al).  On the other hand, co-sleepers are less likely to be bothered by the wakings (Keller and Goldberg), something I will discuss in a future post.  At 29 months, about 10% of children are not sleeping for 6 consecutive hours (Touchette et al), although this figure is probably higher because the study was conducted by questionnaire, and these tend to underestimate night wakings (see discussion of sleep study methods above).

Failure to sleep through the night by 6 months of age does not necessarily indicate a persistent sleep problem (something I will discuss in a future post).

Ideas about a 'normal sleep' varies dramatically between cultures:

Ideas about 'normal' sleep have been developed primarily by studying white, middle-class Americans. But studies of other cultural groups are showing that a 'normal' sleep is a surprisingly variable concept. For example:
  • In a study in Southern Mississippi it was found that 4.9% of white children aged 8 napped, but 39.1% of black children, although both racial groups got the same total amount of sleep in a 24 hour period (Crosby).

  • Italian children typically have no bedtime schedules or rituals. A study of Italian children between 2 and 4 years old found that they sleep less than American children, they stay up later, and usually fall asleep without adult assistance (Jenni and O'Connor).

  • Dutch 3 month old children were found to go to bed earlier than American children and sleep an average of 15 hours a day, as opposed to the American 3 month olds, which sleep an average of 13 hours a day, although there are questions around how much of the time in bed children are actually asleep (Jenni and O'Connor).

  • In Japan, school-aged children have a nap after dinner and then are woken so they can do their homework for a few hours after their parents have gone to bed (Jenni and O'Connor).

  • In a Swiss study comparing sleep duration, bedtimes, and settling difficulties found that children in the 1970s were expected to go to bed earlier and had greater bed-time resistance, compared to children in the 1990s (Jenni and O'Connor).

  • In traditional Balinese society, both infants and adults sleep and wake throughout the day and night when they feel like it, and infants are expected to fall asleep anywhere with any degree of audio and visual stimulation (Jenni and O'Connor).

  • In various countries including Italy, Mexico, China, and Japan, daytime napping is the norm for adults as well as children, and there have been times in their history where it was not only normal but institutionalised (eg. Shops and offices shut down at expected nap times). (Jenni and O'Connor).

  • In the middle ages in Europe, the normal sleep pattern was to have a 'first sleep', a wakeful meditative period in the middle of the night, and then a 'second sleep'. It was believed to promote good digestion and religious behaviour (Jenni and O'Connor).

  • In 19th century America, the recommended sleep for three year olds was 12 hours a day, and 8-9 hours for 7 year olds (Jenni and O'Connor), 'sleep problems' did not emerge as a complaint until around the 1920s when health experts began to state that American children had remarkably little sleep, and parents began to be advised to get their children to bed earlier. This was when sleep routines and methods of getting a child to sleep emerged (Jenni and O'Connor).
  • Mothers of the Ache tribe in South America mothers sleep sitting up with their babies in their laps (Small).

Coming Up Next...

Stay tuned for the next post in this series where I will summarise the latest research on the effectiveness and effects of controlled crying.


Aeschbach et al, 'A longer biological night in long sleepers than in short sleepers', (2003) Journal of Clinical Endocrinology and Metabolism, Vol 88, p26.

Crosby et al, 'Racial Differences in Reported Napping and Nocturnal Sleep in 2- to 8-Year-Old Children' (2005) Pediatrics Vol 115, p225.

Jenni and Carskadon, 'Chapter 1: Normal Human Sleep at Different Ages: Infants to Adolescents'

Jenni, Deboer and Achermann, 'Development of the 24-h rest-activity pattern in human infants' (2005).

Jenni et al, 'Sleep Duration From Ages 1 to 10 Years: Variability and Stability in Comparison With Growth' (2007) Pediatrics Vol 120, pe769.

Jenni and O'Connor, 'Children's Sleep: An Interplay Between Culture and Biology' (2005) Pediatrics Vol 115(1), p204.

Keller and Goldberg, 'Co-Sleeping: Help or Hindrance for Young Children's Independence?' (2004) Infant Child Development Vol 13, p369.

Möllborg et al, 'Bed-sharing among six-month-old infants in western Sweden' (2010) Acta Paediatrica

Small, Our Babies, Ourselves: How Biology and Culture Shape the Way We Parent (1999) First Anchor Books.

Tafti, Maret, and Davilliers, 'Genes for normal sleep and sleep disorders' (2005) Ann Med Vol 37, p508.

Tikotzky and Sadeh, 'Sleep Patterns and Sleep Disruptions in Kindergarten Children' (2001) Journal of Clinical Child & Adolescent Psychology Vol 30(4), p581.

Touchette et al, 'Factors Associated With Fragmented Sleep at Night Across Early Childhood' (2005) Archives of Pediatric Adolescent Medicine, Vol 159, p242.

Werner et al, 'Agreement Rates Between Actigraphy, Diary, and Questionnaire for Children's Sleep Patterns' (2008) Archives of Pediatric Adolescent Medicine Vol 162(4), p350.


  1. Wow! I'm looking forward to the controlled crying post as I've done a lot of reading about sleep methods when I was trying to resolve Eamonn's problems and I got so confused and upset with the contradictory advice, all from reputable sources.

  2. Caroline, I really want to thank you so much for this blog, and tackling the sleep issue in particular. I agree with you that every family has to decide what works for them and if sleep training works, then awesome. But for me it didn't feel right and I was really struggling with whether I was actually doing my daughter a disservice by not 'letting' her figure it out herself. Looking at the studies from another perspective was really, really valuable to me.

    Thanks again!

  3. This is great! I read this as I sit here rocking and nursing my 1 year old to sleep! Thank you for posting this.

  4. You suggest sleep cycles are 45-50 minutes but elsewhere i read 40 mins. My friends have been amazed when I predict, to the minute, what time our ruggie will wake up from his daytime sleep - based on 40 minute cycles. Presumably it varies from one child to the next.