What Was Learned From Studying the Effects of Early Institutional Deprivation

Abstract

The effect of experiences in infancy on human development is a central question in developmental science. Children raised in orphanage-like institutions for their first year or so of life and then adopted into well-resourced and supportive families provide a lens on the long-term effects of early deprivation and the capacity of children to recover from this type of early adversity. While it is challenging to identify cause-and-effect relations in the study of previously institutionalized individuals, finding results that are consistent with animal experimental studies and the one randomized study of removal from institutional care support the conclusion that many of the outcomes for these children were induced by early institutional deprivation. This review examines the behavioral and neural evidence for altered executive function, declarative memory, affective disorders, reward processing, reactivity to threat, risk-taking and sensation-seeking. We then provide a brief overview of the neurobiological mechanisms that may transduce early institutional experiences into effects on brain and behavior. In addition, we discuss implications for policy and practice.

How important are our earliest experiences? This is a central question in developmental psychology and one that has drawn scientists to studying children reared for their first years in institutions and then adopted or fostered into families. While early institutional care is an extreme version of early adversity, it has the advantage of being time limited and thus allows examination of the importance of the earliest years of life. The early years encompass sensitive periods when the neural systems supporting basic skills and competencies are establish through the interaction of genes and stimulus inputs. Once established, these foundational circuits can be difficult to alter. This is sometimes referred to as the early experience hypothesis (see Zeanah et al., 2011); however, it is difficult to test because there tends to be strong continuity between early and later environments. Of course, this is not the case for children adopted or fostered out of institutional care. These children experience a marked change in conditions allowing a test of the early experience hypothesis and an examination of which skills or competencies seem to be established early and difficult to change with subsequent experiences and which are more plastic and open to change as the child grows older (see for extended discussion, Zeanah et al., 2011). Of course, because children cannot be randomly assigned to institutional care and both age of exposure and duration are typically confounded, we must be cautious in interpreting this type of evidence for sensitive periods.

Few experiments of nature are clean. One does not end up in an institution by chance nor is the duration of time in institutional care random. However, the animal model studies (see Sanchez et al., 2001) provide strong experimental support for many of the outcomes observed for previously-institutionalized children. In addition, the Bucharest Early Intervention Project (BEIP: Zeanah et al., 2003) involved random assignment when children were between roughly one and three years to either remain in the institution or whatever care the Romanian government determine (care-as-usual) or placement in study-supported foster care. This study also supports conclusions drawn from studies of children adopted from institutional care. Thus, there is good reason to believe the effects observed for previously-institutionalized children are due to the care infants receive in institutional settings (van IJzendoorn et al., 2011).

This review will cover what we have learned about neurobehavioral development from the study of individuals who spent their first years in institutional care. Although much of the work on previously-institutionalized (PI) children has been approached from a deficit standpoint, using an evolutionary perspective we can question which effects might reflect trade-off’s to protect life and thus reproductive fitness at the expense of later morbidity (Hostinar & Gunnar, 2013). Furthermore, we might find that adapting to such a harsh, deprived environment might provide what has been termed “hidden talents” (Ellis et al., 2020). We will consider these possibilities, as well as recognizing that despite the legacy of early institutional deprivation, many if not most of the children carrying the burden of this legacy adapt, compensate, and ultimately do well, speaking to the resilience of the human species. Finally, this essay will conclude by briefly considering the policy implications of this information.

Nature of Institutional Care

It is estimated that worldwide between two and three million children are in institutional care (Petrowski et al, 2017). While institutions vary in quality, there are commonalities driven by the need to care for a large number of children by a limited number of staff. For infants, interactions with staff center on diapering, feeding, dressing, and toileting, thus creating an assembly line quality to care. Staff frequently rotate and children may receive care from ten or more people in a week. Individualized care in response to the infant’s signals is rare. This means that even when physical needs are met, the need for response-contingent stimulation and relationships are lacking. Because for many months infants cannot get their hands to their mouths, roll over, or do much at all to provide themselves with stimulation except to signal caregivers, the contrast in experiences between a baby reared in even the most average family and one reared in an institution is stark.

Thus, institutional care deprives infants of stimulation (McLaughlin et al, 2017) and responsive care from one or a few consistent caregivers (Zeanah et al., 2005). While it would be convenient to separate stimulus deprivation and lack of relationships into effects on cognitive and socioemotional functioning, respectively, this is too simplistic. A quasi-experimental study of Russian Baby Homes showed that stimulation, while important, is not sufficient for either cognitive or emotional development. They trained workers in one institution to be more sensitive and responsive and to provide more stimulation, while in another institution they provided this same training but also altered staffing so that babies received care from fewer, more consistent caregivers (Groark et al., 2005; McCall et al., 2018). Compared to the control, care-as-usual institution, infants in the training-only institution did marginally better. The best outcomes were in the institution where workers were trained, and staffing was altered to increase the infants’ opportunities to form relationships. While psychologists have tended to emphasize deprivation of stimulation and relationships, it is also important to note that children adopted or fostered from institutions frequently arrive in their families stunted in linear growth and carrying intestinal parasites (Hostetter et al., 1989; Johnson, 2000). Intestinal parasites and stress reduce nutrient absorption contributing to iron and other micronutrient deficiencies noted at adoption (Fuglestad et al., 2008). Thus, infections, parasites and poor nutrition also need to be considered aspects of institutional care through which the impact of early institutional care may flow. This is especially important because infections, parasites and poor nutrition can impact neurobehavioral development, as shown in children growing up in poverty in low and middle income countries (Xie et al., 2019).

Brief History of Research on Early Institutional Care

Orphanages have been around since as early as the 14^th century (Boswell, 1988); however, research on how infants faired in them did not emerge until the mid-20^th century. This early research emphasized the harm that such rearing produced (Goldfarb, 1945; Provence & Lipton, 1962; Skeels & Dye, 1939). The accumulation of these findings led to a shift from institutional care to foster care in the United States, Western Europe, South Korea and Scandinavia; however, orphanage-like institutions remained common in the rest of the world.

The fall of communism in Russia and Eastern Bloc countries brought to light the plight of the many children being reared in institutions, along with a surge in inter-country adoption from these countries and later from China, Vietnam, India and Latin and South America. This wave of inter-country adoption also led to a resurgence of interest in studying the impact of early institutional care. However, attention shifted from documenting behavioral problems and cognitive delays to examining effects on neurobehavioral development (Zeanah et al., 2003). Furthermore, with the increase in non-invasive tools to study potential physiological mechanisms, the goal of this research has evolved into understanding how early deprivation “gets under the skin” to affect physical and mental health.

