Appetite Self-Regulation in Childhood: A Narrative Review and Conceptual Model of Processes and Mechanisms With Implications for Research and Practice

Abstract

The purpose of the present narrative review was to propose a unifying generalized conceptual model of mechanisms and processes in appetite self-regulation (ASR) in childhood. Appetite self-regulation, along with other domains of self-regulation, develops across childhood and contributes to energy intake and balance, diet quality, weight, and therefore long-term health outcomes. There have been efforts to conceptualize and measure components of ASR and associated processes/mechanisms, but, at present, there is no unifying conceptualization of ASR in childhood. A search of key databases supplemented by snowballing was undertaken for definitions/conceptions and theoretical models of ASR with a focus on children. An interpretive synthesis approach was used to identify themes from the definitions and models. The themes formed the basis of the proposed unifying generalized model of ASR in childhood, which is the main contribution of the article. At the center of the model is bottom-up reactivity to food, food cues and hunger, satiation and satiety signals, together with top-down regulatory control. An additional contribution is the proposed 5 interacting and overlapping domains (biological, hedonics, cognitive, behavioral, and traits) that function in and influence both bottom-up reactivity and top-down regulation. The domains also contribute to ASR outcomes of enactment and competence. External contextual and intrapersonal factors are conceived as impacting the domains and the bottom-up, top-down processes. The relevance of the model for explanations of ASR phenomena in childhood and children’s food choice and diet quality, as well as its implications for research directions and approaches to preventive intervention, including food parenting practices, are discussed. The model provides a framework for researchers and practitioners to support and interpret children’s problems and competence in self-directing food choices, energy intake, and nutrition.

INTRODUCTION

Along with domains of self-regulation, such as behavior regulation, cognitive regulation, and emotion regulation, appetite self-regulation (ASR) is a core developmental task of childhood.^1–3 Regulation in the domains of behavior, cognition, and emotion has been linked to major developmental outcomes, such as school achievement and behavioral, social, and emotional adjustment.^1,4,5 Appetite self-regulation has emerged as a central individual ability to promote the consumption of healthy diets and energy balance that develops during childhood. In simplistic terms, ASR pertains to “over-regulation,” as in the case of picky or fussy eating, associated with poorer nutrition and diet variety^6,7 and “under-regulation” or “overeating,” with consequences for excess weight gain in childhood.^8–10 Children’s ability to self-regulate appetite and eating is important in resisting tempting but unhealthy foods, achieving better balanced energy intake and expenditure, and consuming healthier diets.⁸ An environment with a wide availability and promotion of palatable but unhealthy food increases the relevance of ASR.

Interventions to improve children’s dietary intakes and nutrition, sometimes via parents’ feeding practices, have usually been directed to the acquisition of knowledge, attitudes, skills, and behaviors about healthy eating.^11–15 These interventions have had limited impact on children’s diets and weight in the longer term.^14–19 The evidence on the limited sustainability of interventions to promote change in children’s diet, energy balance, and eating behavior highlights the importance of formulating alternative approaches.¹⁶ As noted, ASR is a factor in children’s management of food choices and intakes, suggesting that enhanced ASR has the potential to contribute to sustainable improvements in children’s eating behaviors and diets in a parallel way to “general” self-regulation (GSR), contributing to long-term developmental outcomes.^1,4,5 The sustainable development of ASR in childhood is dependent on theory and evidence on its fundamental mechanisms and processes, which can be better understood through the application of a theory to research. A unifying conceptual model of mechanisms and process in ASR in childhood (a question of “how it works”) is necessary for the translation of basic research into applied efforts and outcomes. For example, the formulation of guidelines for food parenting practices and the design of preventive interventions are enhanced by the clarity of models about how children self-regulate their appetite and factors that influence ASR because these models provide the evidence base upon which such programs and interventions are designed. By understanding when and why some children display ASR enactment or competence difficulties, tailored strategies to support the development of ASR could be devised.

A unifying conceptual model of ASR in childhood is further needed because of a degree of vagueness about definitions and process/mechanisms inherent in ASR and children’s eating behaviors.^20,21 Without a model there are challenges in defining and measuring ASR as an outcome in research, and as a target in interventions. Nevertheless, a recent conceptualization of the core process in ASR is that it involves a bottom-up approach and top-down regulatory control,^22–27 also described as a dual processing approach.^28,29 This core process has also been applied to the “non-food” domains of self-regulation in childhood (eg, behavior, cognitive, emotion), usually described as GSR.^30–32 Briefly, the bottom-up, top-down approach involves “bottom-up” reactive or automatic responses to signals, stimuli, or situations together with “top-down” regulatory or control processes. The bottom-up component includes signals of hunger and satiety and interceptive signals of nutrient availability.^24,25 The reactive and control processes have also been conceived of as “impulsive” and “reflective,” respectively.³³ Bottom-up processes are assumed to be served by subcortical brain structures, with top-down processes served by cortical structures and distinct neural networks.^1,34

In addition to the bottom-up, top-down process, other theoretical approaches to ASR have been proposed—for example, that ASR involves responding to the cues of hunger, satiety, and satiation.^35,36 In addition, others have treated ASR as a multidimensional construct that includes traits such as food responsiveness and processes such as those outlined in the Satiety Cascade, and individual skills or strategies (such as the ability to delay gratification). At present, then, there are many conceptions of ASR and no concurrence about definitions and inherent processes and mechanisms.

