The developmental support hypothesis: adaptive plasticity in neural development in response to cues of social support

Abstract

Across mammals, cues of developmental support, such as touching, licking or attentiveness, stimulate neural development, behavioural exploration and even overall body growth. Why should such fitness-related traits be so sensitive to developmental conditions? Here, we review what we term the ‘developmental support hypothesis’, a potential adaptive explanation of this plasticity. Neural development can be a costly process, in terms of time, energy and exposure. However, environmental variability may sometimes compromise parental care during this costly developmental period. We propose this environmental variation has led to the evolution of adaptive plasticity of neural and behavioural development in response to cues of developmental support, where neural development is stimulated in conditions that support associated costs. When parental care is compromised, offspring grow less and adopt a more resilient and stress-responsive strategy, improving their chances of survival in difficult conditions, similar to existing ideas on the adaptive value of early-life programming of stress. The developmental support hypothesis suggests new research directions, such as testing the adaptive value of reduced neural growth and metabolism in stressful conditions, and expanding the range of potential cues animals may attend to as indicators of developmental support. Considering evolutionary and ecologically appropriate cues of social support also has implications for promoting healthy neural development in humans.

This article is part of the theme issue ‘Life history and learning: how childhood, caregiving and old age shape cognition and culture in humans and other animals’.

1. Introduction

Across animals, there is evidence that social interactions early in life stimulate brain development. For instance, touch plays a key role in neural growth and brain development across mammals (reviewed in [1–3]). Why should development of the brain be dependent on social input, given the importance of neural investment for fitness [4,5]? There are a number of explanations for the observation that neural development depends on social interactions. First, many brain regions are wired through activity-dependent processes, and social interaction clearly plays a role in directing the development of corresponding brain regions [6–8]. For example, touch and social cues direct the development of corresponding brain regions [7–9] and infants deprived of social touch have problems with sensory processing [10–12]. Similarly, social interactions play a role in the development of social cognition [13,14] and regions of the brain involved in touch [15]. However, as we review here, social interaction more broadly affects neural growth and metabolism throughout the brain [16–18], as well as stimulating growth and development more generally [19,20]. The effects of social interactions go beyond just directing neural development through sensory experience and activity-dependent processes.

In this review, we focus on the idea that the social dependence of neural development could have arisen in part as adaptive developmental plasticity where organisms adjust their neural development to match anticipated levels of developmental support (figure 1). In other words, this flexibility in neural development is a ‘predictive adaptive response’ [21–23] where cues of social support predict the developmental environment. We contrast this adaptive explanation with a non-adaptive explanation (figure 1b), that in conditions of compromised social support, decreases in neural growth represent an offspring stress response where individuals are making the ‘best of a bad situation’ [24,25]. Indeed, extreme disruptions of social interaction during development, as in historical orphanages, result in truly maladaptive disruptions to cognitive development [1,10,11]. In reality, there is no doubt that all these explanations of socially induced neural plasticity are true, but evaluating hypotheses around adaptive plasticity offers new perspectives on existing literature, novel routes for interventions to improve neural health and opportunities to integrate multiple explanations.

Figure 1.

Plasticity in neural development with social support: adaptive or not? Variation in climate, predation, nutrition and other factors results in variation in parental care of offspring within and between species. (a) Cues of such developmental support are known to trigger variation in neural and cognitive development: individuals receiving more touch or visual attention from parents experience greater neural and cognitive development and associated changes in behavioural exploration, learning, body growth and stress responsiveness. Is this plasticity adaptive? (b) The developmental support hypothesis posits that cognitive plasticity in response to cues of developmental support are adaptive (e.g. a predictive adaptive response). This would be measured as relatively higher performance in the ‘high support’ environment of individuals exposed to cues of high support and relatively higher performance in the ‘low support’ environment of individuals exposed to cues of low support, in other words, a phenotype–environment matching. (c) Non-adaptive explanations of neural plasticity (in (a)) posit that ‘low support environments' are simply stressful and/or early experiences in high-quality environments have a lifelong benefits (silver spoon effect). These explanations predict that both early life and current experiences in ‘high support’ environments will have benefits which may or may not be additive.

