The short answer
Mostly because of active sleep, the newborn version of REM, when a baby's face and limbs can still move. Researchers have recorded sleep smiles in premature babies and fetuses, long before a baby has seen a face, and no one can say whether she is dreaming. The awake social smile usually appears by the end of the second month.
"Why do babies smile in their sleep?" is not an urgent question. Nothing is wrong. It is just that a sleeping newborn's sudden lopsided grin is one of the strangest and loveliest things you will ever see, and when you are that tired you want it to mean something.
The answers online are mostly confident and mostly unsourced. Some say she is dreaming about you. Some say it is gas. Some say it is the first sign of a social personality. Reading what researchers have actually published, for this page, the honest picture turned out to be more interesting than any of those, and also more modest.
Here is the short version. Newborn sleep smiles happen mostly during active sleep, the infant version of REM. Researchers have recorded them in premature babies and even in fetuses on ultrasound, long before a baby has seen a single face. Nobody can tell you whether a baby is dreaming, and the research does not claim to. The awake social smile, the one aimed at you, is a separate milestone that the AAP says usually appears by the end of the second month.
This page walks through the studies behind each of those statements, with links so you can read them yourself. It is not medical advice, and if anything about your baby's sleep or development worries you, your pediatrician is the right person to ask. For the broader picture of newborn sleep, the homepage is a good place to start.
Need the clock times for your baby?
The calculator builds the whole day from one wake-up.
Active sleep is where most of it happens
To understand the smile, it helps to know what a newborn's sleep looks like from the inside. The Sleep Foundation describes newborn sleep as cycling through three states: quiet sleep, indeterminate sleep and active sleep. Active sleep is the infant counterpart of adult REM sleep, the stage in which adults typically dream.
There is one big difference. In adult REM sleep, most muscles are temporarily paralysed. In infant active sleep they are not. So during active sleep a baby may twitch, make sucking motions, move her arms and legs, frown, and smile. The Sleep Foundation's baby-sleep overview adds that newborns spend around half of their sleep in REM, which may explain why sleep smiles are common in the first few months.
This connection between smiling and REM states is not new. Emde and Koenig published "Neonatal smiling and rapid eye movement states" in 1969 (Journal of the American Academy of Child Psychiatry, PMID 4306576), followed later that year by a sleep-cycle study that also looked at frowning (PMID 4317623). In 1971 Emde and colleagues extended the work to premature infants (PMID 4333441). The titles alone tell you where researchers were looking: at the sleep stage, not at the baby's day.
If your baby also grunts, squeaks and even cries without waking, that is the same active sleep at work. Our page on baby crying in sleep covers the noisy side of it, and the newborn baby sleep schedule explains how much of the day this kind of sleep takes up.
Smiles that start before birth
The most surprising finding in this literature comes from a group of Japanese researchers led by K. Kawakami at the University of the Sacred Heart in Tokyo, who have published a series of studies recording what they call spontaneous smiles: smiles that arise without any outside trigger, mostly during sleep.
In a 2006 preliminary study (Early Human Development, PMID 16185829), they observed spontaneous smiles in 10 newborns and followed six infants over time. In the newborns, one-sided smiles were more common than two-sided ones, but by two months almost all spontaneous smiles were bilateral. That lopsided sleep grin many parents notice in the first weeks is exactly what they described.
In 2008 the same group observed 22 premature babies in a neonatal intensive care unit for one hour each and recorded 95 spontaneous smiles (PMID 18468691). The younger and smaller babies smiled more often and for longer than the older and larger ones. The youngest was 200 days from conception and weighed 511 grams, and she smiled too.
Then, in 2012, F. Kawakami and Yanaihara used four-dimensional ultrasound to observe 568 fetuses, and selected 31 who smiled, between 156 and 214 days of conceptional age (Infant Behavior and Development, PMID 22721744). Their conclusion was blunt: since fetuses smile a lot, fetal smiles "must not result from social effects."
Put together, these studies change the question. A baby who smiles before she has ever seen a face is not smiling at anyone. Whatever produces the sleep smile, it is built in, not learned.
Why would a baby smile with nobody to smile at? The Sleep Foundation reports that researchers believe spontaneous smiles during sleep may help develop the muscles used to smile, and lay the groundwork for smiling and laughter as the child grows. It also mentions studies suggesting these active-sleep smiles are generated in the cerebral cortex, the outer part of the brain. Both are presented as research findings and hypotheses rather than settled facts, and are best read that way. They are, though, a more satisfying answer to why do babies smile in their sleep than either "gas" or "dreaming of you", because they fit what the fetal studies show.