The Nature of Post-Institutionalized Care.

Outcomes for previously institutionalized children depend not only on the nature and duration of institutional care, but also on the quality of post-institutionalized care. Studies of the homes in the United States and Great Britain into which PI children were adopted have shown that these home are generally well-resourced with highly educated parents (Hellerstedt et al., 2008). Adoption disruptions in international adoption are relatively rare (SAMHSA Report, 2015). In the English and Romanian Adoptees study (ERAs), the researchers examined many of the traditional sources of negative family inputs to children’s development and generally found both low rates and no evidence that these adoptive family factors explained negative outcomes associated with early institutional care (Castel et al., 2010).

The exception to the rule of generally high quality and stable post-institutional care is the care experienced by children in the BEIP study. While the study-supported foster care homes were of good quality, once the study ended the children in the foster care group returned to being overseen by government agencies in Romania. Over the course of their childhoods, children in both the care-as-usual and foster care group experienced multiple caregiving disruptions, including movement between foster care homes and between birth families and foster care and institutions (range=2 to 9 by age 12 years). The quality of the care in birth families, foster care and adoptive homes also varied from dangerous to excellent. Several recent reports from this study document the importance of disruptions in care for impairing children’s emotional health and brain electrical activity (Almas et al., 2020; Debnath et al., 2020). In addition, the quality of care predicted children’s executive functioning, reward sensitivity, and internalizing and externalizing symptoms well into adolescence (Colich et al., 2021).

Neurocognitive Functioning

Lower IQ is one of the most frequently reported effects of early institutional care (van IJzendoorn et al., 2020), and is correlated with reduced brain volume in PI children Eluvathingal et al., 2006; Hodel et al., 2015; Mackes et al., 2020; Mehta et al., 2009; Sheridan et al., 2012). However, IQ rebounds remarkably once children are removed from institutional care and placed in families. Furthermore, in the ERA study those children with the lowest IQs at adoption continued to show improvements in IQ even through adolescence (Sonuga-Barke et al., 2017). Despite this, several aspects of cognition that have important ramifications for both cognitive and socioemotional functioning exhibit effects that persist long after family placement, suggesting that the early months of life constitute a sensitive period for the shaping of the neural systems that support them.

Executive Function.

Persistent, seemingly intractable deficits in executive function (EF, i.e., working memory, planning and sequencing, and executive attention) is observed in PI children and adolescents (Colvert et al., 2008; Hostinar et al., 2012; Loman et al., 2013; Pollak et al., 2010). Attention deficit, hyperactivity disorder (ADHD) is also frequently noted (Mueller et al., 2017; van Ijzendoorn et al., 2020). This is the case for children adopted beyond 6 months of age, but not for those adopted before this point (see Colvert et al., 2008). Deficits in EF may contribute to other deficits noted in PI individuals, including poorer language development, problems with theory of mind, and indiscriminate friendliness (i.e., responding with intimacy appropriate for relationship partners when interacting with unfamiliar people) (Bruce et al., 2009; Colvert et al., 2008). These findings are consistent with evidence that EF is a transdiagnostic contributor to psychopathology risk in PI adolescents (Wade et al, 2020) and adults (Golm et al., 2020).

Because EF relies on fronto-parietal circuitry (McLaughlin et al., 2017; E.K. Miller & Buschman, 2013), and attention deficit problems, in addition, reflect alterations in fronto-striatal networks (Casey et al., 1997), we can expect that brain correlates of early institutional care will include these networks. Results of neuroimaging studies support this expectation. In a study of PI adolescents, reductions in brain volume (controlling for total brain volume) were most pronounced in the prefrontal cortex, driven by reductions in surface area rather than cortical thickness (Hodel et al., 2015). In the BEIP study, both children randomly assigned to leave the institution for foster family care and those assigned to care-as-usual when assessed later at age 8 years exhibited widespread reductions in cortical thickness that were most pronounced in the prefrontal cortex and lateral and medial parietal cortices (McLaughlin et al., 2014). Furthermore, cortical thickness in the lateral orbital frontal cortex, inferior parietal cortex, precuneus, superior temporal gyrus, and lingual gyrus mediated associations between institutional care and ADHD symptoms in that study. Prefrontal connectivity is also altered in PI youth (Behen et al., 2009), and a number of studies have shown that prefrontal white matter organization is disrupted (Eluvathingal et al., 2006; Govindan et al., 2010; Hanson et al., 2013).

Studies using electroencephalograms (EEGs) are consistent with the neuroimaging data. While still in institutional care, infants in the BEIP study exhibited greater low frequency theta power over posterior scalp regions, and less high frequency alpha power over frontal and temporal regions, a pattern associated with delayed development (Marshall et al., 2004). Followed up at 42 months, after those randomized to foster care had been in families, children placed at earlier ages exhibited higher alpha power, suggesting more rapid rebound in neural development with earlier intervention (Marshall et al., 2008). In another study, similar results were obtained with 18-month-olds adopted primarily from China, and in addition predicted EF at 3 years of age (Tarullo et al., 2011). There is evidence that with time in the family, EEG power normalizes if children are placed in families early enough, and as it normalizes it predicts better EF as PI youth move into adolescence (Sheridan et al., 2012). Furthermore, normalization of EEG tracked normalization of white matter volume which also showed an age-at-placement effect with less normalization if removal from institutional care happened after 24 months of age (Sheridan et al., 2012). Thus, there may be a sensitive period after which it was more difficult for white matter and EEG maturation to rebound. However, this was not found for gray matter volume in the prefrontal cortex, which did not normalize in the BEIP study for the foster care group regardless of whether they were placed in families before 24 months or not (Sheridan et al. 2012), nor was it different for those adopted before and after 12 months in the study by Hodel et al (2015). This suggests that if there is a sensitive period for the establishment of gray matter structure in the prefrontal cortex it is before the postnatal age of the youngest children removed from institutional care in these studies (i.e., 4 months in Hodel et al, 2015, 9 months in Sheridan et al., 2012). This is consistent with continued problems with EF in children adopted beyond 4–6 months.