Research and practice would both benefit if the diversity of processes and mechanisms about ASR in childhood could be integrated into a unifying model. The model would suggest new research directions, enhance measurement, and assist in the interpretation/explanation of commonly studied ASR phenomena such as eating in the absence of hunger^37–39 and caloric compensation,^37,40,41 delay of gratification for food reward,⁴² and trait or profile measures of ASR such as “avid eaters,”⁴³ where there are large differences between children but where the reason for variability is uncertain. A unifying model could also help explain the observed decline in measures of ASR across childhood,³⁰ which contrasts with improved GSR development over the same ages.

The main objective of the present narrative review was to propose a generalized conceptual model of mechanisms and processes in ASR in childhood. The relevance of the model for explanations of ASR phenomena in childhood and children’s food choice and diet quality, as well as its implications for research directions and approaches to preventive intervention and food parenting practices, are discussed.

METHODS

To identify relevant literature that outlines models, definitions, and conceptions of ASR in childhood research articles, reviews and theoretical contributions were collected from searches in key databases (Google Scholar, PsycINFO, Medline complete) using the terms “appetite self-regulation” (human limited) as well as specific ASR terms (eg, eating in the absence of hunger, delay of gratification). Titles and abstracts were scanned for (1) a focus on children or childhood (ages or stages were not specified) or (2) models/theories/conception of ASR.

These articles were then read, with those that contained definitions/conceptions of ASR in childhood or a model/theory of ASR retained. Citations in these articles were searched for other relevant publications (snowballing). An interpretive synthesis approach was used to thematically organize and represent the definitions/conceptions and models from the search. Definitions/conceptions and models were included in the synthesis when they were judged to add new information, in that they differed from or added new dimensions or aspects of influences on and processes in ASR in childhood from those already included. Definitions/conceptions and models were excluded if they were judged to duplicate, parallel information already obtained. Both authors discussed and agreed upon the definitions, their thematic organization, and the structure and organization of the model.

References involving theoretical descriptions of GSR were also collated by drawing on an earlier review describing GSR and the references contained therein.³¹ There were 3 foundation references on GSR that served the purpose of providing theoretical/conceptual descriptions of GSR, together with 2 more recent supportive publications.

Terminology

The ASR and GSR conceptual models (and associated research) have used several terms to describe the elements in the model. These include “entity,” “variable,” “factor,” “component,” and “domain.” The term “component” is used here to refer to all elements within a model, whether they are substantive domains, factors, or variables, or linking processes, mechanisms, or influences. The components that influence or function in ASR are described here as “domains of influence.”

Definitions/Conceptions of GSR

The literature on GSR was also used in constructing the narrative review because of contrasts and parallels between GSR and ASR and the proposal that ASR be conceived under the umbrella of GSR as a separate domain.^30–32 There were 3 foundation references on GSR that are widely cited in the literature on GSR that served this purpose,^1,34,44 together with 2 additional recent supportive publications.^45,46 Descriptions of the dual processing approach were also consulted. Informed by our prior work comparing and contrasting GSR and ASR,³¹ the references were examined for theoretical aspects of GSR that could inform the development of a unifying ASR model.

Nigg¹ provides a unifying analysis and model of GSR processes and mechanisms. He reasoned that top-down (deliberate) and bottom-up (automatic) aspects mutually influence one another. Nigg argued that self-regulation involves physiological systems as well as homeostatic and allostatic mechanisms in response to stress, challenge, or new information and that it encompasses strategic/deliberative as well as reactive/automatized processes and their reciprocal influences. Wesarg-Menzel et al⁴⁵ used Moilanen’s⁴⁷ definition of self-regulation to emphasize the role of personal goals and motivation in the activation of self-regulation. Blair and Ku⁴⁶ proposed a model of self-regulation with emotional, behavioral, physiological, and genetic components that are reciprocally and recursively related.

Overall, it was evident that GSR is conceived as being multifaceted, with the dynamic interaction and mutual influence of bottom-up reactivity and top-down regulatory control. It includes behavioral and neurobiological, cognitive, and emotional processes and recognizes both behavioral inhibition (avoidance tendencies) and behavioral under control (approach tendencies), internal and external cues, monitoring, and personal motivation. General self-regulation was treated in these publications as a generalized construct without a developmental dimension.

Definitions/Conceptions of ASR

The interpretive synthesis yielded 3 main groups of definitions/conceptions of ASR: (1) those with a focus on a bottom-up, top-down, or dual processing approach; (2) those where the focus was specifically food-related, such as hunger, satiety cues, food approach, and food avoidance; and (3) definitions and conceptions that included multiple components and processes. There was no age- or stage-related thematic dimension to the definitions/conceptions, which have been summarized in Table S1.

