Sleep education

Psychoeducation about sleep has been one of the fundamental and most common approaches to address infant sleep problems. It is usually the main modality in prevention programs, delivered during pregnancy or in the first months postpartum. It is also often incorporated in treatment programs aimed at alleviating existing infant sleep problems, typically in infants aged >6 months (e.g., Gradisar et al. ). Sleep education may cover a broad spectrum of topics, including typical sleep development, circadian rhythms, the effects of light exposure, sleep homeostatic pressure, sleep architecture and spontaneous awakenings between sleep cycles, sleep hygiene, constructive sleep associations, links between sleep and feeding, and parental involvement in the sleep context. The potential benefits of preventing sleep problems before they occur are vast, given the considerable physiological, emotional, and economic costs of such problems. Due to the low intensity of many sleep education interventions, they also have the potential for high cost-effectiveness and can be used as a first step in stepped care approaches.

In the AASM’s 2006 practice parameter paper, parent education/prevention was deemed a Standard treatment approach for young children, based on modified Sackett criteria. This indicated a generally accepted patient care strategy, with high-quality evidence. Still, the recommendation was based on four studies that were available at the time. A 2010 Cochrane review found that postnatal parental education programs resulted in increased infant sleep duration, but not increased sleep consolidation. This review was based on randomized controlled trials (RCTs) alone, yet only four studies were identified, precluding any exhaustive conclusions regarding the efficacy of prevention programs. A meta-analysis published in 2016 identified nine RCTs that examined the efficacy of sleep-focused interventions administered in the perinatal period. Consistent with Bryanton and Beck, benefits were observed for infant sleep nighttime duration but not for nighttime awakenings. Small improvements in maternal mood were also detected in this meta-analysis, although the authors suggested that this result was not definitive, due to a potential risk for publication bias.

Since then, several additional investigations into the effects of psychoeducation on infant sleep have been conducted, generally yielding null to small effects. Galland and colleagues conducted a large ( N = 802) RCT to test the effects of group sleep education delivered antenatally and a home visit at 3 weeks postpartum. These authors found that parent-reported infant sleep problems at 6 months of age were equivalent between groups that received sleep education and those that did not. Moreover, no significant differences were found between groups in infant self-settling and benefits were not documented for parent depression, nor for parent fatigue or sleep. Actigraphic infant nighttime awakenings were significantly less frequent in the intervention group, yet this difference was not deemed clinically meaningful.

Santos et al. evaluated the impact of sleep education delivered at 3 months postpartum on infants’ sleep duration (assessed using actigraphy and parent reports) compared to a feeding education control group. This too was a large-scale RCT ( N = 586) and included follow-up assessments at 6, 12, and 24 months of age. The intervention was delivered in two home visits and two telephone calls when infants were 3 months old. In line with Galland et al.’s findings, no significant differences were found between groups following the intervention, despite some trends in the expected direction (e.g., longer nighttime sleep duration in infants whose parents received sleep education). A third RCT assessed whether a two-session group psychoeducation intervention delivered during the third trimester to first-time mothers would prevent postpartum sleep problems. Mothers who received the intervention and mothers who only received information booklets did not differ significantly in ratings of infant sleep quality. At 4 months postpartum, maternal sleep quality was higher in the intervention group, and insomnia symptoms were lower, yet these effects were small and not apparent at 6 weeks or 10 months postpartum.

The null or very modest effects detected in these three recent large and well-controlled RCTs may be due to several factors, such as low adherence, the exclusion of fathers from prevention programs, and the high potential for floor effects when psychoeducation is provided to educated mothers of typically developing infants. Preventative interventions addressing infant sleep may be more beneficial when delivered beyond the antenatal or newborn period, once parents have gotten to know their infant better and got through the “fourth trimester.” Moreover, given these latest RCTs, it has been suggested that resources such as home visits and face-to-face counseling may be more wisely invested in delivering interventions to families at risk or to those that already report infant sleep problems.

When provided to parents who already perceive their infant’s sleep as a problem, psychoeducation has usually been offered in conjunction with other treatment modalities (e.g., graduated extinction), making it difficult to evaluate its therapeutic effect as a standalone treatment. Sleep education is often used as a control group for other intervention strategies (e.g., Gradisar et al. ), but to the best of our knowledge, has not been compared to a nonactive control when treating infants with insomnia. Moreover, the topics covered may vary significantly between studies and protocols, possibly rendering psychoeducation a “catchall category.” For example, some educational interventions have guided parents to wait a few minutes before attending to the infant when they wake at night, ^, resembling extinction-based approaches. Additionally, many educational interventions emphasize the importance of a consistent structured bedtime routine. ^{, ,} This overlap in content blurs the distinction between psychoeducation and other intervention approaches, making the specific benefit of psychoeducation unclear. More research is needed to determine which topics should be covered in sleep education interventions, perhaps by developing standardized protocols.