Several studies have used event-related potentials (ERPs) to examine EF associated neural activity. In the BEIP study, ERPs to go/nogo and flanker tasks were studied at several time points. At 8 years, using a flanker task, regardless of foster placement, children who experienced early institutional care performed more poorly on incongruent trials and exhibited smaller Pe amplitudes at central and parietal electrodes (an index of error monitoring) than the never-institutionalized children (McDermott et al., 2013). At 12 years, using a go/nogo task, similar results were obtained along with evidence that P2 amplitudes were larger and N2 responses were smaller over medial-frontal scalp electrodes (Lamm et al., 2018). The ERP findings mediated some of the associations between duration of institutional care and behavior problems. Interestingly, at 12 years, children randomly assigned to foster families did not differ from never-institutionalized children in the error related negativity (ERN) signal, although the care-as-usual group did (Troller-Renfree et al, 2016). Studying PI children averaging 11 years of age adopted from many regions of the world, Loman and colleagues (Loman et al., 2013) used both a go/nogo and flanker task. PI youth performed more poorly on these tasks, particularly on congruent and go trials, suggesting problems in sustaining attention. ERP data indicated problems in error monitoring (i.e., N2 on the go/nogo task and ERN on the flanker task). Importantly, this study also included children adopted internationally from countries using foster families. These children did not differ from the never institutionalized comparison group, increasing the likelihood that something about institutional care was disrupting EF development.

Since persistent deficits in EF and reductions in brain volume are so prominent in the lives of previously-institutionalized children, we can wonder whether they reflect some sort of trade-off that supported survival in the institutional context. One quite reasonable suggestion is that the brain and especially the prefrontal cortex is not that critical to early survival. When the environment signals that resources, including human resources, are in short supply, it may be more important to divert or retain metabolic resources to keep the heart pumping, fight infections, and preserve thermal regulation. We will return to this point later when we discuss the impact of early institutional care on physical growth and development.

Memory.

Early institutional deprivation predicts deficits in the hippocampal-based memory functions (McLaughlin et al., 2017). PI children score lower on paired associate learning tasks (Bos et al., 2009; Pollak et al., 2010) which assess visual episodic memory combined with associative learning. Recognition memory is also impacted (Bos et al., 2009; Guler et al., 2012). These effects persist at least into adolescence (Wade et al., 2019). However, despite evidence of problems on hippocampal-dependent memory tasks, evidence of gross structural impacts of early institutionalization are equivocal. No differences in hippocampal volume were noted in two studies (Sheridan et al., 2012; Tottenham et al., 2010), while significantly smaller hippocampi were noted in another two, one reporting a dose response relation with duration of institutional care (Hodel et al., 2015; Mehta et al., 2009).The children in these last two studies were older, which may be relevant. Studies using other neuroimaging measures have more consistently implicated alterations in medial temporal lobe function. Using positron emission tomography (PET), reduced glucose metabolism in the head of the hippocampus was noted in one study (Chugani et al., 2001). Using a test of episodic memory previously shown to be impaired in individuals with hippocampal lesions, PI children exhibited a smaller P300 ERP component over parietal leads (Guler et al., 2012). The P300 is believed to reflect hippocampal activity. Thus, it may be that functional alterations in hippocampal activity are easier to detect in children than structural alterations which might be revealed as individuals age.

Summary of Neurocognitive Functioning.

While IQ for the most part has been found to rebound remarkably once young PI children are placed in families that are supportive and provide consistent care, two aspects of cognition do appear to be impacted in ways that may contribute to deficits in multiple domains. Most clearly affected is EF and the neural systems that support executive functions. Hippocampal memory appears to be somewhat affected, although evidence of gross morphological effects of early institutional deprivation are equivocal and may not emerge until later in development. Again, from an evolutionary perspective this effect may reflect trade-offs to preserve life when the signals reaching the child’s body indicate a marked lack of supportive care and thus an environment where adaptation requires conservation of energy resources.

Socioemotional Functioning

Affective Disorders

The early studies of institutional care emphasized affective problems. A recent meta-analysis, in contrast, concluded that although individuals with institutional histories are more likely to suffer from anxiety and depression, the effect sizes are small (van IJzendoorn et al., 2020). This is the case even though PI youth are more likely to receive mental health services (Juffer et al., 2004). PI children and youth are at much higher risk of disinhibited social engagement disorder (DSED; (Zeanah & Gleason, 2015). This disorder is characterized by intimate social advances to unfamiliar adults. In PI youth, DSED is seen even in those who have formed specific and apparently secure attachments to their adoptive parents. There is currently debate over whether it predicts serious pathology or may co-exist with relatively healthy emotional functioning (Kennedy et al., 2017). Indeed, it has long been suspected that this behavior may have evolved among children deprived of consistent adult attention to increase the attention the child receives in the institution, thus helping children get their needs met (Chisholm, 1998).

Notably, while anxiety and depression among PI individuals increases in adolescence and young adulthood (Golm et al., 2020; Kumsta et al., 2010), recent analyses have shown that this is the result of a developmental cascade in which problems in neurocognitive functioning predicted problems in employment and relationships in young adulthood, which in turn leads to clinically significant emotional problems (Golm et al., 2020). Furthermore, there is some suggestion that adolescence may be a time when it is especially critical for PI youth to experience consistent high quality relationships in their family environments. In the BEIP study, family care quality during adolescence, more so than middle childhood, predicted socioemotional health in early adulthood (Colich et al., 2021). Pathways through relationships and academic problems in adolescence may explain why, in addition to family quality, EF in PI adolescence was found to be a transdiagnostic indicator of mental health problems in the BEIP study (Wade et al., 2020).

Reward Processing.

The frequent diagnosis of ADHD among PI children may signal problems in reward circuitry as well as placing individuals at risk of developing depression. While the dominant theory of ADHD emphasizes executive dysfunction and prefrontal circuitry, some (e.g., Sonuga-Barke, 2003) have argued that there is also a motivational pathway implicating the meso-limbic branches of frontal-striatal circuitry that terminate in the nucleus accumbens (NAcc). In functional imaging (fMRI) studies, there is evidence of hyporesponsivity of the NAcc among PI adolescents and young adults (Goff et al., 2013; Mehta et al., 2010). Hyporesponsiveness to reward may not only contribute to problems in motivational components of attention and behavior regulation, but may also reflect more general issues with associative learning (McLaughlin et al., 2017; Wismer Fries & Pollak, 2017). Indeed, using a modified monetary incentive delay task, PI youth who had spent more of their early lives in institutional care were less sensitive to reward, failing to respond more quickly on high versus low reward trials (Sheridan et al., 2018).

In animals, appetitive stimuli trigger a reflex that cups the ears forward. In humans, the muscle that moves the ears in this way is vestigial, but its response can still be measured. Termed the post-auricular reflex (PAR), it is larger in response to positive, approach-eliciting pictures. Adolescents show larger PARs than children, reflecting their enhanced reward sensitivity. In one study, PI adolescents did not show higher PARs than PI children, while comparison adolescents did (Quevedo et al., 2008). Thus, like the Goff et al. (2013) study, PI youth did not follow the expected developmental increase in reward sensitivity in adolescence.