The definitions and conceptions presented in Table S1 showed that much of the literature on ASR simply refers to and defines “self-regulation” rather than separating ASR from GSR. In those cases, the core of the conceptualization involved the bottom-up, top-down process or similar ideas. The focus was on self-regulation or ASR as a process and when applied to individuals as an ability or competency. Individual differences were dealt with through concepts such as differences in the reward value of food or inhibitory control abilities. When attention turned to ASR specifically, often food-related ideas such as hunger, satiation, and satiety cues were incorporated into a bottom-up approach (eg, food reward/hedonics) or avoidance (eg, food neophobia/fussiness). Nevertheless, new, or different (from GSR), elements to definitions and conceptions of ASR were evident, such as internal cues of hunger, satiety, and satiation. In common with GSR, regulatory control was mainly conceived in terms of inhibitory control or impulse control abilities, or more broadly in terms of executive functions or effortful control.

The examples in the section on “other definitions or conceptions” (Table S1) include multiple components and processes in ASR. This conception of ASR as involving interacting components is especially reflected in person-centered research, which uses multiple measures to identify subgroups of children having different profiles or phenotypes.^20,48,49 It is assumed that the phenotype arises from an interaction of a genotype with the environment⁵⁰ and that ASR differs from 1 phenotype subgroup to another.

The definitions and conceptions demonstrate that ASR has been treated as an individual capacity that is trait-like, as a process, and as a skill that could be developed or changed. Some definitions and conceptions recognized that ASR involves influences and contributions from behavioral, cognitive, emotional, neural, motivational, and social domains. Also, a temporal component to ASR was acknowledged in the sense that food reward or aversion is usually immediate, but the consequences (such as weight gain) can be more delayed. Authors at times mentioned biological elements to ASR, such as neural, biological, or physiological processes, but these were not articulated in definitions.

A definition and overall conception of ASR in childhood were formulated from integrating the definitions and conceptions in Table S1. The definition and conception are intended to apply to levels of ASR competence as a trait-like potentiality or ability as well as to the enactment of ASR in food and food cue situations.

Definition: ASR is a multidimensional latent construct that involves dynamic recursive bottom-up/top-down processes in managing reactive responses to food and food cues and internal states of hunger, satiation, and satiety that, in turn, are influenced by individual characteristics and processes as well as intrapersonal and external contextual factors.

Conception: ASR occurs at multiple levels and is integrated across several interacting domains of influence and functioning (described below) in the context of intrapersonal and external influences. It involves competence in the regulation or management of one’s food choices and intakes. When enacted in situations, it involves the engagement of regulatory processes (especially inhibitory control) to regulate approach or avoidant responses to food cues as well as to regulate hunger, satiety, and satiation cues. In addition to activating top-down regulatory processes to bottom-up reactive food approaches or food avoidance, it involves monitoring and responding to emotions, internal and environmental cues, and feedback from others, as well as applying cognitive strategies and regulatory behaviors. Increasingly across childhood, ASR can be activated and directed by food-related goals, such as being healthy or having a healthy diet. The bottom-up, top-down process is at the core of ASR both in terms of overall competence and its enactment in particular situations. This process is influenced proximally by traits such as food responsiveness or cognitions such as food beliefs, and more distally by intrapersonal factors such as prior learning and food experiences and external contextual factors such as the food and social environment of the child.

This conception broadens earlier perspectives of ASR that had a focus on energy imbalance through the self-regulation of energy intake and ASR as a response to hunger and satiety cues. Rather, ASR is considered here to encompass both the quality of foods consumed as well as energy balance. The conception suggests that ASR can be conceptualized first in relation to its enactment in regulatory situations (state-like) and general competence (trait-like). ASR failures are conceptualized in terms of both “approach” (eg, excess consumption of high-energy/low-nutrient foods, overeating) and “avoidance” (eg, not consuming a range of healthy foods, undereating), with ASR competence ideally reflected by the selection and consumption of a varied, healthy, energy-balanced diet. In this sense, better ASR competence ideally enables children to engage in healthy eating while lower ASR competence can contribute to poor diet quality, overeating, and outcomes such as weight gain and obesity on the one hand (“approach”) or fussy eating and attendant low diet quality on the other (“avoidance”). “Ideally” is used here to indicate that ASR in childhood is within the constraints provided by the external contextual factors.

Theories and Conceptual Models of ASR

The interpretive synthesis suggested 4 themes in the theories and models. The first was general theories of eating and eating regulation. The second was biopsychosocial and biobehavioral models. The third were psychological models, and the fourth were developmental models. There were some additional models outside of these thematic groups. Examples of the models and theories in these groups are set out in Table S2.

It was evident from the review of models that ASR in children (competence or enactment) involves not only bottom-up reactivity and top-down regulation but also a complex interplay between the following domains of influence and functioning: biological (eg, metabolic, neural), behavioral (eg, speed of eating), cognitive (eg, cognitive control), hedonics (eg, reinforcing value of food), and traits (eg, fussiness, food responsiveness) interacting with intrapersonal and environmental/situational factors. Commonalities and distinctions between GSR and ASR were apparent.

The models differed in the degree to which they were related to outcomes (eg, incidences of ASR success or failure, ASR competence, diet quality, weight status). If the focus was on the nature and mechanisms associated with incidences of ASR success or failure, the model involved only the component constructs and suggested processes. Despite this heterogeneity in focus and purpose of the models, several constructs and processes emerged consistently. Six were prominent and provided input for our proposed unifying model: (1) the hedonic/reward value of food and food cues; (2) abilities to inhibit or control impulses; (3) awareness of and responses to internal signals, including of hunger, satiation, and satiety; (4) bottom-up, top-down, or dual processing processes; (5) the importance of individual differences in levels of ASR competence from early in life; and (6) reference to the role of other factors, including biological processes, traits, and cognitions.

