A new study in Early Human Development claims that “responsive settling” helps increase infant sleep without the associated stress (for parent and child) of traditional extinction methods

[1].  The research was spearheaded by Dr. Wendy Middlemiss, who is known for earlier research looking at mother-child synchrony and infant stress in traditional extinction sleep training[2].  In her earlier work, Dr. Middlemiss found that three days of extinction sleep training at a residential sleep school resulted in the loss of physiological synchrony between mother and infant at night, but also saw infant’s cortisol as high at sleep time as it was pre-sleep (as opposed to dropping as it normally would), even when the infant was not crying.  This was taken to provide limited support that extinction methods of sleep training were at least acutely stressful for the infant and this continued even when the infant was not outwardly showing distress.  There were limitations to the research – small sample sizes, no home comparison, no long-term follow-up, and so on – but the study did highlight that perhaps the effects on infants were not as benign as we may have previously thought.

This new research – on 34 mother-infant dyads (infants aged 4 to 10 months) – was aimed to examine the effectiveness of a technique deemed to be gentler and more responsive.  The intervention is described as follows:

“During the period of infant settling, parents are encouraged to provide progressively increasing parental care, based on the infant cues and behaviors… Parents are encouraged to listen to the infant and offer care in response to infants’ communication. This guides parents away from either intrusive or non-contingent care. In the event infants cry, parents are coached to respond with low level support care and build it over a short period of time, e.g., no longer than 15 to 30 seconds. This low level of supportive care can include quiet verbal comforting, stroking the infant, patting the mattress, or other low level calming responses. During this short length of time, parents have the opportunity to watch and listen for cues, and the infant has the opportunity to self-regulate.  Parents are advised that it is important that the infant is not left to cry. If soothing without picking up their infant does not calm the infant, parents are encouraged to provide increasing levels of response and attention until they can soothe their infant, which includes cuddling or feeding to comfort.”

This isn’t my own way of working with families, but from the brief description it sounds okay, especially the point about the infant not being left to cry; however, this is a point we shall have to return to later on (i.e., don’t accept it at total face value just yet, even though you, the reader, like me, would take it as such).

Like all residential sleep schools in Australia and NZ, it is a multi-day program (in this case 4 nights) and parents are free to do as they please during the day.  Nurses help implement the sleep portion at night.  In the current study, sleep was also measured approximately nine days after the school at home.  Saliva cortisol was assessed at wake, onset of bedtime routine, and 20 min after sleep (which equates to 20 minutes earlier for each time point as cortisol takes approximately 20 min to enter our saliva so this means it was right before wake, right before the onset of the bedtime routine, and right as the infant fell asleep) during the few days of sleep school and again at the home visit approximately nine days later.

What did they find?

Well, the short of it (and the focus of the entire paper) is that the authors claim to have found that infant sleep improved across the three in-school days and at home.  These improvements were across all infants as well as with subgroups of moms who scored higher on depression, anxiety, or stress.  This was important as often it is these mothers who struggle more and need the most help as we know that these psychological factors can have negative impacts on mother-infant bonding and attachment (for a meta-analysis, see [3]).

The authors take this as evidence that responsive methods can be just as effective as extinction methods without the associated stress or trauma that can result for parents and infants.  This gives parents support in knowing there are alternatives and that these alternatives can actually work.  For many in the field of helping parents, this is something that we face a lot – parents are resistant to “gentle” because they fear it won’t work and believe (somewhat erroneously[4]) that extinction methods always work and work quickly.

Is it so clear-cut?

Of course not.  It never is, is it?  There are quite a few issues worth discussing here that really limit what we can take from the study.

Before getting to my issues, I should say that the authors are clear that this is a preliminary study.  As such, they acknowledge the usual limitations and generalizability of the study.  Yes, there is a small sample size.  No, they didn’t assess every variable they could have.  This is normal for preliminary research and the authors don’t overstate their findings in this regard.  No one should be calling them out for this either because they have already done so.  However, given this isn’t the first type of study of this kind by these researchers and that we need to be able to really understand what’s going on, there are some additional issues that deserve consideration:

  1. The infant sleep time is parent-report (with nurse help the in-school nights) and there is no objective measure of sleep to complement this. When compared to most of the research on sleep training, this puts it squarely in with the rest; however, recent research suggests a discrepancy between what parents report and what infants are actually doing[5][6].  That is, while parents report improvement, objective measures of sleep don’t.  This is important if we want to examine the reasons behind sleep training and to make it clear that it is often for the parent, not the child.  This is another study in which we have no objective evidence that infant sleep was improved, only that parental sleep was.
  2. We have no idea what most parents and infants experienced in terms of levels of distress, need to respond, and so on. That is, because the program itself isn’t clear cut (like extinction sleep training), we are left without information on how much children cried, how long until parents responded, and so on.  This could affect results as the program may work for those who quickly adapt, but not for those who don’t.  Similarly, we don’t know what parents did once home which is problematic.  If there were families for whom things didn’t work, did they try something else or did they give up?
  3. Perhaps most importantly, there is virtually no discussion of the cortisol levels. The pattern here is that during sleep school, the infants’ cortisol increased post-sleep (well, technically at sleep onset given the timing) meaning the infant was more stressed upon falling asleep than before the sleep routine.  This had disappeared by the home visit with the infants showing the typically continued lowering post-sleep.  This matters because one of the reasons the original Middlemiss study is often cited is that the infants showed this same stress response during the training period (but there was no follow-up at home).  This would lead us to one of the following conclusions:
    1. The original preliminary study provides no evidence that sleep training is at all stressful. Some suggested that the increase in cortisol at nighttime is likely due to the new environment, but not only are there reasons to not believe that (based on other work on stress and environmental changes with a caregiver), the current study would also negate that as the pattern from wake to nighttime routine was typical of what one would see at home.  However, something else could be going on to result in the increases that has nothing to do with acute stress with extinction sleep training.
    2. It suggests that this particular sleep training that is deemed “responsive” is still stressful for the infant. It does seem to be acute instead of long-term (though that’s not necessarily something to rejoice in as acute stress is still something to be weary of) given the reduction in stress by the home visit nine days later.  However, it could also be lower at home if families changed what they were doing at night to truly be more responsive.  This seems more likely than option a given what we know about stress and separation from caregivers[7].
    3. Something else I’m not quite thinking of at the moment (entirely possible).

There was also no discussion of synchrony with Mom which had been a huge focus of the first study on extinction methods in these same residential settings[2].  It’s unclear why it wasn’t assessed as it would tie in nicely with the previous study.  If they found asynchrony again, then it raises the same questions as above about either this method or the conclusions reached on the extinction methods.  Finally, the way in which cortisol was reported in the current study makes it impossible to compare to the previous study so we can’t try to gauge the degree of stress the infants experienced between the two studies.

How do we know the program is “responsive”?  Would this mean that all “responsive” or “gentle” techniques are still stressful?

This is probably the crux of the issue:  Is “responsive settling” as it is used in these sleep programs actually responsive and gentle?  The description certainly sounds like it should be with the emphasis on not crying and making sure parents understand and respond to their children’s cues.  The problem I face in writing this is that I happen to know quite a few people that have gone through the “responsive settling” program in Australia and found it to be not as gentle or responsive as perhaps it’s even intended to be.  Given the cortisol data above, it is a question that has to be asked and we need to determine this before we can start to look at whether or not changing infant sleep is inherently acutely stressful or if it’s the methods being used.

I have to caution I’m moving away from science towards the anecdotal now, but it’s important given the limitations in the study in explaining exactly what the process is for families, how families experienced it, and what it meant for sleep for their infants.  Without this data, we have to rely upon what families report.  I should also add that for every family where there has been something negative happen, there is likely many families who feel saved by this program, just as there are for extinction methods.  The issue here isn’t how parents perceive the program, but how gentle it really is for the infant.

So… what have parents said?  I focus on two mothers here because I have their permission to share quotes from them.  One mother was in the program with her 4.5-month-old and another was in it with her 13-month-old child.  For a fuller description from one Mom, I strongly urge you to read a full account here.

In terms of the “responsiveness” of the program, they had this to say:

Source: Unknown

“‘Responsive settling’ in cot, telling me to walk out if he doesn’t lay down etc.”

“It’s ‘responsive’ in that they tell you to respond but it’s not responsive in that they tell how, when, when not, and when to withdraw the responsiveness and in my experience, clearly overriding the mother’s own knowledge of when her baby needs help and the kind of help that is needed.  I was literally held at the door because my baby was “just protesting” and even once he was apparently emotional, first I had verbally reassure him at the door, then shush him from there before moving next to him to do the same, then pat the mattress next to him and then place my hand of his chest, all the while not looking at him and shushing him. This little routine would see my baby red faced and hysterical, so then I could pick him up and calm him but you guessed it, once he was calm, down he’d go again and repeat. A nurse ended up rocking him to sleep in her arms once they decided enough was enough for one session and they’d just say we’d try it again next time.”