Reactivity to Threat.

In animal models, early deprived care is associated with increased fearful behavior and premature emergence of adult patterns of fear conditioning and extinction (Pattwell & Bath, 2017). Thus, there has been a very strong expectation that PI children and youth would be more fearful and would show alterations in amygdala reactivity and connectivity (Callaghan & Tottenham, 2015). However, the data are equivocal. One study that included PI as well as children in domestic foster care noted larger amygdala responses to fear relative to neutral faces (Maheu et al., 2010). Another reported a larger response to fear expressions than fixation point (Tottenham et al., 2011). In yet another study, PI youth did not show a larger amygdala response, although their responses and not the responses of comparison youth correlated with anxiety scores (Silvers et al., 2017). Turning to amygdala volume, we know of only one study that reported larger amygdala volume and this was only in relation to more prolonged institutional care prior to adoption (Tottenham et al., 2011). Other studies have reported no difference (Hodel et al., 2015), or smaller amygdala volumes (Hanson et al., 2015). This is consistent with evidence from a systematic review of all studies of amygdala volume and all types of early deprivation and neglect (McLaughlin et al., 2019).

Attention bias to threat, another way of examining threat reactivity, is often examined using dot-probe tasks. The lack of a bias is typical in non-anxious populations, and this is what was found among the never-institutionalized youth in the BEIP study (Troller-Renfree et al., 2017). Interestingly, this was also found in the care-as-usual group, while the youth who had been randomly assigned to foster care showed a bias to positive stimuli, and no threat bias. Furthermore, the positive bias appeared to be protective against affective symptoms and stress hormone dysregulation. This could perhaps be an example of a hidden talent emerging among previously institutionalized children.

Emotion modulated eye-blink startle is another way to assess threat reactivity. To our knowledge, only one published study has examined eye-blink startle in PI youth (Quevedo et al., 2015). Controlling for age, fear-potentiated startle did not differ between PI and comparison children who were pre- or early pubertal. However, among participants who were mid/late pubertal, PI youth had blunted fear potentiation. Similar to studies of reward sensitivity, the expected increase in emotion-sensitivity in adolescence (Silk et al., 2009) may not be observed in PI youth.

Finally, we can examine the threat of negative social evaluation. The Trier Social Stress Test (TSST) is a social evaluative stressor that threatens the social self (Dickerson & Kemeny, 2004). In one study, neither children nor adolescence showed differences in self-rated stress/threat between PI and comparison participants (Gunnar et al., 2009). In another, however, PI youth rated themselves less stressed/threatened than comparison youth, and this was particularly so among the adolescent participants (DePasquale et al., 2019). Thus, overall there is little evidence of greater fearfulness or sensitivity to threat among PI youth. It has been suggested that this might reflect a difference in the outcomes of deprivation versus exposure to fear-eliciting early environments (McLaughlin & Sheridan, 2013).

Risk Taking and Sensation-Seeking.

Taking risks is not pathological; indeed, calculated risks may be necessary for healthy development. However, when the risks are not judged wisely, risk-taking does threaten health. Dual-factor models of adolescent risk-taking argue that the increase in adolescent risk-taking is due to heightened reward sensitivity (as discussed above) in the context of more gradual development of impulse-control (Casey et al,, 2008; Steinberg, 2004). Tripartite models, on the other hand, argue that the extent to which the decisions adolescents make are risky depends on the weighing of three processes: reward sensitivity, cognitive control or EF, and affectivity, most notably fearfulness (Crone & Dahl, 2012). We have already discussed PI functioning on the three components in these models. Here we ask whether there is evidence for PI children and youth taking more risks.

One task used to study risk-taking is the Balloon Analogue Risk Task (BART) in which the participant pushes a button to inflate a balloon. The more inflated the balloon, the greater the points received. However, the balloons are programmed to pop at random inflations and when they pop, points are lost. In three studies, PI youth exhibited less risk-taking on the BART relative to comparison youth (Herzberg et al., 2018; Kopetz et al., 2019; Loman et al., 2014). In the Hertzberg et al. (2018) study this was the case for youth adopted above 12 months of age, but not for those adopted before 12 months providing evidence that duration of deprivation matters. Results from a different risk-taking task yielded similar findings with PI youth (Humphreys et al., 2015). One study examined associations between brain structure and the BART using a Region of Interest (ROI) analysis. In that study orbital frontal cortex (OFC) and anterior cingulate cortex volumes correlated with less risk-taking, and these volumes were also smaller in PI youth (Herzberg et al., 2018). Similar results were obtained when reports of actual risk-taking behaviors were measured (Kopetz et al., 2019). While the Iowa Gambling Task is often used to assess risk-taking and risky-decision making, we know of only one study that used this task with PI youth and it yielded no evidence of riskier decision making (Herzberg et al., 2018).

As with risk-taking, sensation-seeking is not necessarily negative, as there are socially acceptable risk-taking motives that can encourage psychological growth (e.g., I like to do things that are frightening), as well as socially unacceptable motives (e.g., I like to do new things even if they are illegal). In the BEIP study, sensation-seeking was examined in the same report in which the BART was used (Kopetz et al., 2019). Both the care-as-usual and foster care youth scored lower on sensation seeking than the never-institutionalized youth, but this was true only for the unacceptable motives in the foster care group. In another study, independent of motives, PI youth who were pre/early puberty scored lower on sensation-seeking than did comparison youth, while mid/late PI youth did not (Loman et al., 2014). In another study, youth adopted after 12 months, but not before scored lower on sensation-seeking than comparison youth (Herzberg et al., 2018). Thus, the results on sensation-seeking are nuanced, with PI youth scoring lower on interest in sensation-seeking, but this effect seems stronger for less physically mature youth and those with longer periods of early institutional care. Whether this serves an adaptive function for PI youth is not clear. Being cautious, avoiding risk and sensations related to unacceptable motives might well improve survival under conditions of psychosocial deprivation. If so, then this might be an example of the earliest years of life being a sensitive period for establishing the neural circuitry regulating risk and sensation seeking, which might be difficult to recalibrate once life conditions are not threatening or indeed are highly supportive.

Summary of Socioemotional Functioning:

While some PI youth exhibit marked psychosocial problems, as a group the effect sizes for these problems are small, although they may increase by early adulthood. Furthermore, there is reason to believe that when these problems are significant in adolescents and young adults, they reflect earlier emerging neurocognitive deficits and failures leading to problems in relationships, academic failures, and bleak employment prospects. It should be noted, however, that many PI youth do exceedingly well. It is also noteworthy that PI youth do show significant alterations in component processes that contribute to socioemotional functioning. Thus, PI youth are less sensitive to reward, especially perhaps failing to show the expected adolescent increase in reward-sensitivity and sensation-seeking. As a result, they do not report being interested in or actually taking as many risks as adolescents without early histories of deprivation. Notably, despite strong expectations from animal models, there is mixed evidence on whether PI youth respond more intensely to threat or are more generally fearful and anxious.