In summary, almost all forms of behavioral interventions for infant sleep problems include a psychoeducation module of some kind. Within the context of prevention programs, these have generated null to small benefits for infants and parents. Providing the education after the first few months, when parents have gotten to know their infant, may lead to superior outcomes. When provided later on in development, in families who report problems with their infant’s sleep, psychoeducation usually forms the basis for other treatment components (e.g., education about sleep homeostatic pressure to provide a rationale for bedtime fading). Despite its modest efficacy, psychoeducation could be used as a low-intensity, low-cost, first-step approach to address mild to moderate cases of infant insomnia. Future work is warranted to determine the efficacy of psychoeducation approaches for diverse and at-risk populations, which content modules should be included in these interventions, what would be the optimal timing of delivery, and whether low-cost highly accessible delivery modes (e.g., digital programs) yield equally beneficial effects.

Bedtime routines

Implementing consistent and appropriate bedtime routines is a highly ubiquitous pediatric sleep health recommendation. Bedtime routines may include 2 to 4 activities, such as bathing and reading stories, lasting no longer than 30 to 40 minutes, followed by lights out time. Consistency is key, and parents are advised to repeat the same activities at the same time each night, for as many nights a week as possible. Building on stimulus control therapy techniques, and accordant with the idea of “sleep hygiene,” bedtime routines aim to strengthen associations between sleep-promoting stimuli (e.g., a massage) and sleep onset and to avoid stimuli that may interfere with sleep onset (e.g., technology devices). Sleep then becomes an anticipated and rewarding experience, rather than an unpredictable and potentially unpleasant occurrence. Consistent routines hold the potential to regulate not only behavioral and emotional responses to the sleep context but also the child’s circadian rhythms, provided that bedtimes are consistent across nights.

The links between consistent implementation of bedtime routines and young children’s sleep have been demonstrated in multiple studies. Cross-sectional and longitudinal investigations have identified associations between bedtime routines and earlier bedtimes, shorter sleep onset latencies, longer nighttime sleep duration, and more consolidated sleep. Benefits extend beyond the sleep domain, and include reduced bedtime resistance, and improved child and parent mood, emotion regulation, and child literacy outcomes.

Intervention studies have further established the role of bedtime routines in pediatric sleep. A recent scoping review of behavioral treatments for pediatric insomnia found that bedtime/positive routines were the most commonly studied approach, included in 29 (61.7%) of the 47 studies investigating treatments for insomnia in young children. Most of these studies examined the use of bedtime routines as part of multicomponent behavioral sleep interventions, not allowing for evaluation of their individual therapeutic potency. Yet, the efficacy of bedtime routines or individual prebedtime activities as independent interventions has been demonstrated in a few studies. Mindell and colleagues randomly assigned mothers of children aged 7 to 36 months old with perceived sleep problems to either a bedtime routine or to a control condition. Instructions for the bedtime routine group included implementing the following three steps over approximately 30 minutes: (1) bathing; (2) massage or lotion; and (3) quiet activities, such as cuddling or singing a lullaby. Mothers in the control group were instructed to continue their child’s usual routine. Two weeks following the intervention, significant reductions in maternal reports of sleep onset latency and in the frequency and duration of nighttime awakenings were found in the bedtime routine group, but not in the control group. Mothers who implemented bedtime routines also reported having improved mood and decreased child bedtime resistance.

Studies into the effects of prebedtime massage have also found beneficial effects. In a study of 23 infants and toddlers, a 15-minute massage before lights out time resulted in lower bedtime resistance and shorter sleep onset latencies compared to reading a bedtime story. In a later study, massage with lotion was found to be superior to massage without lotion and to a no massage condition, in reducing mother-reported nighttime awakenings and increasing sleep duration in newborn infants. The mechanisms by which prebedtime massage may exercise its benefits may include positive parent-child interactions, muscle relaxation, and circadian regulation. Evidence for the latter is indicated in a study by Ferber et al., revealing that 30-minute prebedtime “massage therapy” for 2 weeks, starting when infants were 10 days old, led to higher levels of nocturnal melatonin secretion (6-sulfatoxymelatonin excretions) at 12 weeks of age, compared to control. These authors concluded that the massage served as a zeitgeber, facilitating the alignment of infants’ circadian rhythms to the 24-hour day. While these findings provide preliminary evidence for a physiological pathway through which bedtime routines may promote sleep, further investigations are warranted to establish these effects and examine them in infants with pediatric insomnia. Furthermore, more research is needed to explore other potential mechanisms of bedtime routines and their individual components, and so are studies examining diverse infant populations (e.g., infants with neurodevelopmental conditions and diverse cultural backgrounds).

Bedtime fading

In the late 1980s Sleep Restriction Therapy (SRT) was developed as a behavioral intervention for chronic insomnia in adults. SRT involves implementing a prescribed time-in-bed that matches the individual’s baseline average total sleep time. This new restricted schedule is intended to regularize and entrain circadian rhythms and assure that homeostatic sleep pressure is sufficiently elevated when sleep is attempted, so that time in bed is mostly spent asleep rather than awake. This in turn reconditions sleep-related associations, so that the bed and bedroom environment are more closely associated with feeling sleepy and falling asleep.