This review focuses on the hypothesis that the dependency of brain development on social cues could have originated as adaptive plasticity, what we term the ‘developmental support hypothesis’. We first outline the necessary conditions that favour the evolution of adaptive developmental plasticity. We then review the evidence that the development of large brains and cognition is incredibly costly, in terms of time, energy and exposure [26]. If the ability of parents to provide care during this time is compromised, there could be benefits to offspring investing less in neural development, instead being more independent. This review explores the components of this argument, contrasting it with existing hypotheses and finally discussing the implications of such a perspective for future studies and promoting cognitive health in humans.

2. The evolution of adaptive developmental plasticity

Variable environments tend to select for plasticity in development, especially when different traits are favoured in different environments and there are reliable cues [27,28]. For example, chemical defences protect plants from herbivores, but because they are costly to produce, most plants have evolved the ability to induce the production of such chemicals only in the presence of predator cues [29]. Why would plasticity in neural development arise, given the benefits of neural investment and cognition [4,5]? There is no doubt of maladaptive plasticity in conditions of reduced parental care, for instance, smaller body size and reduced adult cognition following early-life nutritional stress [30,31]. Is it possible that some of the variation in neural development represents adaptive plasticity in response to developmental environment (figure 1)? We argue here that environments with compromised developmental support favour a plastic reduction in neural investment because neural development is a costly process in terms of time, energy and risk. We review factors that favour the evolution of plasticity and discuss their relevance with respect to this ‘developmental support hypothesis’.

(a). Neural development is costly

Developing a large brain and associated cognition is incredibly costly for individuals in terms of time, energy and exposure. In terms of development time, the total duration of brain development scales exponentially with relative brain size across species [32]. For large-brained species, much of this development necessarily occurs after birth, which may come with additional costs of offspring feeding and protection. Even with adaptations to the pelvis for large brains [33,34], most of human brain growth happens after birth, with volume increasing threefold from birth to adulthood [35]. Across mammals and birds, large-brained species have extended developmental periods [36,37].

Building a large brain takes time owing to processes playing out at both the neural and behavioural levels. At the neural level, building a brain with trillions of functional neural connections requires juvenile periods with active wiring of the entire brain [38–40]. In humans, synaptic density of brain regions involved in attention peak at 1 year and do not reach levels comparable to adults until 16 years [41,42]. In other mammals, there are analogous peaks in synaptic density three to six months after birth, with densities much higher than in adults, reflecting periods of active synaptic pruning [43–45]. Post-natal periods are associated with a refinement of connections through apoptosis, with up to 30% of neurons lost in the developing cortex of the mouse [46,47], and 15% loss in monkeys [43]. At the behavioural level, during juvenile periods, large-brained species are immersed in learning. Humans and other primates learn through watching and listening, copying actions and sounds made by other individuals, through trial-and-error and active play [48,49]. Skilled foraging behaviours in birds, such as tool use in some crows, can take months to years to develop [50]. In humans, the development of expert-level performance on a task, such as surgery, large game hunting or concert piano playing, can take years to decades to develop [51,52]. Species with more post-natal brain development engage in more play behaviour, presumably as an important part of brain development [53]. Building a large brain with billions of functional connections takes time.

Neural development is costly not only in terms of time, but also energy and trade-offs with attention. Metabolic demands of the brain peak at approximately 2–8 years in humans, corresponding to time of active neural pruning and learning [54,55]. Attention trades off with neural development in part owing to the need for offline processing: sleep is necessary for processing information and rewiring the brain without active sensory input, and need for sleep is much higher during periods of active neural plasticity in infancy [56–58]. At the behavioural level, as individuals focus their attention on play or learning, there are trade-offs with attention to risks and dangers [59,60]. Focus on cognitively demanding tasks takes attention away from other tasks such as predator detection, increasing risk [60,61].