Physical Health Impacts

Although psychologists have focused on the cognitive and emotional sequelae of early institutional deprivation, health effects are beginning to be reported. Linear growth delay is one of the most robust effects of early institutional care (van IJzendoorn et al., 2020). The pattern of growth delay is reminiscent of psychosocial dwarfism in that linear growth is affected, with weight remaining proportional to height. Once placed in a family, rebound growth occurs, often so rapidly as to outstrip the child’s iron stores (Fuglestad et al., 2008). Rapid catch-up growth may be a risk factor for obesity and poor cardiovascular health among PI youth (Tang et al., 2018). However, we have shown that PI youth who were moderately to severely stunted at adoption remain slender well into adolescence (Reid et al, 2017). Nonetheless, by early adolescence they exhibit higher systolic blood pressure, higher total cholesterol, higher triglycerides, higher insulin, and evidence of acceleration of arteriolar stiffening (Reid et al., 2018). Cardiovascular health and risk of obesity might also be associated with early onset puberty, which was initially reported to be one of the risks of institutional care for girls (Mason & Narad, 2005). Nonetheless, better controlled studies have failed to find evidence of pubertal advance in either PI girls or boys (Reid et al., 2017; Tang et al., 2018). The study of health consequences of early institutional care has increased as the ages of those adopted from institutions has increased. It is very likely that the health outcomes of early institutional rearing will become more apparent as these individuals age.

It should be noted that the slowing of linear growth under conditions of deprivation is likely an evolutionary adaptation (Hochberg et al., 2011). This slowing of growth reflects interactions between the stress system (notably the hypothalamic-pituitary-adrenocortical (HPA) system) and the growth hormone (GH) system. In cases where the deprivation is psychosocial, administering growth hormone does not produce growth; instead, removing the child from the stressful context, even without medicating with growth hormone, allows growth to rebound. Presumably, the slowing of growth supports survival by diverting energy from future-oriented activity (i.e., becoming tall) to the service of immediate survival needs. The trade-off may be a heightened risk of later physical and, since the brain is also delayed/reduced in growth, mental health problems.

Pathways Through Which Early Experiences Impact Development

We now turn to considering the physiological and psychological processes that help explain how early institutional care “gets under the skin” to influence later development. Earlier we noted that two aspects of institutional care appear to be critical for its effects on infants and young children: lack of developmentally appropriate stimulation, especially response-contingent stimulation (McLaughlin et al., 2017) and lack of consistent adults (i.e., attachment figures). This latter, we argued, leaves the infant and young child open to experiencing excessive stress activation (Hostinar et al., 2014). These two are intertwined because for much of the infants’ young life, stimulation contingent on the child’s actions and signals is mediated by responsive adults. Such individualized care is rare in institutional settings, but common in even the most average family. Mechanisms through which early institutional care impacts neurobehavioral development likely flow out of these two aspects of institutional deprivation.

Deprivation Induced Over-Pruning and Apoptosis

There are several mechanisms through which stimulus deprivation may affect neurobehavioral development. The most widely considered is accelerated synaptic pruning. As the brain processes information, it confirms neural pathways and prunes back those that are not used. When there is a dearth of stimulation, this process can result in a reduction in cortical thickness. In rat separation models, the dearth of stimulation doubles the cell death in the cerebral and cerebellar cortices, leading to reduced brain volume. Reduction in white matter has also been found in these studies. We do not know that this is occurring in human infants during institutional care, but it would be consistent with overall reductions in gray and white matter volume, as reviewed earlier. Deprivation-induced apoptosis in the brain, however, is usually discussed in the context of stress. Indeed, there has been much more attention paid to mechanisms associated with stress than mechanisms associated with stimulus deprivation.

Institutional Deprivation and Stress.

To understand the mechanisms transducing early institutional care into neurobehavioral effects, researchers have often turned to rodent models that manipulate maternal care as it affects the HPA axis (Levine, 2005). Of course, the HPA axis forms only part of the mammalian stress system. In this section while we will focus on the HPA axis, we will also include studies of the sympathetic-adrenomedullary (SAM) system. This leaves out neurohormonal systems in the brain that are also important in transducing the impacts of stress (Joëls & Baram, 2009), but which are not accessible in studies of children.

While there are many aspects of institutional care that might activate stress, the lack of responsive care is the most likely. This is because responsive care regulates activity of the HPA axis as early as the newborn period (e.g., De Bernardo et al., 2018). In contrast, the lack of it increases cortisol, even in response to everyday caregiving activities like bathing (Albers et al, 2008). By the time the infant has formed a specific attachment, the presence and availability of the attachment figure can prevent elevations in cortisol to normative events that frighten the infant, even as the child cries and seeks contact (Gunnar & Donzella, 2002). Thus, in the absence of a responsive, consistent caregiver, the infant is very vulnerable to stress.

There are only a few studies of HPA axis activity in infants and young children while they are still in institutional care. Two studies revealed altered diurnal rhythms with lower than expected cortisol levels early in the morning and slightly higher late afternoon levels (Carlson & Earls, 1997; Kroupina et al, 1997); however, studies of somewhat older (3–6 year old) children found higher overall cortisol production among institutional children (Dobrova0Krol et al., 2008). Among PI children and adults, the most consistent finding has been of a blunted cortisol rhythm and blunted responses to stressors prior to puberty (Hengesch et al., 2018; Koss et al., 2016; Kumsta et al, 2017; McLaughlin et al., 2015). This hypocortisolism pattern is associated with attention and conduct problems (Koss et al., 2016). Furthermore, the blunted rhythm is predicted by institutional deficits in social care (Koss et al., 2014). It should be noted, that among rhesus monkeys, those randomly assigned to nursery care (no mother) have low cortisol levels and blunted HPA axis responses both to psychosocial stressors and pharmacological stimulation (Capitanio et al, 2006). Thus, the correlational PI data are consistent with the experimental nonhuman primate data in showing early parental deprivation produces hypocortisolism. Hypocortisolism may result from changes at many levels of the axis (Herman et al., 2012). Reduced activity of the axis may help protect stress-sensitive brain regions, such as the hippocampus and amygdala from toxic effects of glucocorticoids. However, the brain needs a responsive HPA axis to function normally.