Development of the Unifying ASR Model

The ASR models from the literature (Table S2) were constructed through the integration of existing theories and empirical evidence. The same process was used for the present unifying conceptual model. That is, prominent aspects of the model were incorporated from existing models and the proposed definition and description of ASR. Empirical research findings that support the model’s basic framework and operation were also used to inform its development. Much of this research is cited in supporting publications provided throughout the description of the model, as shown below. Nevertheless, by its nature, there is substantial novelty in the model that unifies much of the existing scholarship.

PROPOSED UNIFYING CONCEPTUAL MODEL OF ASR IN CHILDREN

Overview of the Unifying ASR Model in Childhood

The proposed unifying ASR model (Figure 1) represents ASR as a dynamic and recursive process and is at a generalized level, meaning that it does not specify variables associated with the components. At the center of the model is top-down regulatory control of bottom-up food approach and avoidance reactivity. The bottom-up component pertains to the initial prepotent or reactive and automatic responses to food, food cues, or hunger cues (and other environmental cues such as time of day or social context). In the recursive aspect of the bottom-up, top-down model during eating episodes, bottom-up signals, such as hunger, satiation, and satiety, interact with top-down regulatory responses. The top-down component refers to regulatory processes that are self-generated and involve inhibition or control of reactive approach or avoidant responses. The bottom and top are conceived as engaging in cross-talk or as mutually influencing over episodes involving responding to food and food cues, eating, and self-regulation.

Figure 1.

Unifying Model of Appetite Self-Regulation (ASR) Constructs and Processes in Childhood Incorporating Domains of Influence and Functioning

The model includes 5 domains that influence and are functionally related to or involved in bottom-up and top-down processes: biological, cognitive, behavioral, traits, and hedonics. This is shown in Figure 1 by every domain intersecting with the bottom-up, top-down core of the model. For example, top-down cognitive processes in ASR could be influenced by and/or functionally involve genetics, temperament, and the strength of different appetitive traits, as well as intrapersonal (eg, emotions) or contextual factors such as the social situation or expectations from family or peers. These same types of influences could impact the strength of bottom-up reactivity to food or food cues and be functionally involved in reactivity. The model suggests that both the bottom-up and top-down processes are multidimensional with a complex interplay of influences and functioning from the domains. This means that bottom-up processes involve more than “food approach” or “food avoidance.” Similarly, top-down processes involve more than “inhibitory control.” The model therefore presents ASR as a construct involving multiple, interacting characteristics and processes influencing or functioning in bottom-up reactivity and top-down regulation. Furthermore, as explained below, the interacting characteristics and processes appear to differ across subgroups of children, and individual differences in ASR competence of enactment could arise from all components and processes in the model.

The model also recognizes wider influences from intrapersonal and external (environmental) contextual factors. Through the domains of influence and functioning, the model incorporates biopsychosocial/biobehavioral perspectives. Boundary conditions can moderate the extent to which self-regulation is balanced more to bottom-up reactivity or to top-down regulation.⁵¹ These conditions could be situational (eg, the social context) or dispositional, such as an overriding attraction to a particular food that inhibits the execution of self-regulatory behavior. In the present model, boundary conditions could arise from any of the 5 domains, meaning that the balance to reactivity or regulation could be shifted by many factors in the domains.

ASR Outcome

The outcome of the model is ASR. ASR pertains to (1) the enactment of ASR (in situations involving reactivity to food, food cues, and hunger, satiation, and satiety signals) and (2) ASR competence. Competence is conceived as including constructs from the domains of functioning. For example, ASR competence encompasses relevant cognitions, traits, hedonics, behaviors, and biological factors (such as brain connectivity and brain responses to palatable food). ASR enactment is a narrower construct than competence and involves situational instances of ASR success or failure (eg, Eating in the Absence of Hunger, EAH or failure to demonstrate caloric compensation). The domains of influence and functioning are implicated in ASR outcomes in terms of both enactment and competence (eg, the eating behavior of meal termination influences the amount of food consumed but is also an indicator of ASR enactment; similarly for a trait such as food responsiveness). The separation of enactment and competence has parallels in relation to GSR, where specific behavioral measures and traits from parent-report, self-report, or teacher-reports have been used to measure self-regulation.^1,34,52–56

Domains of Influence and Functioning

Components of the domains are assumed to influence bottom-up and top-down processes. They are also conceived to be functional elements of the top-down and bottom-up processes themselves. For example, components of the domains could be involved in bottom-up reactivity (eg, the trait of food responsiveness is an element in the bottom-up food approach) and top-down regulatory control (eg, executive function is inherent in top-down control). Components of the domains could also operate as moderators and mediators in the bottom-up, top-down recursive process. For example, regulation of food avoidance or food approach could be mediated by cognitions about food healthiness or moderated by temperamental impulsivity. On the other hand, levels of bottom-up reactivity could be partly mediated by domain components (eg, food avoidance partly mediated by beliefs about food energy levels) or moderated by domain components (eg, taste sensitivity or temperamental fearfulness could increase bottom-up avoidance).