Across these examples, we see that neural investment and behavioural development is associated with a range of developmental costs, from long development times [32,62] to periods of active exploration and neural plasticity that trade off with other tasks (e.g. sleep and foraging [63,64]). These developmental costs result in the increased dependency of relatively large-brained species on parental investment [65–68]. Similar trade-offs in parental investment and neural plasticity are even seen in insects, paralleling patterns see across mammals [69,70]. These patterns overall support the idea that neural development is costly and such costs play out in patterns both within and across species.

(b). Environmental variation can compromise parental care

The evolution of developmental plasticity is favoured in environments that vary in time or space [71–73], but where the rate of environmental change is moderate enough that developmental adjustments can keep up [74,75]. As we have just reviewed, neural development can be incredibly costly in terms of time, energy and risk; to survive this period, individuals tend to rely on increased investment from parents and other relatives. However, the capacity for parents to invest in offspring care varies widely within species. Environmental variation in food, temperature and other factors can influence whether parents should invest in care of current offspring or save efforts for future opportunities [76–78]. Theory suggests that, as environments are improving predictably, parents should provide increasing care to successive offspring [76], as seen in owls that adjust parental investment with year-to-year changes in prey abundance [79]. Birds also shift care intensity, such as nest visitation rates, as predation pressure varies [80,81]. In many birds and mammals, parental care varies with age, tending to increase over time to a maximum level at middle age, probably owing to combined effects of improvements with experience, changes in value of current offspring and later-life senescence (e.g. [82,83]). In species where parents do not directly care for offspring, variation in parental investment occurs through variation in resources deposited in eggs or placement of eggs in high-quality locations (e.g. [84,85]). Across all of these examples, we see that the ability of parents to invest in offspring varies. This variability in parental care within species sets up conditions that favour the evolution of plasticity in neural development: offspring should adjust costly neural investment depending on the availability of developmental support.

(c). Cues of developmental support as triggers of neural development

Plasticity is more likely to evolve when there are cues that reliably indicate the future environment [86–88]. In the case of the developmental support hypothesis, the future environment is not adulthood, but the environment in which neural development occurs, which may be weeks to decades, depending on the species. What are cues of future developmental support? Here, we focus on many aspects of parental behaviour that go beyond the basic needs of offspring, as potential cues of future developmental support. For instance, many instances of grooming, licking and touch behaviours go beyond the needs of the offspring in terms of thermoregulation or parasite removal [89]. While we primarily focus on non-nutritive cues (touch, visual attention, etc.), one could even argue that parent feeding of offspring results in not only a current nutritive benefit, but also acts as a cue of future feeding, similar to the argument that courtship feeding not only has direct benefits, but also acts as a signal of paternal care abilities [90,91].

Across animals, cues of parental investment and social support have broad effects on brain growth and behaviours that support brain development, such as exploration. In rodents, maternal licking affects the expression of genes involved in neural growth, synapse survival and metabolic support in multiple regions throughout the brain [92–95], resulting in higher learning and memory in a range of spatial and recognition tasks [94]. Maternal separation reduces the expression of neural growth factors and neurogenesis in multiple regions throughout the brain [16–18], resulting in decreased neural growth and increased cell death [96] in the brain (e.g. brainstem, cerebellum, cortex), reduced cognition [97] and even smaller organs, such as the heart and liver [20]. A recent meta-analysis of almost 100 rodent studies shows consistent negative effects of maternal separation on memory, nerve growth factors, body weight and growth [98]. In human infants, caressing touch stimulates brain regions involved in memory, language and cognitive function, not just those involved in processing mechanosensory input [99,100]. The stimulating effects of touch on brain development can even correct for experimentally induced cortex lesions in rats [101].