The same research group noticed one more thing parents sometimes hear: a sleep laugh. In the 2006 study, spontaneous laughs, which the authors define as smiles accompanied by vocal sounds, were always two-sided, unlike the early smiles, and the authors suggested that spontaneous smiles and spontaneous laughs might be different behaviours from the beginning. In the 2007 single-baby case study, which logged more than 329 hours of observation over six months, they recorded 565 spontaneous smiles and 15 spontaneous laughs. So if your baby occasionally makes a small giggling sound in her sleep, researchers have recorded that too; it is simply much rarer than the smile.
So is she dreaming?
This is the question most parents are really asking, and the honest answer is that nobody knows.
The Sleep Foundation's article on this question puts the uncertainty plainly. It notes that adults smile in response to positive dream imagery, but that it is difficult to say whether the same is true for babies. It also says that older research described infant sleep smiles as reflex-like, and that until more research is done it is difficult to say for sure why infants smile during sleep. Its key takeaways allow that as babies grow, smiles may also occur during dreams, while adding that researchers cannot say for certain what babies are dreaming about.
Messinger and Fogel, in a 2007 review of how social smiling develops (Advances in Child Development and Behavior, PMID 17682330), state that "neonatal smiling has no clear emotional content." Their review also notes that infant smiles emerge even without visual feedback. That fits the fetal and premature findings: the smile exists before the experiences that would give it meaning.
In 1972, Emde and Harmon wrote about "endogenous and exogenous smiling systems" in early infancy (PMID 4337500). The vocabulary is useful even if you never read the paper. Endogenous means coming from inside: driven by the baby's own internal state. Exogenous means coming from outside: a response to something in the world, like your face. Sleep smiles in newborns belong to the first group.
None of this rules out that something pleasant is going on. If you came here hoping the answer to why do babies smile in their sleep would be "because she is happy", the research does not say that is wrong. It says nobody can measure it. It simply means a sleep smile cannot prove a dream, a happy feeling, or a memory of you. I find that freeing rather than disappointing. You do not need the smile to mean anything to enjoy it.
The smile she aims at you arrives later
The awake, social smile is a different behaviour with its own timeline. Writing for the AAP's parent site, pediatrician Christine Pagano says a baby's first social smile usually appears by the end of the second month. She describes the earlier grins of the first weeks as primitive and often random, says "we don't quite know" where they come from, and notes that they often seem to respond to an internal signal like gas or hunger.
The AAP's guide to emotional and social development from birth to 3 months traces the shift the same way: at first babies experiment with simple grins and grimaces, and by the second month those expressions begin to turn into genuine signals of pleasure and friendliness. Among its social-emotional milestones by three months is smiling when you talk to or smile at them.
The old assumption was that sleep smiles are a temporary precursor that social smiles replace. Kawakami's group challenged that. A 2007 longitudinal case study followed one boy from birth to the end of his sixth month and still recorded spontaneous smiles in the sixth month (PMID 17292787). A 2009 study found spontaneous smiles in infants between 10 and 15 months old, and its authors argued that theories treating spontaneous smiles as replaced by social smiles must be revised (PMID 19665804). So the sleep smile does not simply switch off. It continues alongside the social one.
According to the sources above, the two kinds of smile tend to differ in a few ways:
- When: sleep smiles appear mostly in active sleep from the first days, while the social smile appears awake, usually by the end of the second month (AAP)
- Trigger: sleep smiles arise without an outside cause; social smiles respond to a face or voice
- Shape: early spontaneous smiles are often one-sided, becoming two-sided by about two months (Kawakami, 2006)
- Mouth: the Sleep Foundation describes REM-sleep smiles as usually closed-mouth and awake social smiles as open-mouthed
When a sleep expression is worth mentioning
For almost every family, a sleep smile is simply a sleep smile. This section is here because the Sleep Foundation's article raises one rare exception, and because it is better to read it calmly here than to stumble on it at 3am.
The Sleep Foundation describes gelastic seizures as an extremely rare type of seizure that causes uncontrollable laughter, generally without the amusement that normally comes with laughing. It says infants may have one on the brink of falling asleep, that they tend to last 10 or 20 seconds, and that they may come with other signs such as squirming, grunting or lip smacking. Its advice is that caregivers who notice concerning symptoms should talk to their pediatrician. A website cannot tell a seizure from a smile, and you should not have to either: if a pattern bothers you, a short phone video is something you can show your pediatrician, who can judge it properly.
The other thing worth raising at a well visit is the social smile itself. The AAP lists smiling when you talk to or smile at your baby among the milestones by three months. Babies vary, and missing a date by a little is not a diagnosis, but development questions are exactly what those appointments are for.
For everything else, the research gives you permission to simply enjoy it. If you want to understand more of what is going on in those early weeks of sleep, the guide to how many hours newborns sleep and our wake windows for a 2-month-old page cover the practical side.