In the early 1990s, a pediatric variation of this treatment was introduced as “bedtime fading” by Piazza and Fisher. Using a similar technique, in which the sleep window is restricted and titrated (by first delaying bedtimes and subsequently “fading” them earlier by 15 minutes at a time), these authors successfully treated insomnia in two young children in an inpatient unit. Bedtime fading has since yielded promising results in several other case reports in both typically and nontypically developing preschool children. Still, the efficacy of this intervention remained unsubstantiated due to the relatively limited number of investigations assessing its effects, many of which were of poor quality. Importantly, most of these reports used multicomponential treatment packages, in which bedtime fading was delivered together with other treatment modules (e.g., positive routines), making it impossible to evaluate the specific contribution of each technique. Consequently, the 2006 AASM practice parameters recommended bedtime fading as a Guideline treatment, indexing only a moderate level of clinical evidence for this treatment.

In contrast to the increasing recognition and popularity of this approach in the treatment of adult insomnia, bedtime fading has remained a rather overlooked approach in pediatric sleep research and practice. This is quite surprising, given the considerable advantages this intervention may have in lowering the probability of child distress and bedtime resistance. These advantages may be particularly important when treating infants and toddlers, whose parents often seek gentle intervention approaches.

Two more recent investigations have reinstituted the efforts to establish an evidence base for bedtime fading in young children. In an open trial, Cooney et al. provided two group sessions to mothers of 1- to 4-year-old children with behavioral insomnia. Treatment included basic sleep education and a bedtime fading protocol. Following treatment, large reductions were observed in children’s sleep onset latency, nighttime wakefulness, and the frequency of bedtime tantrums. Gradisar and colleagues tested the efficacy of bedtime fading for infant sleep problems in an RCT. Infants aged 6 to 16 months were randomized to receive either bedtime fading, graduated extinction, or sleep education control. Treatment was delivered in one individualized session, with the option of subsequent phone call support. At posttreatment, parents of infants who received bedtime fading reported large reductions in sleep onset latency, which were comparable to graduated extinction and greater than those observed in the control group. Bedtime fading also generated reductions in maternal stress compared to control, but not in nighttime awakenings. This study provided the first high-quality evidence for the efficacy of bedtime fading to improve the sleep of infants. Additional RCTs are warranted to replicate these findings and potentially form a sound evidence base for the routine use of this intervention in clinical practice.

Scheduled awakenings and dream feeding

The scheduled awakenings procedure was developed to enhance sleep consolidation in infants. In this approach, parents wake their child 15 to 30 minutes prior to the timing of their typical spontaneous awakening to soothe, feed, or facilitate the onset of a new sleep cycle. When spontaneous awakenings diminish, parent-induced awakenings are faded out by gradually increasing the time increments between them. This technique was first introduced in 1980, and was subsequently evaluated in a series of small studies. While benefits were documented, not all parents adhered to the technique, and the latency to see improvement, as well as its extent, were found to be inferior to extinction. Based on this evidence, scheduled awakenings were considered “probably efficacious” and regarded as a Guideline treatment by the AASM 2006 task force. This approach has received very limited research attention since and is seldom recommended in clinical practice. The reasons for this may include the lack of consistently timed awakenings in infants, making it difficult for parents to time their induced awakenings. In addition, parents may prefer not to disrupt their infant’s and their own sleep, in case spontaneous awakenings do not occur.

Despite the dearth of studies into scheduled awakenings over the past three decades, a derivative of this technique—“dream feeding”—seems to have accrued much more popularity. In this method, parents initiate a nighttime feed before going to bed themselves. As opposed to the original scheduled awakenings protocol, the timing of the “dream feed” is determined by the parents’ sleep schedule, aiming to reduce the chance of infant awakenings while parents are asleep. Moreover, parents are not advised to wake their infants per se, but rather feed them while asleep. This technique is meant to reduce awakenings triggered by hunger, especially in young infants who still need to be fed throughout the night.

Dream feeding is commonly suggested in self-help books, websites, and colloquial advice for infant sleep problems. ^, Yet, scientific evidence for its efficacy is lacking. To date, two controlled trials have incorporated dream feeding for young infants (aged <4 months) as part of a multicomponent treatment. ^, While both studies documented benefits for infants receiving intervention, the use of comprehensive “package” treatments did not allow conclusions regarding the efficacy of dream feeding as a standalone component in enhancing sleep consolidation. Hence, given its apparent ubiquity, systematic research is clearly warranted to determine the efficacy of this technique for improving infant sleep.

Extinction-based methods

Based heavily on classical and operant conditioning principles, extinction-based interventions aim to provide parents with skills on how to modify and reduce their involvement in the infant’s sleep context. These changes are expected to foster children’s ability to fall asleep independently at bedtime and after waking up at night (i.e., to self-soothe), thus promoting consolidated sleep. To that effect external regulation—such as feeding, holding, or rocking the baby to attain sleep—is targeted and decreased. Extinction-based approaches correspond with evidence for concurrent and longitudinal links between increased parental involvement in the infant’s sleep context and infant sleep fragmentation and disruption (see Chapter 12 for a comprehensive description of these links). Infants usually require similar conditions to initiate sleep at the start of the night and throughout it. Hence, given that spontaneous awakenings frequently occur in the first year of life, facilitating the development of self-soothing to sleep has the potential to reduce nocturnal wakefulness and allow more consolidated sleep for both infants and parents.