Touch and other cues of social support not only affect brain development, but also behaviours normally associated with brain development, namely, exploration and play. Maternal touch in rats is linked to offspring's exploration of novel environments and novel foods [102–104] and active play behaviour [105]. In rodents, effects of maternal separation on exploration and novelty seeking can be rescued by experimental touch [106,107] or cross-fostering to mothers with more social interaction [103]. Across rodents, primates and even pigs, early-life touch is associated with later exploration, and approach of novel spaces and objects [14,102,103,108]. Similar patterns are seen in humans where maternal attentiveness predicts later exploratory behaviour in infants [109–111]. Widespread links between touch and behavioural plasticity is consistent with the idea that cues of parental or social support signal to offspring that there is a ‘safe space’ for active exploration, neural plasticity and learning, recalling discussions in the child development literature on secure mother–child attachment and resulting play and exploration [112,113].

While much of the existing research has focused on maternal touch, touch from conspecifics or interaction with other individuals in a social group may also act as cues of relevant social support. Being in a group greatly reduces risk [114], and in many species, larger social groups may care for related or even unrelated offspring [115,116]. The importance of broader social support could explain why being reared in a larger litter can rescue rat neural development from deprivation of maternal touch [117] and increased exposure to peers can enhance effects on neural growth factors [118]. Similarly, rearing in ‘enriched environments' where rodents interact with rat pups from other families (in addition to having access to more space, running wheels and other objects) stimulates insulin-like growth factor expression and the enhanced development of several brain regions, even the visual cortex [119,120]. As discussed below, these ideas have further implications in humans, with their extensive social networks involved in child rearing [121,122].

(d). Cues of social support as triggers of growth and stress responsiveness

Cues of social support affect the development of not only the brain and exploratory behaviour, but also aspects of overall growth and stress responsiveness. Maternal touch and interaction tends to reduce later-life stress responses, while individuals that were deprived of cues are more stress reactive (e.g. [17,93,102,103,123]; reviewed in [124,125]). Indeed, massage therapy in infants may reduce stress and enhance cognitive responses through similar mechanisms [126,127]. Plasticity in stress responsiveness may be an adaptive response to variation in an organism's physical and social environment, including the degree of parental investment in offspring, termed the ‘adaptive calibration model’ [128]. Individuals that perceive an environment with less social support may need to be more self-dependent; in this case, having a relatively more active stress response could be adaptive [129]. Indeed, rodents experimentally deprived of maternal touch are relatively better learners under stressful conditions later in life [130,131]. The adaptive calibration model is similar to the developmental support hypothesis because it speaks to adaptive plasticity in compromised developmental environments. However, the adaptive calibration model focuses on stress responses, whereas the developmental support hypothesis focuses on neural and behavioural plasticity, but also includes stress reactivity as part of an alternative strategy in environments with compromised parental care.

Social support cues also stimulate overall growth and immune development, suggesting they may tie into plasticity in life-history strategies (i.e. patterns of developmental transitions and offspring investment). Maternal touch directly triggers the expression of enzymes and hormones that regulate cell growth in the brain and major organs (reviewed in [19,20]). Maternal separation per se (not associated changes in nutrition) results in tissue insensitivity to growth hormone [132] and reduction in body size [106]. Increases in growth rate are also seen as a result of therapeutic touch applied to pre-term humans in intensive care units [133]. These changes in growth could be an adaptive mechanism of reducing overall size and associated developmental demands, in the face of less available developmental support [133]. Others have argued individuals in environments with less developmental support shift towards a faster life history, where less energy is invested in self-maintenance and more energy is shifted towards earlier life reproduction [128,134]. In support of this idea, infant rodents and primates that experience less touch have lower investment in immunity [135,136].

Links between social interactions and growth, stress responsiveness and life-history strategy may be an adaptive response to variation in the social environment, and the degree of parental care [128]. If parents are less able to care owing to a low-resource or high-predation environment, it would be beneficial for offspring to be more sensitive to stress, and reduce not only brain growth, but overall body growth, and shifting to a faster life history [128]. Differences initially induced by variation in parental care may also be transmitted to the next generation as stress reactive individuals can be less responsive parents [103,106]. These ideas are similar to broader discussions about ‘adaptive programming’ of stress responses [129,137,138].

