INTRODUCTION

Obesity is the most common nutritional disorder among infants and adolescents in both industrialized and emerging societies (¹). The prevalence of obesity has steadily increased recently. This increased prevalence is associated with the increased availability of food and changes in lifestyle, including increased physical inactivity, decreased physical activity in children and adolescents, and poor eating habits (²).

In particular, childhood obesity is a risk factor for cardiovascular, orthopedic, respiratory, digestive, dermatological, psychosocial, physical, neurological, and endocrine complications. Childhood obesity is also associated with certain forms of cancer and lower life expectancy, which justifies the prevention and early treatment of childhood obesity (³).

Research indicates that obese people have greater difficulty controlling their impulses than thin people (⁴). In this regard, obese individuals in weight reduction programs receive higher scores on scales that assess impulsivity, and they more frequently engage in impulsive behaviors, including substance abuse, than thin people (⁵,⁶). Impulsive individuals are less likely to engage in various behaviors that contribute to healthy eating patterns (⁷).

Impulsiveness might affect the ability of an individual to plan meals, eat regularly, and resist impulses to enjoy foods that are high in fat and sugars (⁸). In addition, research shows parallels between obesity and potentially addictive behaviors in adults – both obesity and addiction are characterized by impulsive behaviors and increased reward sensitivity (⁹).

This relationship is particularly important because of the prevalence of food availability in industrialized societies. In addition, a variety of environmental factors have changed over the last 30 years that are associated with increased obesity; for example, the increasing number of fast food restaurants enables greater access to cheap meals that are high in calories and fat (¹⁰). Consequently, the chance of overeating increases when greater food reward sensitivity exists (¹¹). For example, both laboratory and real-world studies have shown that exposure to appetizing environments is more effective at increasing food intake but only among participants who are highly responsive to reward (¹²,¹³). In this sense, it is more difficult for obese people to resist the temptation to overeat, despite the negative consequences that this behavior might lead to in terms of health and quality of life (¹⁴). This finding suggests that reward sensitivity is a causal mechanism for overeating in an obesogenic environment.

In this regard, several explanations for the mechanisms that underlie the ability to control impulsiveness have been proposed. One potential and suitable mechanism is reinforcement delay, which refers to the relative preference for immediate but small rewards compared with larger but delayed rewards (¹⁵). Reinforcement delay has considerable potential to improve our understanding of the behavioral processes associated with numerous problematic behaviors, including those associated with overweight and obesity (¹⁶). Research on this matter shows that obese children choose small but immediate rewards and exhibit difficulty with postponing gratification when large meal portions are presented as alternatives (¹⁷) primarily because they are more sensitive to rewards than children of normal weight (¹⁸). In addition, recent research has demonstrated that the inability to delay gratification predicts weight gain in children between the ages of 3 and 11 years (^{19 20}-²¹). The use of this basic paradigm emphasizes the value of the reward depending on delayed access to said reward as compared with another selection presented concurrently.

Research in this area has been limited to programs associated with different amounts of the reinforcer and variations in the delay of the reinforcer of a greater magnitude. In some cases, however, the delay of the latter may involve additional enhancing dimensions available within the individual's environment, such as the quality of the reinforcer, the effort of the task, and the reinforcement program. The procedures developed by Neef and Lutz (²²) were successfully used to assess the relative influence of reinforcer dimensions on children's impulsive behaviors and to develop self-control and decrease impulsive behaviors through the use of basic development models of self-control, such as the fading procedure (²³).

However, the literature on impulsive behavior in obese and non-obese people has focused primarily on evaluation rather than treatment because, in many cases, impulsivity is evaluated using inhibition response assignments or psychometric instruments. Currently, the focused evaluation of impulsive choice affected by reinforcer dimensions has enabled the development of procedures that increase self-control based largely on the fading procedures of late delivery of a reward of lesser value or gradual increase in the delay of a reward of greater value (^{23 24 25}-²⁶). Therefore, the objectives of this study were to evaluate the relative influence of different reinforcer dimensions (e.g., rate, quality, effort, reinforcement delay) on the choice behavior of overweight and obese children, as well as to identify relevant combinations of these dimensions to promote self-control.

METHODS

RESULTS

For each child being evaluated the number of choices was quantified per competing dimension for each of the evaluation conditions throughout the trials. The individual responses of each participant are presented in table I. The proportion of choices for the option associated with a dimension is presented on the left side, while the selections for the competing dimension are presented on the right side. For example, at baseline, when the evaluation condition was the direct competition of high-quality against low-quality reinforcement (HQvsLQ), child 3 assigned 60 % of his choices to healthy foods associated with higher-value stickers, while the remaining 40 % were assigned to unhealthy foods associated with lower-value stickers. After the intervention, the same child assigned 100 % of his choices to the healthy foods associated with the high-quality reinforcer. In this sense, the sensitivity to the quality of the reinforcer was more acute after the intervention. In the quality against immediacy evaluation, one option consisted of unhealthy food associated with the immediate delivery of a low-quality reinforcer, while the other option consisted of healthy food associated with the delayed delivery of a high-quality reinforcer. In this condition, the same child assigned 60 % of his choices to the condition associated with the higher-quality delayed reinforcer (exhibiting self-control), while he assigned the remaining 40 % of his choices to the option with the immediate delivery of the lower quality reinforcer. Subsequent to treatment, the child assigned 100 % of his choices to healthy food associated with the highest-quality reinforcer, despite a delay in access to the reinforcer of up to 24 h.