The HPA axis follows a diurnal rhythm with high levels in the early morning and very low levels in the evening. This rhythm dictates which glucocorticoid receptors in the brain are occupied: all mineralocorticoid (MR) and some glucocorticoid (GR) in the early morning and primarily MR late in the day. Hypocortisolism, especially in the early morning as is typical in PI youth, may impair cognitive functioning (deKloet, Vreugdenhil, Oitzl, & Joels, 1998). Although in studies of humans, cortisol is the product of the HPA axis that is measured, in animal studies, corticotropin-releasing hormone (CRH) is accessible. Tallie Baram has long argued that many of the neural effects attributed to elevated cortisol are being produced by CRH by brain regions outside of the hypothalamus (Korosi & Baram, 2008). Extra-hypothalamic CRH may increase following chronic stress in concert with a down-regulation of CRH production from the hypothalamus, and through this pathway hypocortisolism may co-exist with hyper-CRH activity in the amygdala and hippocampus (Rosen & Schulkin, 1998).

Changes in cortisol and CRH should affect the hippocampus, prefrontal cortex and amygdala (Korosi & Baram, 2008; Vyas, Mitra, Rao, & Chattarji, 2002). In the case of both the hippocampus and prefrontal cortex, reduction in volume may reflect apoptosis and/or reduction in spine length, number, volume and total number of apical dendrites (see review, (Hodel, 2018). For the amygdala we might expect a dendritic remodeling that increases neuronal activity (Vyas et al., 2002). Notably, in the animal literature, these effects are reversible in adults, but not in the young (see Hodel, 2018). In addition, stress has marked effects on microglia and the integrity of white matter, effects which may be reversible during development. While it is clear that CRH and glucocorticoids have impacts on the developing brain, what is not clear is whether those impacts always produce deficits. There is some evidence in animal models that fear learning may be enhanced in early life stress models (Davis & Burman, 2020). In addition, there is increasing evidence that development of the circuitry of fear learning may be accelerated by early stress (Callaghan & Richardson, 2011; Nieves et al., 2020). In PI children, a developmental acceleration of task-based negative coupling between the amygdala and medial prefrontal cortex has been reported that is protective of the development of anxiety symptoms (Gee et al., 2013). Notably, Gee et al. (2013) found that cortisol mediated the association between early institutional care and negative amygdala-mPFC coupling.

There is recent evidence that down-regulation of the HPA axis may not be permanent. Specifically, in PI children adopted into well-resourced homes, hypo-responsiveness to the Trier Social Stress Test (TSST) was noted among children, but not adolescents (Hostinar, Johnson, & Gunnar, 2015). Studied longitudinally, this pattern was apparent prior to puberty, but as puberty advanced the cortisol response increased until it matched that of youth who had no history of early adverse care (Gunnar et al, 2019). In terms of age, the normalization was obtained around the mid-point in puberty. The normalization of reactivity, however, did not accompany any reduction in externalizing symptoms, and indeed, was associated with an increase in internalizing ones (Perry et al., 2020). )

To our knowledge, there have been only three studies of the sympathetic system following early institutional care. All three studies noted higher sympathetic tone at baseline, despite the fact that one study examined preschool-aged children within their first years post adoption (Esposito et al., 2016), while two examined children in late childhood many years after adoption (Gunnar et al., 2009; McLaughlin et al., 2015). None of these studies noted group differences in parasympathetic activity. Whether sympathetic reactivity is associated with early institutional deprivation is not clear, as one study using the TSST noted a blunted response (McLaughlin et al., 2015), while the other noted no difference in reactivity (Gunnar et al., 2009). It is not clear what role elevated baseline sympathetic tone plays in the effects of early institutional care, but it seems likely that there would be cardiovascular health-related impacts of chronically elevated sympathetic tone.

Epigenetics and Telomere Length.

Epigenetic modification is another likely route through which early adversity affects subsequent development. Another process, telomere shortening, will also be considered in this section because epigenetic modifications shape telomere structure and influence its maintenance (Adwan-Shekhidem & Atzmon, 2018). In addition, telomeres can regulate epigenetic activity affecting gene expression. Thus, it is important to consider them together.

Two decades ago, Michael Meaney and colleagues reported that the effects of early adverse maternal care in rats was mediated by a change in methylation of the glucocorticoid receptor (Weaver et al., 2001). From this finding emerged a body of research showing that early postnatal experiences alter the epigenome in many cells and for many genes. Gene regulation was being adapted to the early postnatal environment presumably to enhance survival (Meaney & Szyf, 2005). The challenge of studying epigenetics in humans is that epigenetic modifications are tissue and cell specific. In humans, the tissues studied are buccal cells and lymphocytes, not brain cells. Nonetheless, there have been studies of PI youth that are suggestive that alterations of the epigenome are an important pathway for early experiences to “get under the skin”.

In the only study we know of that examined epigenetic alterations in infants and toddlers still in institutional care, 29 children in a Russian institution were compared to low-income infants and toddlers living in their families (Naumova et al., 2019). Examining whole genome methylation, a number of significant differences emerged on 172 genes. Gene network analysis and functional enrichment identified genes involved immune system regulation and cellular response to stress. Several of the methylated genes were correlated with the duration of institutional care, including a gene required for organizing myelinated axons during brain development. This last finding is highly reminiscent of evidence of reduced white matter volume in institutionalized children. Importantly, 7–14% of the behavioral deficits observed in these children were predicted by their differences in DNA methylation profiles.

A second study examined 50 adolescents adopted from Russia and Eastern Europe at a median age of 12 months (Esposito, Jones, et al., 2016). Whole genome methylation of peripheral lymphocytes was also examined in this study. These children, however, were compared to never institutionalized youth whose families were of comparable high income and education to adoptive families. The results indicated 30 differentially methylated sites on 19 genes. Functional enrichment analyses indicated the differences clustered into two categories, neuronal (neuron, action potential, and a large regulation cluster) and developmental (organ induction, nephron tubule development, bone morphogenic proteins, etc.).

Finally, a third study used buccal cells to study DNA methylation in young adults adopted between 6 and 43 months from Romanian institution in the early 1990s (R. Kumsta et al., 2016). These individuals had been exposed to incredible deprivation prior to adoption. The sample was small (roughly 16 per group) and none of the 400,000 sites survived false discovery rate adjustments. However, the researchers did note elevated methylation in the promotor region of the CYP2E1 gene that seemed consequential. This gene is involved in breaking down many toxic environmental chemicals and in metabolic functions such as fatty acid oxidation. Elevated methylation of this gene was correlated with impaired social cognition in the PI participants.