3. Predictions and implications of the developmental support hypothesis

Here, we have discussed what we term the ‘developmental support’ hypothesis, the idea that cues of social support early in life can stimulate brain development and associated exploratory behaviours involved in neural development (figure 1). In environments where parental care is compromised, offspring invest less in neural development and develop into smaller, more stress reactive individuals. The developmental support hypothesis builds on existing ideas and is consistent with the wide literature on neural development in animals (especially mammals). However, this interpretation of existing data suggests a variety of interesting avenues for additional research, as we outline here.

(a). Neural plasticity in response to cues of developmental support: adaptive plasticity or maladaptive stress response?

There are two broad categories of explanations to explain variation in fitness-related traits in response to early developmental conditions. Adaptive plasticity hypotheses argue that changes in development have evolved as a result of environmental variation and represent a strategy for matching one's phenotype to environmental conditions (figure 1b), as we have primarily discussed here. Non-adaptive explanations focus on the idea that some environments are simply ‘bad’ and result in lower fitness (figure 1c). These explanations show up in the literature on stress responses, sometimes described as ‘making the best of a bad lot’ [24,25], the ‘silver spoon effect’ [139,140] or ‘deficit models’ [141]. Adaptive explanations can be found in the literature on adaptive developmental plasticity and ‘predictive adaptive responses' [21,22]. There are many examples of ‘silver spoon effects’ [139,140], most telling from reciprocal experiments that can tease apart both hypotheses (figure 1; [142,143]). There is no doubt that as the environment worsens, there is some point when adaptive plastic responses begin to break down and stress effects kick in. However, the preponderance of examples of adaptive plasticity [144] suggests that there is at least some range of environmental variation where adaptive developmental plasticity allows organisms to buffer potentially poor situations (e.g. [145–148]).

In explaining neural plasticity in different social environments, how would one tease apart the developmental support hypothesis and stress explanations? The developmental support hypothesis falls into the category of a predictive adaptive response [21–23], where the cues are predictive of the developmental environment, but not necessarily the adult environment. The developmental support hypothesis predicts that individuals growing in situations of reduced developmental support should have higher fitness in these environments than those initially exposed to a high-quality environment and then switched to a low-quality developmental environment (figure 1b). For instance, we would predict that reduced neural growth in maternally separated rats would be associated with a decrease in associated neural costs, such as a decline in brain metabolic rate and the need for long periods of sleep early in development. Similarly, the developmental support hypothesis predicts that smaller, highly stress-responsive offspring are more likely to survive in conditions where parental care is compromised and environments are inherently more stressful. There are a handful of studies showing adaptive changes in learning ability and neural development in response to early-life stress [130,149]. Indeed, as predicted by the developmental support hypothesis, the development of neural connectivity seems to speed up in humans that experienced early-life stress [149] potentially leading to offspring independence earlier in life. However, we are very much at the beginning of considering the suite of adaptive responses in environments with comprised developmental support [129,138,141].

In many ways, the developmental support hypothesis lies at the intersection of both adaptive and non-adaptive explanations of how the developmental environment shapes fitness. Adjusting neural development in conditions of low social support may be adaptive in terms of surviving early development (adaptive plasticity or predictive adaptive response), but maladaptive in terms of adult fitness or performance (representing a silver spoon effect). This may be especially pronounced in long-lived species like primates where evidence for predictive adaptive responses to poor developmental conditions are mixed [150–153]. It is also possible that neural plasticity in response to the social environment represents a predictive adaptive response to a future individual state, rather than future external environment (see [23]). In this case, it could be beneficial for individuals with less energy reserves to have lower neural investment to avoid metabolic costs of a large brain later in life. Studies could tease these explanations apart by switching individuals between environments at different life stages.