Several variations of extinction-based approaches have been developed and studied. The three most common variations are unmodified extinction, modified (or gradual) extinction, and extinction with parental presence. These will be described in the following sections.

How well do behavioral interventions compare?

The various behavioral sleep intervention approaches described above were presented as separate standalone entities, yet in clinical trials and practice they are often combined or delivered in succession. In their recent scoping review, Meltzer and colleagues reported that in roughly 82% of the investigations into these treatments, a combination of multiple treatment approaches was used. Furthermore, Honaker et al. found that approximately 25% of parents using extinction-based techniques start with one approach and then switch to another, with the most common transition occurring from modified to unmodified extinction. Thus, the treatment of infant sleep problems is rarely unidimensional, and optimal efficacy might require flexibility, and a willingness to try various components.

Several studies have compared the efficacy or effectiveness of behavioral sleep interventions, aiming to gauge whether some methods are superior to others. Most of the RCTs that compared the efficacy of techniques found null to small differences between methods. For example, in a trial of 36 young children, Adams and Rickert found that delaying bedtimes in addition to positive routines was as beneficial in reducing young children’s bedtime tantrums as modified extinction, both of which were more beneficial than the nonactive control. However, improvements were achieved more quickly following bedtime fading with positive bedtime routines. Moreover, the latter approach yielded improvements in marital satisfaction, which did not occur in the other two groups. The authors attributed these gains to the additional opportunity for family “quality time” when bedtimes are delayed and routines include positive parent-child activities, as opposed to the distress that is often involved in modified extinction.

As mentioned earlier, the RCT conducted by Gradisar and colleagues compared the efficacy of modified extinction, bedtime fading, and a sleep education control. Here too benefits were demonstrated following both intervention groups compared to control, yet there were some indications of superior benefits for infants undergoing modified extinction. Specifically, while sleep onset latency was similarly reduced in both intervention groups, sleep consolidation metrics (nocturnal wakefulness and the number of awakenings) indexed greater improvement in graduated extinction compared to bedtime fading (which did not differ from control). In their conclusions, the authors proposed that these two intervention approaches could be delivered in succession, with bedtime fading increasing sleepiness, shortening the procedure of parents checking in and out, and thus facilitating the learning of more independent sleep associations.

Different variations of extinction-based approaches were compared in three randomized trials. ^{, , ,} Matthey and Črnčec compared modified extinction with and without parental presence in 16 infants aged 6 to 18 months and found that equal proportions of infants had reductions in the number of nighttime awakenings in both groups. Sadeh conducted an RCT to compare the efficacy of modified extinction (“checking”) and unmodified extinction with parental presence in 50 infants aged 9 to 24 months with sleep disturbances, using actigraphy in addition to parent-reported infant sleep. Both assessment methods revealed equivalent improvement in infant sleep between both intervention groups at posttreatment. Kahn et al. ^, similarly used multimethod assessment in an RCT comparing modified extinction (“checking-in”) and modified extinction with parental presence (“camping out”). This trial had a larger sample ( N = 91) and included a 6-month follow-up assessment. Corresponding with Sadeh’s findings, benefits were indexed for actigraphic and parent-reported infant sleep following both extinction-based approaches, with no significant group by time interaction effects. ^, Of note, these two RCTs revealed a similar pattern, in which benefits were greater and more robust according to parent-reports compared to actigraphic monitoring. ^{, ,} This pattern points to the mechanistic pathway by which benefits are exercised, suggesting that during “sleep training” with or without parental presence, infants acquire self-regulatory capabilities, and are thus less inclined to signal to their parents when waking at night.

Comparisons between extinction-based approaches were also examined in Honaker et al.’s retrospective survey of parents who had used these interventions with their infants. Here too, differences were not found between approaches on most of the assessed outcomes. Specifically, unmodified extinction, modified extinction, and each of these combined with parental presence did not differ in reported improvements in infant sleep-wake patterns and parental stress, nor did they differ in infant age, the amount and intensity of infant crying, or parental stress at time of implementation. However, differences were found in the rates of parent satisfaction, with only 76% of parents who used modified extinction with parental presence reporting being satisfied, compared to 94%, 91%, and 87% in the three other groups. Furthermore, parents using unmodified extinction reported the greatest first-attempt success rate. Yet, the sample for this study was recruited via social media peer support groups for parents using behavioral sleep intervention, and thus findings may not be representative of the general parent population.