While these findings are interesting, it is notable that there has been little to no correspondence across studies in DNA methylation sites. This could be due to differences between studies in the ages of the individuals studied (infants, adolescents) and tissues used (buccal, peripheral lymphocytes). Also, none of these studies reported finding that the GR gene was methylated; the initial finding in the work by Meaney and colleagues.

The association of early institutional deprivation with telomere length has been examined in the BEIP study at multiple points up to age 15 years (Drury et al., 2012; Humphreys et al., 2016). At every point, children who had been in institutional care had shorter telomeres on buccal cells than the never-institutionalized group. Furthermore, examined up to age 8, telomere length was negatively correlate with duration of institutional care. In a third analysis of the BEIP data, Wade and colleagues (Wade et al., 2020) examined the direction of effect between telomere shortening and psychopathology. Counter to expectations it was psychopathology symptoms that predicted telomere shortening in middle childhood and not the reverse. Thus, overall, early deprivation and its duration predicted telomere shortening up to age 8 in the BEIP study, while from then on it seemed that the psychiatric distress these children experienced further accelerated telomere aging. Notably, in a study of young adults adopted early from institutional care, no differences in telomere length were noted in whole blood samples consisting of a mixture of cell types (Elwenspoek, Sias, et al., 2017). The adults in this sample had been stably placed in supportive families from the point of adoption, unlike the situation for many of the foster care youth in the BEIP study.

Immune System: Inflammation and Immune Senescence.

There is increasing interest in whether the impacts of early adversity on physical and affective health are mediated, in part, by adaptations of the immune system to a harsh, early environment. Early life adversity is hypothesized to create a phenotype characterized by inflammation, impaired cellular immunity, and immunosenescence (Elwenspoek, Sias et al, 2017). Much of the focus on early life stress and immunity has been on inflammation and an inflammatory phenotype that is expects to increase the risk of cardiovascular disease and depression (G.E. Miller et al., 2011). However, while studies of adults with adverse early life histories tend to support this hypothesis (Baumeister et al., 2016), findings in children and adolescence have been less consistent (Kuhlman et al., 2019). To date, there is little evidence that PI youth are in a subclinical inflammatory state, as with one exception increased circulating levels of proinflammatory cytokines and C-reactive protein have not been reported (Reid et al., 2019; Slopen et al., 2019). In the one exception, the researchers noted elevated circulating levels of Tumor Necrosis Factor alpha, but not IL-6 or IL-1 or its receptor (Engel et al., 2020).

It has been argued, though, that in children and adolescence the proinflammatory tendency will be more apparent if cells are stimulated with antigens (Slopen et al., 2013). However, for PI youth, stimulating their cells with three commonly used antigens (LPS, PHA, and PWM) has failed in two studies to yield a larger inflammatory response compared to controls (Elwenspoek, Sias, et al., 2017; Engel et al., 2020 ). In one study, stimulation with an antigen more targeted to T-cells (PMA/IO) did reveal evidence of a stronger response in PI than comparison youth (Engel et al., 2020). This latter finding may be because of consistent evidence of an impact of early institutional care on T-cell populations.

Immune compromise, rather than inflammation, was the consistent theme in early animal studies of maternal deprivation and prolonged separation. Even a two-week maternal separation in Pigtail Macaques suppressed T-lymphocyte proliferation to antigen stimulation in adulthood (Laudenslager et al., 1985). Nursery rearing also skewed the CD4+ CD8+ lymphocyte ratios in the direction suggesting immune compromise (Coe et al., 1989; Lewis et al., 2000). Although in PI youth evidence of inflammation has been meager, four studies have reported a lower ratio of CD4+ to CD8+ cells associated with early institutional care, a pattern associated with immune compromise (Elwenspoek, Hengesch, et al., 2017; Esposito, Jones, et al., 2016; Naumova et al., 2019; Reid et al., 2019). In addition, two studies have noted a higher percentage of terminally differentiated T cells with cell surface markers reflecting replicative senescence (Elwenspoek, Sias, et al., 2017; Reid et al., 2019). Both of these studies also reported that PI individuals were more likely to be seropositive for cytomegalovirus (CMV), with evidence that CMV mediated the association between institutional care and immunosenescent T cells. Immune compromise is also consistent with evidence that PI youth are less able to contain the herpes simplex virus, than are comparison youth (Shirtcliff et al., 2009). Thus, immune compromise and early senescence may be a consequence of early institutional viral load. Whether this leads to health consequences among PI individuals remains to be seen. It is the case, though, that neither current activation of the HPA axis, health risk behaviors or immediate health can explain the above findings (Elwenspoek, Hengesch, et al., 2017; Engel et al., 2020 ).

Microbiome.

The gut microbiome plays an important role in educating the immune system (Belkaid & Hand, 2014) and there is increasing evidence that there is a microbiome-immune-brain pathway that influences neurobehavioral development (Callaghan et al., 2020). In non-human primates, early deprivation of maternal care impacts the microbiome (e.g., Dettmer et al., 2019). Two recent studies have examined the microbiome in PI youth (Callaghan et al, 2020; Reid et al., 2020). Both found differences in the diversity of microbiota with higher relative abundance of Prevotella and Bacteroides in PI children (Callaghan et al., 2020) and adolescents (Reid et al, 2020) relative to controls. Relative increases in Prevotella have been associated with chronic infections, while increases in Bacteroides have been associated with prefrontal cortex activation to emotional faces in PI youth (Callaghan et al., 2020). Notably, in PI youth Bacteroides and six other taxa were associated with the prevalence of immune senescent (terminally differentiated) T cells (Reid et al, 2020). Thus, in PI youth there is reason to expect that some of the health and neurobehavioral impacts of early institutional care are mediated through alterations in the gut microbiome that can still be observed years post-adoption, despite the fact that for years the PI youth have eaten a comparable diet to the comparison youth.

Gene by Environment Interaction.

Individual differences in response to early institutional care are remarkable. While nearly every child reared from infancy in institutional care falls behind in developmental milestones, there is striking variability in how children rebound and ultimately thrive or not once placed in supportive families (Woodhouse et al., 2018). Several have argued that there are common gene variants that have survived selection pressure because of their increase susceptibility to both negative and positive environmental input (Ellis et al., 2011). The gene coding brain-derived neurotrophic factor (BDNF) is considered a susceptibility gene, and the polymorphism associated with reduced trophic factor release has been found to correlate with more attention problems than the polymorphism associated with more trophic factor release in children adopted late from institutional care, but with fewer attention problems from those adopted early (Gunnar et al., 2012). PI youth who experience peer victimization are at risk for depression. FKBP5, another potential susceptibility gene, moderated the relation between peer victimization and depression for PI girls, increasing susceptibility at higher but reducing it at lower levels of victimization (VanZomeren-Dohm et al., 2015). ADHD z-scores for PI Romanian youth revealed significantly higher scores for those with the risk haplotype for the dopamine transporter gene 1 (DAT1), than for those without this genotype (Stevens et al., 2009). Thus, not surprisingly, genotype likely moderates the impact of early institutional deprivation. The evidence so far, however, is simply proof of concept and does not provide a comprehensive picture of gene by environment interaction in regulating how early deprivation affects neurobehavioral development. It is also worth noting that in all of the studies of GXE for early institutional deprivation, the impact of early institutional rearing is observed even for the genotype group considered low risk.