The multiple roles of components in the domains are depicted in Figures  2–4. Figure 2 shows direct influences from the domains on bottom-up and top-down processes. Figure 3 captures the moderating and mediating roles of components in the domains on bottom-up reactivity processes and top-down regulatory processes. Finally, Figure 4 represents the idea that components of the domains are inherent in the bottom-up and top-down processes themselves. The proposed domains of influence and functioning are consistent with the literature on GSR where scholars include behavioral and neurobiological, cognitive, and emotional processes in GSR.

Figure 2.

Influences on the Bottom-Up and Top-Down Components of the Unifying Child Appetite Self-Regulation (ASR) Model

Figure 3.

Domains of Functioning Acting as Mediators and Moderators Between Top-Down Regulation and Bottom-Up Reactivity

Figure 4.

Components of the 5 Domains of Influence and Functioning Are Also Inherent in Bottom-Up Reactivity and Top-Down Regulation

The 5 domains are presented as overlapping (Figure 1). This can arise because a construct or measure is relevant to more than 1 domain. For example, the behavioral domain contains eating behaviors, but at the same time there are behavioral measures in the other domains such as the traits domain. Alternatively, the overlap could arise because constructs in 1 domain underpin constructs in another domain (eg, biological factors underpin traits and behavior). Intrapersonal and external (environmental) factors are assumed to influence the domains, together with bottom-up and top-down processing.

The 5 domains of influence and functioning are explained below to outline their role in the model, including their interrelationships and contributions to bottom-up reactivity processes, top-down regulatory processes, and ASR outcomes.

Biological Domain

Biological processes and characteristics might be construed as the “infrastructure” that underpins and contributes to ASR processes.³⁰ The biological domain of the ASR model includes (1) processes or systems of appetite regulation as well as (2) biologically based individual characteristics. The biology of appetite regulation and ASR is complex, involving a multitude of process, systems, and characteristics.^57,58 These include metabolic processes involving interactions between the gut and brain (the gut–brain axis, which includes signals of satiation and satiety⁵⁹), the role of the neuroendocrine system in governing appetite and satiety,⁶⁰ hormones and peptides associated with satiation and satiety signaling, cephalic phase response, and neural processes and pathways such as dopaminergic reward pathways, sensory processing pathways, and activity in the prefrontal cortex. It also encompasses cross-talk between homeostatic and hedonic systems,⁵⁹ preferences,⁶¹ and subliminal gut–brain reward pathways.^62,63 The biological domain also includes physiological responses to emotions and stress (eg, hypothalamic-pituitary-adrenal axis activation^64,65).

There are also individual biologically based characteristics that influence and function in ASR. They include innate sensory preferences (eg, towards sweet taste, away from bitter taste⁶⁶), sensory acuity or sensitivity,^67,68 genetics and epigenetics,^69,70 temperament (eg, inhibitory control, impulsivity^30,71), food cue and reward sensitivity,^72,73 food texture preferences,⁷⁴ gastric sensation (eg, interoceptive ability⁷⁵), and brain structure and development (eg, development of the prefrontal cortex^76,77).

Appetite self-regulation can be both cause and consequence of unhealthy diets and/or weight gain, as consumption of unhealthy foods and weight can affect brain structure and functioning, which, in turn, affects ASR through disruptive effects on the inhibitory-related prefrontal cortex.⁷⁸ This can disrupt gut–brain pathways for satiety signaling,⁷⁹ and reducing sensitivity to homeostatic and interoceptive cues.⁸⁰ Longitudinal studies that enable the causal pathways explaining ASR development and disruption are needed to provide further information on how diets affect and are affected by ASR.⁸⁰

Biological processes and characteristics can affect the strength of a bottom-up approach and avoidance. For example, interoception, responses to food taste and texture, as well as gut–brain signaling could enhance bottom-up reactivity and be a factor in individual differences. The capacity for top-down regulation is related to brain development and connectivity, as suggested from research on executive function and brain networks associated with responses to food.^30,81 Biological measures (eg, brain responses to palatable foods) are also indicators of ASR competence or enactment.

Behavioral Domain

The behavioral domain encompasses eating behaviors such as food choices (including for portion sizes and food types), energy compensation behaviors (eg, meal termination/initiation), oral processing (eg, speed of eating), eating frequency, and behaviors associated with responses to food environments and contexts (such as disinhibited eating in contexts with peers and palatable food). Behaviors are also functionally integrated with bottom-up and top-down processes. For example, indicators of bottom-up food approach and food avoidance comprise several eating behaviors (eg, avoiding unfamiliar foods, preferring energy-dense foods over less-energy-dense foods) and the termination of eating is a top-down regulatory behavior. Indicators of top-down regulatory control include behavioral measures such as EAH, Delay of Gratification (DoG), Go/No-Go tasks or Stroop tasks.⁸² The biological domain (eg, temperament), the hedonics domain (eg, pleasure from eating), and cognitive domain (eg, decision making, impulsivity, and inhibitory control) can include behavioral measures. This is reflected in the model with overlaps between the behavioral and other domains.