While small differences in efficacy and effectiveness have been found in the studies described above, several factors make it difficult to draw conclusions regarding their relative effects. First, different studies have used different treatment protocols, including varying checking-in schedules (e.g., length of waiting periods, implementation including bedtime only versus nighttime and naps as well), different bedtime fading schedules, and varying content covered in psychoeducation interventions. While real-world implementation of these interventions requires flexibility and pragmatism, developing standardized protocols that entail certain degrees of freedom would allow more careful evaluation of their comparative effects. Similarly, the use of different designs, populations, and outcome measures across studies limits the ability to assess the relative efficacy of these interventions. In particular, only a few studies have included objective assessment of sleep in addition to parent reports. Given the risk of recall bias when it comes to nighttime events, and considering the potential for objective measures such as actigraphy and auto-videosomnography to evaluate sleep-related events that parents were not aware of, ^, including these measures provide a much more comprehensive account of the infant’s sleep-wake patterns. Future investigations should thus use standardized protocols and standardized multimethod assessment of outcomes. This would allow future meta-analytic studies to draw more accurate conclusions regarding efficacy that would better inform clinical practice guidelines.

Notwithstanding these suggestions, it might be the case that focusing on the “horse race” question might not be the best use of research resources. Namely, instead of directing our efforts at identifying which interventions work best, we might have more to gain by attempting to identify how , for whom , and under what circumstances do they work best. Scientific knowledge regarding predictors, moderators, and mediators of treatment has advanced considerably in the realm of behavioral interventions for various child psychopathologies. ^, However, pediatric sleep research has only recently begun to address these questions. The following section reviews findings from these preliminary attempts and outlines a path for further investigations in these directions.

At what infant age are (which) behavioral sleep interventions most efficacious?

One of the most important questions regarding behavioral interventions for infant sleep has to do with timing. Parents often ask whether there is a youngest appropriate age to intervene and what approach is suitable for different stages in development. Despite the gravity and ubiquity of this question, there is limited evidence to answer it. The dramatic development of sleep-wake patterns and underlying physiology during infancy ^, has led many to doubt the value of intervening at an early age. As discussed at length later in this chapter, it has been argued that behavioral sleep interventions “pathologize” the normal developmental trajectories of infant sleep, particularly in the early stages of development. A systematic review focusing on interventions implemented during the first 6 months of life indicated that despite evidence for increased self-regulated sleep following these interventions, they do not improve other infant or mother outcomes. Other systematic reviews and meta-analyses found increases in infant sleep duration, as well as benefits for maternal mood, yet these included studies of sleep in the first 12 months and beyond.

Importantly, most studies of behavioral interventions implemented during the first 6 months have evaluated either psychoeducation, bedtime routines, or a combination of both. In their meta-analysis of RCTs conducted within the first year of life, Kempler et al. note that none of the interventions administered within the first 6 months employed extinction-based approaches. We have identified only one prospective investigation of such methods which included infants aged <6 months. Middlemiss et al. assessed the effects of unmodified extinction in 4- to 10-month-old infants, finding that by the third intervention day all 25 participating infants settled to sleep independently without signaling distress. Other studies of extinction-based approaches have included infants who were either ≥6 or ≥9 months old. Hence, it seems that while preventative programs—focusing mostly on educating parents about realistic expectations and positive routines—have been shown to be beneficial early on, the efficacy of extinction-based methods has been demonstrated primarily from the second half of the infant’s first year.

Importantly, to the best of our knowledge no study has yet to examine whether infant age acts as a predictor or moderator of treatment outcomes. Namely, it has not been tested whether older infants derive more benefit from a certain intervention compared to younger infants (potentially confirmed using infant age-by-time interaction effects in repeated measure designs), nor has it been tested whether this potential age effect might vary as a function of the specific intervention implemented (potentially confirmed using infant age-by-time-by-intervention type interaction effects in repeated measure designs). Given the widespread use of these interventions with young infants, and the dearth of studies examining their effects on this age group, research is warranted assessing whether infant age predicts and/or moderates outcomes of behavioral sleep interventions.

For whom are (which) behavioral sleep interventions most efficacious?

Other child-related factors may also impact the outcomes of behavioral interventions for insomnia. For example, since extinction-based methods usually entail a certain degree of separation from the parents, heightened infant separation anxiety might impede compliance to these interventions, as infants may express increased distress when left on their own, making it more difficult for parents to delay their response to the infant. Indeed, in an RCT of 91 infants aged 9 to 18 months old focusing on predictors and moderators of modified extinction versus camping out, Kahn et al. found that infant separation anxiety moderated treatment efficacy. While significant reductions in nighttime wakefulness were documented for low-anxiety infants in both intervention groups, and for high-anxiety infants in the camping-out group, no significant reductions were observed for infants with high separation anxiety in the modified extinction group. This moderation effect was found for both parent-reported and actigraphic nighttime wakefulness. Other sleep metrics, such as sleep onset latency and the frequency of nighttime awakenings, were reduced following both interventions, regardless of infant separation anxiety. Still, these findings suggest that interventions involving a smaller “dose” of separation from parents may be more beneficial for infants who exhibit heightened separation anxiety.