Summary of Neurobiological Pathways.

With the exception of stress, the many pathways through which early institutional deprivation can impact brain and behavioral development are just beginning to be explored. PI youth provide an important model for this work as they help pinpoint early periods prior to adoption as critical for shaping processes that organize neurobehavioral development. Furthermore, these data make it clear that our earliest experiences are woven into the fabric of our being from the bacteria we carry in our guts to the methyl marks on our DNA to the population of T cells we have available to fight infection and the integrity of our neural systems. Genes likely affect how sensitive we are to the early environment, but so far, the genetic variants studied appear to reduce, but not eliminating the risk of deprived early care.

Practice & Policy

What we have learned in recent neurobiological and behavioral studies of PI children and youth confirms and extends what was already known: specifically, institutional care is a poor option for infants and young children. Ethically, we need to justify continued study of the effects of early institutional care divorced from concerted efforts to prevent or improve the care for children without permanent parents. This is especially the case for studies involving random assignment to care-as-usual compared to study attempts to improve children’s care. Importantly, researchers involved in these studies are making concerted efforts to change practice and policy (see Goldman et al., 2020). Indeed, there are a number of practice and policy recommendations that can be implemented in the short and long term in order to improve conditions for children who might otherwise end up in institutional care.

Multi-system Approach to Prevention of Family Separation

As outlined in the United Nations General Assembly 2019 Resolution on the Rights of the Child, children have the right to grow up in a family (Goldman et al., 2020). Because there are often circumstances other than the death of parents that cause children to be given up—personal, political or economic situations (Waddoups et al., 2019)— steps should be taken to prevent children from being separated from their families in the first place. Preventing children from losing their families requires healthcare, education access, housing, financial aid, social support, and food support to work together. Beneficial processes include everything from making birth control available to reduce births of children the family cannot care for to providing parents with the support and resources they need to support all their children. While such large, systemic changes may take many years to implement, as countries work on these changes, there are a number of individual, local, and national initiatives that may improve institutional care and produce better developmental outcomes for children.

Stop Unnecessary Volunteering

Over the years, tourist volunteering to provide care for children in institutions has become popular and is certainly well-intentioned, but all too often results in iatrogenic effects ( Zeanah et al., 2019). Volunteers generally lack the specialized training necessary to work with at-risk populations. They are not certified or trained in dealing with victims of child abuse or sex trafficking. Volunteers also come and go at their own convenience, resulting in unstable relationships and care. The popularity of orphanage volunteering has turned it into an industry involving millions of people every year (Goldman et al., 2020). In order to make money from the industry and attract tourist volunteers, low-quality institutions are created, and families are often deceived into sending their children to these institutions. It is important that individuals, organizations, and governments stop the promotion of this kind of volunteering.

Deinstitutionalization

When separation of parents and children is unavoidable, the goal should be deinstitutionalization, and employing alternatives such as family-based foster care. Quality foster care allows children to receive the kind of individualized attention that enables them to thrive. While deinstitutionalization and switching to more family-based care should be the ultimate goal, achieving it can be quite difficult as existing policies and institutions for the care of infants and children are long-established and integrated into the community, meaning many local actors maybe dependent on them for their livelihoods.

Improving Current Institutional Care

An initial first step can be to work with the existing infrastructure and revamp programs within the institutions. Promoting practices and policies such as lower child-to-caregiver ratios and adjusting schedules can provide children with more individualized consistent care. In addition, adequate training for institution staff in best practices for early childhood care and development, with a focus on attachment, is essential in order for children to get the sort of quality simulating care they need in order to thrive (Groark et al., 2005). An example of systemic transformation in the quality of institutional care on a national scale through a training of trainer’s model has been underway in China over the past 20 years. The Chinese government, in partnership with U.S. NGO, OneSky for all children and its local Chinese subsidiary, Chunhui Children, developed programs in Chinese orphanages wherein they hire local women and provide them with training on researched based best practices focused on responsive care https://onesky.org/what-we-do/in-china/. Depending on the age group the caregiver will be assigned a certain number of children whom she will care for, teach, and play with every day. As each child grows up that caregiver helps them transition to preschool programs, youth programs, or a home-based foster family. All teachers, mentors and foster parents are also trained to provide responsive care, thus, care is consistent across caregivers. Critically, every child receives one-on-one responsive interactions with a consistent adult, allowing the formation of attachment relationships. Furthermore, transitions are carefully planned and the child has someone they can rely on as they move to different programs and new caregivers.

Assessment and Accountability

Whether in institutions or, preferably, in family settings, programs for children without permanent parents need measurement and evaluation in order to hold them accountable. Consistent organized evaluations of programs are needed to see program weaknesses and strengths and also maintain the quality of the program over time. Social context, economy, natural disasters, and many other factors change over time and can alter the needs of the population. Programs must be able to adapt to the cultural and social context of the time. Not only can measurement and evaluation provide essential feedback for program development and maintenance but also inform policy and new practices on a larger scale. Demonstration with data to back up the positive impacts of a program can be very informative for governments and NGOs when developing programs of care for children separated from their families.

Conclusions and Future Directions

Overall, the research studies on institutionalized and previously institutionalized children have provided a wealth of information about the importance of our earliest experiences. They have also provided tools with which the impact of practices and programs to improve outcomes for children who experience periods of early deprived care. These findings also have important implications for practice and policy. Clearly, policies that effectively prevent family breakdown and children that families cannot care for should be the highest priority, with the development of the most family-like arrangement supportable in the context as the alternative when children cannot be cared for by their parents.

Highlights.

• Review of the long-term effects of early deprivation of institutional care on child development and capacity to recover from this type of early life adversity.

• Examination of the behavioral and neural evidence for altered executive function, declarative memory, affective disorders, reward processing, reactivity to threat, risk-taking, and sensation-seeking in post-institutionalized children.

• Policy and practice recommendations that can be implemented to improve conditions for children in institutional care.