Questions parents ask
Is my baby smiling in her sleep because of gas?
Possibly, but nobody can say for sure. An AAP pediatrician writing on HealthyChildren.org says the grins of the first month often seem to respond to an internal signal like gas or hunger, while also saying 'we don't quite know' where they come from. Research by Kawakami and colleagues describes them as spontaneous smiles that occur without an outside trigger, mostly in sleep, and they have been recorded even in fetuses. Gas is a plausible explanation people reach for, not a proven one. AAP.
Do newborns dream when they smile in their sleep?
The research cannot answer that. Sleep smiles happen mostly in active sleep, which is the infant version of REM, the stage in which adults typically dream. But a 2007 review by Messinger and Fogel states that neonatal smiling has no clear emotional content, and the Sleep Foundation says researchers cannot say for certain what babies are dreaming about. Fetuses smile on ultrasound too, which suggests the smile itself does not depend on experiences to dream about. Sleep Foundation.
Do premature babies smile in their sleep?
Yes, according to the studies that have looked. Emde and colleagues published a study of smiling in REM states in premature infants in 1971. In 2008 Kawakami's group observed 22 premature babies in a neonatal intensive care unit and recorded 95 spontaneous smiles, with the younger and smaller babies smiling more often and for longer. The youngest baby in that study was 200 days from conception and weighed 511 grams. PubMed.
My baby smiles in her sleep but not at me yet. Should I worry?
The two smiles run on different timelines. Sleep smiles can appear from the first days, while the AAP says the first social smile usually appears by the end of the second month, and lists smiling when you talk to or smile at your baby among its milestones by three months. A newborn who smiles asleep but not awake is following the expected order. If your baby is past those ages and you are concerned, bring it up with your pediatrician at the next visit or sooner. AAP milestones.
Sources
- Sleep Foundation — Why Do Babies Smile in Their Sleep? (medically reviewed by Nilong Vyas, MD)
- Sleep Foundation — Baby Sleep overview
- AAP / HealthyChildren.org — Ask the Pediatrician: When do babies first smile?
- AAP / HealthyChildren.org — Emotional and Social Development: Birth to 3 Months
- Emde RN, Koenig KL. Neonatal smiling and rapid eye movement states. J Am Acad Child Psychiatry 1969 (PMID 4306576)
- Emde RN, Koenig KL. Neonatal smiling, frowning, and rapid eye movement states. II. Sleep-cycle study. 1969 (PMID 4317623)
- Emde RN, McCartney RD, Harmon RJ. Neonatal smiling in REM states, IV. Premature study. Child Dev 1971 (PMID 4333441)
- Emde RN, Harmon RJ. Endogenous and exogenous smiling systems in early infancy. 1972 (PMID 4337500)
- Kawakami K et al. Origins of smile and laughter: a preliminary study. Early Hum Dev 2006 (PMID 16185829)
- Kawakami K et al. Spontaneous smile and spontaneous laugh: an intensive longitudinal case study. Infant Behav Dev 2007 (PMID 17292787)
- Kawakami K et al. Roots of smile: a preterm neonates' study. Infant Behav Dev 2008 (PMID 18468691)
- Kawakami F et al. Can we observe spontaneous smiles in 1-year-olds? Infant Behav Dev 2009 (PMID 19665804)
- Kawakami F, Yanaihara T. Smiles in the fetal period. Infant Behav Dev 2012 (PMID 22721744)
- Messinger D, Fogel A. The interactive development of social smiling. Adv Child Dev Behav 2007 (PMID 17682330)
Know a tired parent who needs this?
Send it their way — it is free and always will be.
Keep reading
- Baby Crying in Sleep: Sleep Noise or a Real Wake-Up?
- How Many Hours Do Newborns Sleep? The Sourced Numbers, Week 1 to 12
- Newborn Baby Sleep Schedule: What Weeks 1 to 12 Really Look Like
- 2 Month Old Sleep Schedule: Wake Windows, Naps and a Loose Day Built Around Feeds
- Why Do Babies Fight Sleep? Seven Causes and What Each Looks Like
- Newborn Grunting in Sleep: Noise, Poop or Breathing Trouble?
Building a full day from scratch? Start with the baby sleep schedule guide and calculator on the homepage.

Written by
Sukie Gao
Mom of a daughter, and the person behind every tool on this site.
Sukie Gao has a daughter, and when her daughter was a newborn she was exhausted and completely out of her depth. She spent months piecing together sleep advice from scattered corners of the internet, and built My Baby Sleep Schedule so the next tired mother would not have to search as hard. She writes as a parent from lived experience, not as a clinician.
Last reviewed and updated . Sukie is a parent, not a clinician — check anything that matters with your pediatrician.