Traits Domain

The traits domain includes appetitive and other traits constituting relatively stable patterns of behaviors, cognitions, and emotions that are consistent over time and context. The main appetitive traits associated with ASR are linked to food approach and food avoidance as well as regulatory control.⁷⁰ Traits such as food responsiveness or fussiness are expected to impact or be functionally related to bottom-up reactivity. They could also have a role in top-down regulatory processes by moderating inhibitory control. Relevant traits are also evident in other domains of the model, including temperament traits (biological domain), impulsivity (cognitive domain), and the motivational domain (food approach and avoidance traits). Traits could act to enhance or inhibit eating behaviors. For example, trait “relative reinforcing value of food” could contribute to daily energy intake.⁵⁰ The trait of emotional eating could evoke overeating in situations involving emotions and palatable food.⁸³ As suggested above, elements from the traits domain would also be included in the ASR competence construct.

Cognitive Domain

Cognitive factors influence and function in ASR in multiple ways. They directly influence bottom-up and top-down processes or act to mediate or moderate the effects of top-down regulation on bottom-up reactivity. Cognitive factors are implicit or explicit. They include (1) beliefs such as about the importance of self-regulation or the energy density of different foods; (2) perception, awareness, and interpretations of internal cues of hunger, satiation, and satiety; (3) attitudes (eg, towards healthy foods); (4) knowledge about food and its regulation, including knowledge of ASR strategies; (5) possible eating and weight goals; (6) self-perceptions (eg, about body shape and weight) as a self-regulator; (7), expectations (eg, about taste or satiation effects of different foods); and (8) information-processing and decision-making styles (eg, information source used for decisions), and rational (eg, reasoning about food sweetness) versus emotional (eg, how nice the food will taste) decision-making.^84,85 Executive functions are widely researched as contributing to ASR.^42,86–88 Especially important in the latter research has been inhibitory control and impulsivity.^82,89,90 Cognitive factors and their role would be expected to increase across childhood—for example, in relation to food knowledge or literacy^91–93 and cognitive capacities associated with the development of executive functions.^86,94–96

In the model, cognitive factors are considered to have a role in both bottom-up reactivity (eg, via impulsivity, beliefs about the palatability of food, knowledge about consequences of a diet high in sugar and fat) and top-down regulatory control processes (eg, via inhibitory control capacities, self-efficacy beliefs). Cognitive elements reciprocally interact with factors in other domains, especially the biological domain and the hedonics domain in contributing to ASR (both ASR enactment and competence). For example, metabolic state and hunger/fullness cues (pertaining to interoception) influence the way food is perceived.⁶² Also, gut and brain responses to food and the motivation to eat are likely to be enhanced by beliefs and perceptions about palatability, and cognitions can increase or decrease the hedonic value of food.⁹⁷

Hedonics Domain

This domain is about reward and pleasure aspects of food and eating.^98,99 This includes the affective liking and disliking, the motivational wanting/desire for food, food preference, and the dynamic nature of hedonics.^100–103 These constructs are expressed mainly via behaviors and traits. Because many behaviors and traits have a hedonic element, and hedonics is central to ASR, it is conceptualized here as a separate domain. In relation to the other domains of influence and functioning, cognitive processes are involved in (contribute to or underpin) hedonics, and biologically there are taste, gut, and brain processes associated with hedonics. There are multiple influences that can enhance (eg, social contexts, increased hunger) or reduce (eg, sensory-specific satiety) the hedonic value of food and eating. The hedonic properties of food influence dietary decisions and choices.⁶²

In combination with factors from each of the other domains (eg, trait food responsiveness, gut and brain response to palatable food, beliefs, and expectations about pleasure from food, the choice of palatable food), hedonics has an impact on bottom-up reactivity to food and food cues to increase attraction and expectations of pleasure and the reward value of food and eating, including consciously and subliminally or unconsciously.^62,63 Greater sensitivity to the reward value of food and the consumption of energy-rich foods could weaken inhibitory control and activate neurocircuitry and physiological effects such as salivation that further enhance bottom-up reactivity to foods rich in fat and sugar.¹⁰⁴ Subjective evaluations of food reward (or aversive evaluations) contribute to food choice and quantity.¹⁰⁵ It is assumed that hedonics influences top-down regulatory control (eg, more robust inhibitory responses would be needed with strong food reward reactivity). This is consistent with the idea of boundary conditions affecting the balance between bottom-up reactivity and top-down regulation.⁵¹ Decreased hedonics via sensory-specific satiety or loss of novelty, for example, could contribute to the control of eating.¹⁰⁶ Hunger (biological domain) influences food reward/wanting, with the drive to eat and for food reward being higher when children are hungry. It has been argued that hedonic concepts should be integrated with biological factors associated with homeostatic regulation.¹⁰⁷ Measures of hedonics should also be contained in a broadly conceived ASR competence construct and in ASR enactment, such as the ability to limit the intake of snacks or willingness to try a food that has prompted aversive reactions.