(b). Expanding the range of systems: comparative and experimental approaches to studying adaptations to variation in parental care

The developmental support hypothesis suggests new systems for study with respect to neural plasticity. If variation in parental investment drives flexibility in brain growth and learning, then species that encounter greater variation in environmental threats to parental care may harbour particularly interesting adaptations. In birds, year-to-year climatic variation is associated with cooperative breeding, suggestive of the importance of larger social groups in buffering variation in parental investment [154–156]. Fishes are another promising system for comparisons across species as they have pronounced variation in parental care across sexes [157] and in response to external factors such as predation [158]. There are also a number of insect systems with flexibility in parental care, offering additional situations for experimental manipulation. For instance, burying beetles directly feed offspring, but the degree of care varies with season, and interactions with siblings and mites [159–161]. While existing research has focused on mammals (especially rodents), studying a greater diversity of species may uncover unique adaptations in certain species or broad-scale patterns across species.

It is likely that, as we consider a greater range of species and environments, we will begin to see significant variation in the degree of neural plasticity in response to cues of developmental support. Much existing literature focuses on the fact that there is plasticity, but the broader literature on plasticity general reveals substantial variation within and between species in the degree of plasticity [28,162,163]. We can quantify this variation in plasticity through reaction norms (figure 2a) and use it to measure associated costs and benefits of plasticity ([164]; figure 2b). Are there some genotypes or species with neural development that is robust to perturbations in the social environment? What are the underlying proximate or ultimate mechanisms? For instance, there is a literature which suggests that long lifespans and variable environments favour the evolution of large brains and behavioural plasticity [165,166]—given the costs of large brains and variability of parental care in variable environments discussed above, this could create a positive correlation across species between brain size and neural plasticity with cues of developmental support. Taking a reaction norm represents a powerful toolkit for considering plasticity in an ecological and evolutionary framework [28].

Figure 2.

Variation in the degree of neural plasticity in response to cues of developmental support. (a) Most studies to date have focused on the fact that there is plasticity in cognitive development in response to cues of social support (figure 1a); however, it is likely that this degree of plasticity varies. The dots and lines show a genotype, population or species reaction norm: the dotted lines indicate a less plastic genotype and the solid lines indicate a more responsive or more plastic genotype. (b) Variation in the degree of plasticity can be used to measure associated costs and benefits of this plastic response. For example, contrasting the performance of genotypes from (a) in the high environment would reveal a cost of the ability to be plastic—does the more plastic genotype suffer a fitness trade-off when in the high-quality environment relative to the less plastic genotype that is expressing the equivalent phenotype? Contrasting the performance of these genotypes in the low environment would reveal a cost of mismatch (or a benefit of plasticity): does being unresponsive to the environment result in a phenotype poorly matched to those conditions?

(c). The range of cues indicative of developmental support and implications for human health

Probably, the most exciting future direction suggested by the developmental support hypothesis is the possibly of expanding the range of cues we consider as reliable signals of developmental support. Much of the existing research on the importance of social support cues focuses on touch [1–3,167] and interactions between offspring and mothers [98,133,168]. However, the developmental support hypothesis suggests that there may be a wide range of other potential cues that could act as signals of developmental support. In humans, parental attentiveness and responsiveness, through visual, verbal or physical contact, is important in cognitive development [169,170]. Interactions between parents and children are key to language development and school outcomes [171–175]. The present review suggests that such parental attention may be acting as a cue of social support that is promoting general brain development, not just learning of specific things. Later in development, the types of relevant cues may shift—tickling [176], grooming [89], hugging [177] and play may become more important signals of social support than direct maternal touch. In non-human animals, cues other than touch may act as important indicators of developmental support. In some birds, exposure to conspecific song stimulates the expression of genes upstream of neural growth factors in brain regions involved in song learning [178], but it is possible that exposure to paternal song could be acting as a more general cue of developmental support influencing growth and neural development throughout the brain [179]. This discussion also suggests that mechanisms of social and affiliative behaviour, such as oxytocin signalling [180,181], could have implications for understanding plasticity in neural development.