But it wasn’t just that the methods weren’t as responsive as the mothers were led to believe.  It was the after effects of how their babies responded to these “responsive settling” techniques that have left these mothers devastated, as you can read here:

“We went to sleep school last weekend and we’re both traumatised from it. Nothing I can do comforts him anymore.”

“Nearly three years on, and I am just trying to claw myself back out of the hell that is left in my head by the trauma of sleep school and sleep training, again. I responded to a question… that went into great detail to try and explain why Responsive Settling isn’t Responsive and it triggered an episode for me that sees me unable to stop reliving the experience and emotions. My head pounds, I get a weird tension in my neck and shoulders that twitches, my heart races and I can’t think of anything else.”

As you can hopefully see, what’s described is regimented and likely quite stressful for some babies (and parents!).  It’s difficult to see this as wholly responsive, though obviously it’s not as non-responsive as something like cry-it-out.  I will add that my own reading and research on anxiety wouldn’t support such a method and it’s not one I would use with families.  However, much more needs to be known about the program and how it is being implemented before any firm conclusions can be made about the program in question.

[Update: I will say that the authors of the research firmly believe the following anecdotes do not represent the methods they utilized despite the mothers firmly believing they were.  There is a known program in Australia that is in many maternal & child health units and uses this term, but it is possible that there is a new program designed that is using the same “responsive settling” name with different programming from the original.  This puts me in a difficult situation.  I don’t want to misrepresent the methods, but I also want to make sure that all voices are heard.  I have offered to put the parties in contact to each other to try and sort this out.  At the heart of this, though, it highlights that the term “responsive settling” can clearly mean very different things and this in and of itself is an issue.  If this is resolved, I will make sure to update this accordingly.  As is, my comments about what I have heard with the “responsive settling” program stand.]

So where does this leave us?

In a big, messy gray zone.  Honestly.  I know some people will want to use this to show that “responsive” techniques can increase sleep too.  And for the parent who is exhausted and at his/her wit’s end, this may be welcome news, but the cortisol data and the anecdotes raise questions about the degree to which this method is truly responsive.  Other gentle methods don’t have any science because they often aren’t used in residential settings like the one studied herein, in part because truly gentle methods take time and can’t be done in 3-4 nights.  I know my own methods aren’t “tested” because of these issues (though if I win the lottery, I’ll fund a blind third party to study it), but until then parents will have to make their own decisions about the degree of responsiveness they are hoping for and find someone or some method that respects that.


[1] Middlemiss W, Stevens H, Ridgway L, McDonald S, Koussa M.  Response-based sleep intervention: helping infants sleep without making them cry.  Early Human Development 2017; 108: 49-57.

[2] Middlemiss W, Granger DA, Goldberg WA, Nathans L.  Asynchrony of the mother-infant hypothalamic-pituitary-adrenal axis activity following extinction of infant crying responses induced during the transition to sleep.  Early Human Development 2012; 88: 227-32.

[3] Martins C, Gaffan EA.  Effects of early maternal depression on patterns of infant-mother attachment: a meta-analytic investigation.  The Journal of Child Psychology and Psychiatry and Allied Disciplines 2000; 41: 737-46.

[4] Loutzenhiser L, Hoffman J, Beatch J.  Parental perceptions of the effectiveness of graduated extinction in reducing infant night-wakings.  Journal of Reproductive and Infant Psychology 2014; http://dx.doi.org/10.1080/02646838.2014.910864.

[5] Hall WA, Hutton E, Brant RF, Collet JP, Gregg K, et al. A randomized controlled trial of an intervention for infants’ behavioral sleep problems.  BMC Pediatrics 2015; 15: 181.

[6] Gradisar M, Jackson K, Spurrier NJ, Gibson J, Whitham J, et al. Behavioral interventions for infant sleep problems: a randomized controlled trial.  Pediatrics 2016; 137: e20151486.

[7] See here for a discussion of hyporesponsivity: http://evolutionaryparenting.com/its-just-a-little-cortisol-why-rises-in-cortisol-matter-to-infant-development/