Intrapersonal and External Influences

In parallel with GSR models,¹⁰⁸ both external (from the child’s environment) and intrapersonal (from the child) factors, and their interrelations, contribute to or influence ASR. Contextual influences on ASR include distal factors (eg, wider food environments including the availability of healthy and unhealthy foods, food advertising, social influences on eating, media, social norms) and proximal factors (eg, the presence of others when eating, the location of eating, time of day).¹⁰⁹ Contextual factors can be intrapersonal, such as the child’s emotional state, or external, such as food advertising. Intrapersonal contextual factors include prior learning (eg, arising from parent feeding practices) and memories of foods (from prior experiences of eating), which influence how children react to foods and food cues, perceive, and understand foods and eating.^110–113 In the model, as well as the domains of functioning, the intrapersonal and external factors are assumed to impact bottom-up reactivity and/or top-down regulation. For example, food advertising and the availability of highly palatable snacks impact the bottom-up food approach,¹¹³ especially for some children, as shown by the individual differences in EAH noted above. Prior learning and exposure to models of overeating would be expected to increase food-approach reactivity.^111–113 Intrapersonal and external influences in the model are assumed to act through the domains of influence and functioning to impact both the core bottom-up, top-down process and ASR outcomes.

DISCUSSION

The model provides several directions for future research on child eating and ASR, as well as generating insights for explanations of ASR phenomena. In addition, the model has implications for practice, including with respect to children’s diet and food choice, approaches to intervention, and for food parenting practices. Finally, the model has implications for theory and research about the development of ASR. Each of these is discussed below.

Implications for Research, Including Explanations of ASR Phenomena

The model serves as a stimulus for research designs that test hypotheses about the influences on ASR competence and enactment outcomes, as well as how and why ASR influences diets and weight. To test aspects of the model, attention needs to be directed to how the bottom-up, top-down process is integrated with and influenced by components in the domains. Situational and dispositional boundary conditions⁵¹ could be examined as factors that contribute to a balance more to bottom-up reactivity or top-down regulation. For example, temperament shyness could increase bottom-up avoidance or stress could increase a bottom-up approach.

The multidimensional conceptualization of ASR has implications for measurement. It suggests that the examination of ASR requires increased use of multi-method and multivariate methodologies. Further, consistent with multidimensionality, ASR outcomes should be measured in multiple ways, including measures of children’s abilities to stop eating when full (satiation response), waiting to start eating until hungry (satiety response), selecting and consuming a range of healthier foods, and selecting and consuming appropriate portion sizes.

In the model, ASR processes involve interactions among constructs from 5 domains of influence and functioning. For this reason, person-centered research designs are suited to the investigation of ASR as a multidimensional construct and especially individual differences in the formulation and functioning of ASR.^20,49,114 The phenotypes generated by this approach show that ASR can be differently constituted and expressed from 1 child to another. Research on phenotypes has included behavioral measures such as EAH and caloric compensation,^37,115,116 parent-report measures such as the Child Eating Behavior Questionnaire,^43,49 biological measures,¹¹⁷ and executive function.⁸⁷ The approach has yielded profiles of children who display combinations of traits or behaviors suggesting ASR difficulties^43,49,50,118—for example, “avid eaters” (which could be low ASR competence, approach)⁴³ or food-avoidant phenotypes (which could be low ASR competence, avoidance).⁴⁹ The understanding of ASR could be enhanced if the phenotypes included genetic and neural measures.⁷⁰ The analyses should also include intrapersonal and external contextual factors as covariates to better understand influences on ASR processes.

The model can also assist in the interpretation and explanation of research findings. For instance, protocols such as EAH, caloric compensation, and DoG reveal substantial individual differences in ASR enactment.³⁰ There are comparable individual differences in food-avoidance ASR such as food fussiness^119,120 and food neophobia¹²¹ as both enactment and competence. The model provides several options for explanations of these differences—for example, in relation to the trait of food fussiness, which would be considered low ASR competence (avoidance). According to the model, high levels of the trait food fussiness could be explained by, for example, interactions between high sensory sensitivity (eg, to tastes, textures; biological domain) and/or temperamental fearfulness (biological domain) interacting with strong behavioral responses to feelings of satiation or satiety (satiety responsiveness; behavioral/traits domain), which could act in tandem to produce strong bottom-up avoidance responses, which, when combined with food beliefs contributing to weak top-down regulatory control (inhibitory control; biological/cognitive domain), result in food-avoidant behaviors. Low food fussiness (high ASR competence), in contrast, could be explained by strong top-down regulatory control along with weak bottom-up behavioral food-avoidance responses, possibly associated with low temperamental fearfulness and high impulsivity in tandem with higher food-approach tendencies. Using the model as a framework, similar explanations could be generated for other forms of ASR competence or enactment success or failure.

The model could also facilitate the study and explanation of individual differences in constructs such as EAH. Higher EAH as measured with the EAH laboratory paradigm (ie, ASR enactment failure, approach) could be interpreted in terms of high food cue responsiveness (hedonics domain) interacting with weak behavioral responsivity to satiation cues (behavioral domain) in the context of prior learning about food rewards (intrapersonal contextual influences) to produce low regulatory control (contributing, in turn, to increased energy consumption and weight gain). Indeed, there are many possibilities in relation to both the constructs of relevance within each domain and how they interact that could explain high food fussiness or EAH. Including these constructs in studies of ASR will advance understanding of individual differences in ASR competence and enactment.