It is important to consider not only the types of cues we are considering, but where they are coming from. In much of the human research, the literature often focuses on the critical role of mothers in development, from the role of maternal touch on neural development to the importance of maternal mental health and attentiveness on cognitive development [1–3,168–170,182]. However, we also know that cues of social support from other individuals in a social group are important in neural development [117–120]. Indeed, developmental support or alloparenting from individuals other than parents occurs, is critical in humans [122,183,184] and found in hundreds of other mammals and birds [185]. Human carer networks include family members in extended families [186], same-gender parents in primary parenting roles [187], children as carers [188], kin created from friend and neighbourhood networks who provide critical support [189] and hired carers including nannies and teachers [190].

The developmental support hypothesis may suggest new approaches or interventions for improving human neural development. In humans, socioeconomic conditions can create novel environments where cues of developmental support are compromised [190]. For instance, mothers balancing multiple jobs to support their family are at risk of depression, and reduced attentiveness in depressed mothers can have negative impacts on cognitive development of their children [191,192]. Ideally, longer-term policy interventions would alleviate the causal socioeconomic conditions [193,194], but how could more immediate interventions provide salient cues of developmental support? The present discussion suggests a few strategies. First, we know direct physical interaction from a range of carers is important. Such interactions can come from extended families [122,183,184], or from carers at daycares and schools. However, while touch is emphasized as an important cue in the literature (especially early in development), the social appropriateness of touch varies widely across context [195,196]. Cultures also vary in the way that carers interact with children and the degree of physical interaction [197–199]. Given the wide range of cues of social support discussed above, it is likely that there is no one best way to indicate support and that care from a range of individuals is important. This also shifts the discussion away from the mother as responsible for child outcomes [191] towards considering the entire child environment, from preschool to community and kin social networks [200].

When human interactions are compromised, we often look for appropriate substitutes. The developmental support hypothesis enables us to evaluate substitutes as worrisome or acceptable because of their capacity to emulate relevant and salient social cues. Which cues are particularly important for humans? Touch is particularly important, especially early in development [1–3,167]. However, direct visual gaze and face-to-face interactions are also incredibly important in humans and other highly social species [201–204]. Noting these salient cue calls raises alarms for increasing use of screen time, both in entertainment of young children [205–207] and in distracting parents and reducing direct visual interaction [208,209]. Interestingly, there are efforts to use virtual reality to augment electronic experiences with touch [210,211]; however, whether this will act as a true substitute for touch remains to be seen.

Are there more promising substitutes for comprised developmental support? Social animals used as companions may provide many of the same relevant social support cues—direct visual contact, attention, social exchanges of information, touch and play behaviours. There is a large and long-standing literature supporting the importance of human–animal interactions in reducing stress and improving health outcomes [212,213]. Such animal interactions are also important in child development, impacting social and emotional well being in addition to more general cognitive development and school performance [214–216]. A range of hypotheses and literatures can explain why such child–animal interactions are so impactful [217,218]. The developmental support hypothesis points to the importance of social cues of developmental support, which could explain why dogs are particularly effective therapy animals [213] with their evolutionary history of social interactions. This reasoning also points to why other social animals have been successfully used in therapy (dolphins, elephants, parrots, [219–222]) and reinforces non-human therapies (e.g. robots) that incorporate cues of real social animals [223]. This perspective offers interesting lines of future research with pets as facilitators of cognitive development. For instance, are pet interactions very early in childhood, during periods of more extreme neural plasticity (see above), more impactful in terms of effects on cognitive development?

Despite the promise of interventions that provide appropriate social cues, when considering human cognitive health, we must not forget the importance of broader structural interventions. What are the socioeconomic drivers of compromised parental support in the first place? Supporting cognitive health of children is important, but we must also seek solutions that improve maternal mental health, increasing their long-term impact [191]. Interventions that target the broader environment shaping social interaction have been more efficacious. These programmes have provided comprehensive links to human services, not just instruction on social interaction, and produce better outcomes for family stability by strengthening support for mothers as well as children [191]. Other studies show that educational preschool—a structured, supported environment with multiple carers—can produce some of the same positive outcomes in education and job attainment as interventions that focus on strengthening maternal interaction with children [200]. These studies indicate the effectiveness of a multi-level approach to facilitating social cues through a structural intervention on the environment. This includes strengthening family dynamics (e.g. parenting skills; relationship support), but also increasing access to counselling and mental health services, improving the supportiveness of schools, creating programmes in the community and improving socioeconomic conditions (housing, security, inequalities) [193].