Parent factors were also examined in this RCT, considering the important part parents play in the development of infant sleep, and given that behavioral interventions for infant insomnia primarily target parent behaviors. Two related parent factors were examined in this trial: parental cry tolerance and sleep-related cognitions. These constructs have previously been implicated in infant sleep problems both concurrently and longitudinally. ^{, ,} In this RCT, both low cry tolerance and distress attributions predicted inferior treatment outcomes. Namely, reductions in the frequency of nighttime awakenings and in parent-reported sleep problems were greater when parents exhibited higher tolerance for crying and lower distress-attribution cognitions at baseline. These effects were demonstrated across modified extinction and camping-out groups, with no evidence for one treatment to be more beneficial than the other for parents with a certain baseline profile of cry tolerance or sleep-related cognitions. Given that both extinction-based methods require parents to refrain from providing immediate comfort and tolerate some extent of infant distress, parents who are highly sensitive to these stimuli may be less able to successfully implement the intervention. Interestingly, following the intervention, cry tolerance significantly increased, and distress-attribution cognitions decreased. This suggests that a positive feedback loop may have emerged, in which the increased ability to tolerate distress and the reciprocal increase in child independent sleep were mutually reinforcing.

Findings from this trial represent a first step in the transition to a precision medicine paradigm in the field of behavioral infant sleep interventions. They inform clinicians by indicating which parent populations may require more intensive support when attempting these interventions. They also advise practice guidelines as to which infant and parent factors should be taken into account when recommending a certain treatment method. As opposed to the one-treatment-fits-all approach, this knowledge can be used to personalize treatment so that it meets the needs of specific patient populations. More research is clearly needed to identify further parent and infant factors that may predict and/or moderate treatment success. Specifically, factors that have been shown to be associated with infant sleep problems should be considered, including infant temperament, physiological reactivity, atypical development, breastfeeding, cosleeping, as well as parent culture and ethnicity, health literacy, emotional availability, and psychopathology. ^, These factors should be examined within RCTs that include multiple types of interventions to determine which approach would most likely be appropriate and beneficial for each specific treatment-seeking family.

Is there actually such a thing as “infant sleep problems”?

Firstly, we wish to address the argument that there might not be a need for behavioral interventions for infant insomnia at all, but rather that normal infant sleep is culturally constructed and pathologized into an unwarranted diagnostic category. ^{, ,} These claims stem from biological and anthropological theories of development. From a biological standpoint, human infants are born at the most immature neurological stage compared to all other primates and are completely dependent on external care for many months. More specifically, the physiological processes governing sleep (i.e., the homeostatic and circadian processes) gradually mature over the first year of the infant’s life. ^, Thus, sleep in newborn infants is typically fragmented into multiple brief sleep episodes distributed around the 24-hour day and gradually consolidates into one major nocturnal sleep episode over the course of the first year. Studies of typically developing infants indicate that although the number of night-wakings gradually declines and most infants develop self-soothing skills during the first year, nighttime awakenings do not entirely disappear. In fact, they are the norm throughout this year and beyond, with most infants presenting an average of approximately 1 to 3 awakenings between the ages of 3 and 18 months. ^{, ,} Thus, from a biological standpoint, one could argue that the expectation that babies “sleep through the night” is not realistic and discounts their normal biological developmental progression.

Sociocultural and anthropological arguments augment these claims by stating that prior to the 19th century, infant sleep was not considered a major concern for parents. Families would cosleep in communal beds, and infants would be attended to quickly upon awakening. Similar perceptions of and practices around infant sleep have been documented in modern times in nonindustrial nonwestern settings. The evolution of cultural norms around infant sleep was shaped by expansive developments, such as the strengthening of individualistic values, the industrial revolution, the increasing entrance of women into the work force, the emergence of human milk substitute formulas, the sexualization of the parents’ bed, and the adoption of “scientific motherhood” according to which medical professionals became a more trusted authority regarding child rearing than parents themselves. ^{, ,} It is thus argued that “infant sleep problems” are a modern invention, whereas the real problem lies in holding unrealistic expectations and trying to force a 21st-century Western industrialized lifestyle on an infant brain which has failed to evolve at the same pace. According to this viewpoint, infant-parent cosleeping with responsive breastfeeding should be recommended, and parents should allow their children’s sleep to become consolidated in their own time.

When contemplating these arguments, we believe there are a few points on which agreement could easily be achieved. These include the fragmented nature of infant sleep, especially during the first months of life, making it inherently misaligned with the sleep-wake patterns of most adults in contemporary modern society. They also include the dramatic evolution of cultural norms around sleep in these societies in the past centuries and the acknowledgment that there are infinite ways in which infant sleep can be perceived and constructed. We should, however, be wary not to over-romanticize the past. Historically, cosleeping was practiced for reasons such as assuring appropriate body temperature, as well as protection from predators that nighttime crying may attract. Moreover, historic evidence indicates that it was a widespread practice to drug babies to sleep, with indications of opium use in infants dating back to the 16th century BCE and continuing well into the 20th century. ^, Infants who were “sleepless,” “restless,” or displayed “fits of crying” were regularly provided opiate “soothers,” to the point that opium intoxication was one of the main causes of infant mortality. Therefore, it seems that the perception of infant sleep as a problem should not be treated as a modern invention since parents in ancient times were sometimes willing to use life-endangering measures to get their infants to sleep.