Implications for Practice: Children’s Diet, Interventions, and Food Parenting

Each of the domains of influence and functioning in the model generate implications for enhancing ASR via interventions or guidelines on food parenting practices. Multiple influences on bottom-up reactivity and top-down regulatory control and their recursive interplay are associated with each of the model’s domains. These include the following: (1) consumption of highly processed food, levels of stress, temperament characteristics, and impulsivity from the biological domain; (2) eating behaviors, food choice and preferences, and food avoidance from the behavioral domain; (3) food responsiveness, food fussiness, and emotional eating from the traits domain; (4) food perceptions and belief, including food literacy from the cognitive domain; and (5) food liking and wanting together with the reward value of food from the hedonics domain.

Each of these influences could be a target for intervention or for food parenting guidelines to impact ASR outcomes (including competence and enactment). For example, in relation to the cognitive domain, and especially associated with cognitive development, increased “food literacy”⁹² is a possibility for enhancing ASR competence and enactment in children. This would include knowledge about types of food (eg, being able to recognize and categorize foods and having courage to try new foods), food ingredients (and comprehension of food labels), ethical aspects of food production, food sustainability and health, as well as the influences of media and advertising on their food and diet.^92,93,122 Knowledge might be extended to include aspects of meta-cognition, such as knowledge of strategies to improve diet, to assess foods and to regulate appetite, monitoring, and the self-evaluation of ASR. Additionally, parental feeding practices that disrupt ASR development (eg, nonresponsive feeding practices)^123–125 could be targeted to improve ASR outcomes, such as reduced food responsiveness,¹²⁶ through the activation of the hedonic appetite system and a weakening of inhibitory control.

The person-centered approach to investigating ASR phenotypes is well suited for adaptation to approaches to intervention and guidelines for food parenting practices. The notion is that ASR is constituted and configured differently across individuals or groups of children. Acosta and colleagues¹²⁷ used this idea to target obesity intervention according to 4 phenotypes. The implication for ASR is that interventions and food parenting should be tailored to the child’s particular ASR phenotype, such as “avid eating,” “emotional eating,” or “avoidant eating.”^43,123,128 The effectiveness of this approach would be dependent on whether measures used to describe the phenotypes included components from each of the domains of influence and functioning—for example, whether levels of stress and impulsivity, eating behaviors and traits, cognitive factors such as food literacy, and hedonics such as the reward value of food were measured. This would assist in the tailoring of intervention targets or food parenting practices to improve ASR. Interventions could also be helped if measures of the family eating behaviors and the home environment were included.¹²⁹

Development of ASR

While an examination of the developmental course of ASR was out of the scope of the current review, ASR processes and mechanisms could be expected to emerge and change during development. Taking account of children’s cognitive, social, and emotional development and changes in contextual influences with age (eg, a decline in the role of caregivers and an increase in the child’s self-regulation) is important for the understanding of the development of ASR and the relevance of the model at different developmental stages. This would involve the assessment of the model at different ages—for example, infancy, early childhood, and middle childhood. This includes possible differences in the relative importance of the domains across development and whether there is a general developmental course of ASR, such as a decline with age.³⁰ Temperament measures in the biological domain, for example, could be more important in infancy than in middle childhood.

The early course and origins of ASR development in infants and toddlers and the examination of the applicability of the model to this age group need to be investigated. With age and attendant brain development, executive functioning and GSR improve, including the capacity for effortful and inhibitory control.³⁰ This will impact ASR competence and enactment. In contrast, exposure to palatable food cues, food advertising, and other external influences across childhood also means that the prepotency of a bottom-up approach is likely to be stronger in older children,³⁰ which will affect ASR competence and enactment. The interplay between improved regulatory control and increased food approach could determine whether and why ASR declines for some children across childhood, unlike the general trend for improved GSR.³⁰ As evidence accumulates on age-related changes in ASR processes and mechanisms, a developmental element to the model could be formulated. A developmental perspective could investigate trajectory-based phenotypes to determine whether subgroups of children displayed different ASR phenotype trajectories.¹³⁰ A developmental focus could also facilitate the construction of age-appropriate preventative interventions better matched to the characteristics and processes associated with individual children at different ages.

CONCLUSION

The present model unifies research and theory about ASR in childhood by conceptualizing ASR as involving multiple interacting components and processes that influence and function in bottom-up reactivity and top-down regulation. The model assists in the interpretation, understanding, and explanation of ASR competence and difficulties. It generates new hypotheses and research design in the investigation of ASR in childhood and could provide a foundation for the examination of developmental changes in ASR. It can facilitate the understanding of children’s food choices, provide opportunities for increasing food parenting practices to enhance ASR, and suggests that, with a focus on ASR in preventive interventions, it could contribute to sustained behavior change. The model provides a framework for researchers and practitioners to support and interpret children’s problems and competence in self-directing food choices, energy intake, and nutrition.

Author Contributions

Both authors contributed to the work’s conception and participated in the writing and critical revision of the article in a manner sufficient to establish ownership of the intellectual content. Both authors read and approved the version of the manuscript.

Supplementary Material

Supplementary Material is available at Nutrition Reviews online.

Funding

None declared.

Conflicts of Interest

None declared.