(d). Building underlying theory of integrated plastic responses to suites of cues of developmental support

The literature on the evolution of phenotypic plasticity has a rich theoretical history (reviewed in [162,224]) and expectations derived from modelling have greatly enriched empirical investigations (e.g. [225]). However, many of the ideas reviewed here represent verbal models that deviate from existing theory in important ways, arguing that development of relevant theory is imperative. First, classic theory in plasticity considers responses to single environments variables and cues of environmental states. However, organisms are confronted with a plethora of cues of environmental state [226]—indeed, we argue in the above section that we should consider an even broader range of cues of developmental support. Models of plasticity that incorporate multiple cues produce varying results, sometimes suggesting organisms should attend to only the most reliable cues [227], and others which suggest organisms should use a range of cues to improve predictions of optimal response [228]. Variation in the availability and relevance of cues over developmental time can further shape the evolution of developmental sensitive periods (reviewed in [229]). Distinguishing between social and non-social information and incorporating more realistic data represent important frontiers in developing theory on the range of cues that favour the evolution of adaptive plasticity [230].

Second, much of the classic plasticity theory considers plastic variation in single traits, but organismal responses are integrated across many traits. In the meantime, there have been many verbal models of covariation in suites of coordinated traits (e.g. pace of life syndromes, behaviour–cognition syndromes [231,232]), some of which have produced mixed empirical support [233]. Indeed, discussions above argue that a number of behavioural, neural and physiological traits could shift in an adaptive way in response to early-life cues of developmental support. We need development of more models of correlated plastic responses, especially because existing theory suggests that often assumed trade-offs are not necessary to generate some of the suites of correlations between life history and behaviour (reviewed in [234]). In addition, many of the hypotheses around variation in suites of correlated traits comes from extrapolating interspecific patterns to explain intraspecific variation, and making such connections can be fraught from a theoretical perspective [235,236]. One way forward includes more developmentally realistic models of plasticity that incorporate multiple traits that interact throughout development (e.g. [237,238]). Building models forces precise articulation of hypotheses and assumptions and generates testable predictions, leading to empirical data, which can feed back to challenge and further develop the theory.

4. Conclusion

In this review, we have asked why neural development is sensitive to variation in touch and other cues of social support—why should a trait as important as neural investment be dependent on such external factors? Experimental work suggests that it is not a matter of nutrition, as effects of separation on brain development can be rescued with touch. While we know touch affects the development of specific brain regions through activity-dependent processes, interactions also have much broader effects on neural development, exploration, stress reactivity and the growth of the overall body size. We argue that environmental variation has shaped the evolution of neural development to be sensitive to cues of developmental support, given the extreme costs of brain development. This perspective implies a range of other cues may be important, and possibly redundant, effectors of neural development and that research to date has not fully considered the potential adaptive value of how development proceeds when social support cues are reduced. We have attempted to integrate literature from neurobiology, evolutionary and developmental biology, with anthropology, social science and child development to gain insights on the evolutionary origins of these patterns, along with implications for supporting and understanding human health. Such an evolutionary and ecological perspective re-shapes some of our conversations, stressing the importance of environmental variation in driving parental behaviour, and highlighting how a range of salient cues and social networks of individuals are important in neural development.

Data accessibility

This article has no additional data.

Authors' contributions

E.S.-R. drafted sections on ecology/evolution and C.S.-R. drafted sections on public health/anthropological perspectives. Both E.S.-R. and C.S.-R. edited and revised the entire paper.

Competing interests

We declare we have no competing interests.

Funding

We received no funding for this study.