Furthermore, we must pose the question—would recommending practices that are harmonious with 16th-century social norms, necessarily do justice with parents living in capitalistic, industrialized 21st-century settings? Infant sleep problems are a source of major difficulty for present-day parents. They are the most common issue for which professional assistance is sought in the first 12 months of the infant’s life. Mounting evidence associates infant sleep problems with a multitude of parent adverse consequences, including depression, anxiety, stress, poor physical health, and reduced quality of life. ^{, ,} Parent sleep disturbances in the postpartum period have been associated with negative caregiving emotions and behaviors. In extreme cases, these problems have been associated with child neglect and abuse, such as infant shaking, and even filicide. ^, Behavioral sleep interventions have been shown to significantly reduce maternal depression and benefit parental well-being. ^, The mechanisms through which these benefits are achieved are evident, given the vast literature demonstrating the devastating consequences of sleep deprivation and fragmentation in adults. Thus, helping parents obtain sufficient consolidated sleep may benefit both themselves and—indirectly—their infants, who gain more rested, contented, and regulated caregivers. Of course, infants also stand to gain direct benefit from behavioral sleep interventions, given that sleep problems in infancy have been shown to predict later cognitive, behavioral, and emotional difficulties, as well as the incidence of obesity. ^{, , ,}

Therefore, while infant sleep problems may indeed be a culturally constructed notion, they nevertheless may cause substantial adverse consequences, and thus call for attention and intervention. In addressing these issues, we argue for a middle ground in which parents are offered education regarding the “fourth trimester,” the normal progression of infant sleep, realistic expectations, and the potential influences of their specific cultural context. For those who experience these issues as burdensome and distressing, the full spectrum of available treatment approaches should be proposed.

Are behavioral sleep interventions for infants safe?

The second major argument against extinction-based interventions focuses on their safety. Concerns have been raised as to whether these interventions generate undue stress, compromising parent-infant attachment and the infant’s long-term emotional well-being. At the basis of these claims stands the notion that crying is a universal form of early communication, which promotes parental proximity and fulfillment of the infant’s needs. ^, At the preverbal stage of their development, crying is the most effective means by which infants can communicate their distress and call for external support. Crying is also considered by attachment theorists to be part of a biological system that activates the attachment system and establishes a stable reciprocal bond of trust between infant and parent. ^, Parents are thus biologically primed to infant cry signals, as demonstrated in neuroimaging studies of “parental-brain network” activation in response to infant crying stimuli. ^,

Attachment researchers have claimed that ignoring an infant’s cry means defying fundamental evolutionary drives and disregarding the infant’s invitation to develop a shared relationship of trust and understanding with their parent. Extinction-based methods in which parents do not immediately respond to their infant’s crying at night are thus seen as unnatural, unethical, and harmful in that they may cause unnecessary stress in the short term and hinder the development of secure attachment between infant and parent in the long term. ^{, ,} Notably, the argument here is not that extinction-based methods do not improve infants’—and by proxy parents’—nighttime sleep consolidation and duration for the majority of families who implement them. Rather, whether these improvements should be the criteria by which treatment “success” is determined and whether they outweigh the potential “side-effects” is called into question.

The links between parental responsiveness and infant behavior and development have stimulated vast theoretical and scientific discussions over the past decades. In their classical longitudinal Baltimore study of 26 infants during their first year of life, Bell and Ainsworth reported that nonresponsive parenting early on predicted more crying in subsequent months. This was viewed as evidence to support the attachment-based approach, endorsing immediate parental responsiveness to crying. Approximately 30 years later, van IJzendoorn and Hubbard attempted to replicate and extend this study, using a larger sample ( N = 50) and more sophisticated methodology (e.g., measuring crying bouts objectively via audio recordings, and statistically controlling for important intervening variables). Astoundingly, not only did these authors fail to replicate Bell and Ainsworth’s Baltimore findings, but rather the direction of relationship between responsiveness and later crying was reversed. More frequent ignoring of infant crying early on predicted less crying later on. Pure behaviorists would presumably interpret this finding in terms of operant conditioning, arguing that immediate and frequent parents’ responsiveness reinforces crying behavior, subsequently making it more recurrent and persistent. van IJzendoorn and Hubbard, on the other hand, suggested a more intricate interpretation, according to which “Benign neglect” of crying and fussing may stimulate infants’ emergent abilities to self-regulate and cope with distress. Further evidence for these findings was reported in a recent longitudinal study of 178 infants followed from birth throughout the age of 18 months. Maternal reports of the use of “leaving the infant to cry it out” were associated with subsequent lower frequency and shorter duration of crying. Moreover, no adverse impact of occasionally or even frequently leaving infants to cry in the first 6 months was documented on observed infant behavior (e.g., attention, social referencing, aggression) or infant-mother attachment at 18 months of age.

Bilgin and Wolke justly emphasize that their results neither suggest that leaving infants to cry it out should be generically recommended nor that it should be avoided. Instead, they embrace the concept of “differential responsiveness” to convey the complexity of the relationship between parental responsiveness and infant emotion regulation. From this standpoint, greater and prompter parental responsiveness is not always better. Rather, parents are required to constantly adapt to the ever-changing infant needs and abilities and the ever-changing external circumstances, intuitively providing both high levels of emotional warmth and appropriate limits and challenges. These perceptions are in line with several psychodynamic theories of child development, including Donald Winnicott’s concepts of the “good enough mother” and the “capacity to be alone,” as well as Heinz Kohut’s idea of “optimal frustration.” It is hence the delicate, dynamic balance between fulfillment and nonfulfillment of the child’s urges that provides the foundation for self-regulation. Studies demonstrating the associations between authoritative parenting and positive child outcomes provide evidence for this viewpoint. ^, Furthermore, the natural progression of parents’ responsiveness throughout infant development suggests that different stages in development require differential responsiveness. For example, Bilgin and Wolke report that the “leaving infant to crying” was uncommon at term, yet it increased across the next 18 months of the infant’s life. Correspondingly, parental tolerance for infant crying has been shown to be lower in parents of infants compared to childless controls and to significantly decrease from pregnancy to 6 months postpartum.

Coming back to behavioral sleep interventions, to date, systematic investigations have failed to find support for negative impact on infant-parent attachment or child behavior and emotion at posttreatment or up to 5 years later. ^{, , ,} Thus, interpretations according to which the infant-parent relationship is harmed, or that infants acquire “learned helplessness” in the process of behavioral sleep interventions lack any scientific foundation to date.

With regard to potential short-term harm, two studies have assessed infants’ salivary cortisol levels before and after undergoing an extinction-based sleep intervention. The first study assessed the effects of a 5-day inpatient unmodified extinction program in 25 infants aged 4 to 10 months. Salivary cortisol levels were sampled from mothers and infants on days 1 and 3 of the intervention, both before initiation of the nighttime sleep routine and 20 minutes after infant sleep onset. Results revealed that infant cortisol levels did not change significantly following unmodified extinction on either day. Namely, the expected increase in stress hormones following the intervention was not observed. Similarly, mothers’ cortisol levels did not change from before to after the intervention on day 1, whereas on day 3 there was a significant decrease in their cortisol levels following the intervention. Authors attribute this decrease to the reduction in infant crying on the third day, in which all infants settled to sleep independently without signaling distress. In line with these patterns of results, the associations between infant and mother cortisol levels were significant on day 1, and nonsignificant on day 3, which was interpreted as asynchrony between mother’s and infant’s stress levels. Noteworthy is that a lack of correlation between infant and mother cortisol levels has not been associated with any negative child or parent outcome in research work performed thus far. Moreover, the authors conclude that infants’ cortisol levels “remained elevated” during unmodified extinction; however, at the time there was no normative data to describe normal infant cortisol levels over the course of a day. Since then, reports have been published demonstrating normal cortisol levels (e.g., Ivars et al. ), suggesting that cortisol levels of babies in Middlemiss et al.’s investigation were high (i.e., above the 75th percentile) at baseline. This could be explained by the fact that infants were transitioned to the inpatient unit and nurses—rather than mothers—implemented the interventions, which could be stressful in itself. Nevertheless, and most importantly, infant cortisol levels were not affected by unmodified extinction, which is arguably the most intense extinction-based approach.

Gradisar and colleagues also investigated the effect of behavioral sleep interventions on infant salivary cortisol levels in their aforementioned RCT. These authors randomized 43 infants aged 6 to 16 months to either graduated extinction, bedtime fading, or education control. Salivary samples were collected at baseline, and again at 1 week, 1 month, and 12 months following treatment. While infant sleep latency and fragmentation significantly diminished in both intervention groups, no significant increases were found in infant cortisol levels from baseline to the 1-week assessment in any of the groups. Thus, in line with Middlemiss et al.’s findings, infant stress levels remained unchanged following behavioral sleep interventions. Moreover, at the 12-month follow-up no differences were apparent between groups in parent-reported infant emotional and behavioral problems, or in infants’ attachment style, as assessed using the strange situation procedure.

In conclusion, research to date has not provided any indication of short- or long-term harm resulting from behavioral interventions for infant sleep problems. Consistent high-quality evidence demonstrates the efficacy and effectiveness of these treatments. This evidence should be made clear to parents, and efforts should be made to revise any communications of misinformation. At the same time, we should clearly acknowledge that behavioral interventions for infant insomnia may not be appropriate for everyone. Many parents feel that they do not fit well with their emotional resources, belief systems, or life circumstances at that time. Clinicians should make parents aware of the broad range of evidence-based treatments available and respect their choice to implement any or none of them, as research continues to refine existing interventions and develop novel approaches. Importantly, we should all be wary not to let ourselves get caught in a “Judgement of Solomon” scenario, wherein amid our attempts to prove that we are the righteous authority, we are willing to risk rendering parents torn between two sides of a highly polarized and depleting dispute.