Table I. Percentage distribution of choices along reinforcer dimensions for evaluation conditions and experiment phases 

HQ: high quality; LQ: low quality; I: immediacy; D: delayed; HE: high effort; LE: low effort; HR: high reinforcement rate; LR: low reinforcement rate; CvsI: quality versus immediacy; QvsE: quality versus effort; QvsR: quality versus reinforcement rate; IvsE: immediacy versus effort; IvsR: immediacy versus reinforcement rate; IvsE: effort versus reinforcement rate.

The average percentages and standard deviations for each condition across the four reinforcer dimensions, which indicate the changes related to the sensitivity to reinforcer dimensions and their effects on food choice, are shown in figure 1.

Figure 1. Average percentages and standard deviations of choice allocated to response alternatives across assessment conditions (HQ: high quality; LQ: low quality; I: immediacy; D: delayed; HE: high effort; LE: low effort; HR: high reinforcement rate; LR: low reinforcement rate; QvsI: quality versus immediacy; QvsE: quality versus effort; QvsR: quality versus reinforcement rate; IvsE: immediacy versus effort; IvsR: immediacy versus reinforcement rate; IvsE: effort versus reinforcement rate). 

In the upper panel, the quality of the reinforcer is shown, and the relative preferences (before and after) are plotted for the baseline and evaluation conditions. The choice of quality exhibited high variability among participants before treatment; that is, they had a greater number of impulsive behaviors before treatment. In contrast, after treatment, most participants preferentially assigned their answers to the quality of the reinforcer as opposed to the dimensions in competition, showing greater self-control. With regard to immediacy, the same effect as with quality was observed but in favor of self-control; that is, a preference for other dimensions rather than immediacy of reinforcement was shown. In the case of effort and reinforcement rate, the effect was clearer when competing against quality. Among immediacy, reinforcement rate and effort, however, the effect varied across participants.

After the delayed reinforcement procedure all frequencies of the total selections of each reinforcer dimension were graphed (see Fig. 2). As shown, all dimensions showed differences before and after the intervention.

Figure 2. Comparisons before and after the intervention with regard to the children's choices for each reinforcer dimension. 

To identify whether these differences were statistically robust, a nonparametric Wilcoxon test with repeated measures was conducted to determine the existence of differences before and after the procedure with regard to the total number of selections in each reinforcer dimension. The results of the analysis are shown in table II. Furthermore, comparative analyses were performed before and after the intervention to ensure that the process of self-control development led to changes in the number of impulsive behaviors and a devaluation of the quality reinforcer of a higher value; a nonparametric Wilcoxon test was used for each evaluation condition. The results are shown in table III.

Table II. Comparisons before and after the implementation of the development of self-control process in the children's selections for each dimension 

Table III. The nonparametric Wilcoxon analysis of the differences in number of self-control conditions in the evaluation of impulsive behaviors and quality devaluation selections 

DISCUSSION AND CONCLUSION

Our results showed that children with overweight or obesity exhibited changes in their preferences. For example, children who chose low-quality but immediate reinforcers with less effort or with high reinforcement rates (associated with unhealthy food) changed their choice to higher-quality reinforcers associated with healthy food despite the additional delay, and this change involved greater effort or a lower reinforcement rate.

These results are similar to those described in the literature (²³,²⁸,³¹). For example, the children's selections showed a greater reinforcement delay even with regard to the dimensions that were not manipulated directly via the fading process. Importantly, however, many children associated some of their responses with unhealthy foods at baseline, perhaps because of their initial bias for high-calorie foods.

In general, and for most participants, the fading strategy led to a change in preferences for the conditions in which the gain was a reinforcer of higher quality despite its delay (24 hrs). During the intervention phase, participants generally showed changes in their choices of unhealthy food over healthy food, as well as with regard to the dimensions during the intervention phase (immediacy, reinforcement program and effort), through the gradual increase in reinforcer delay (²⁴,³²).

According to the results collected during the follow-up and post-evaluation phases, the participants maintained their preferences for healthy food and showed self-control with regard to the dimension in which they were trained, even for those that were not manipulated directly via the delayed reinforcement procedure (²³).

The current study extends the research on self-control in humans with the use of procedures of choice influenced by reinforcer dimensions among overweight and obese populations. The current data extend the generalization of tasks and procedures to the development of self-control as well as to the problems of overweight and obesity, which are associated with low impulse control (⁸,³³) and the devaluation of delayed rewards (³⁴).

In terms of applicability, the present investigation started with the assumption that the value of a reinforcer is relative to the options of reinforcement that are available and the reinforcer dimensions associated with each option. This position is easily applicable to feeding behavior, which is viewed as a choice that depends on the motivational state of the individual and the availability of food. In the current scenario, a wide variety of reinforcement options are modeled between consumption and synergy, with the food industry offering (among other things) foods with high calorie content (Q), foods that are easily accessible (E), foods offered more frequently (R), and foods that are ready to eat (I). However, sensitivities to these dimensions are susceptible to modification via behavioral procedures such as fading. The process of fading of the reinforcer delay or of the value of the competing dimension is one example of how an individual's experience can determine his or her subjective estimation of time and organize his or her choices in favor of self-control or impulsivity, in the sense that individuals with little exposure to delayed conditions with a high probability of reinforcement tend to behave more impulsively. In the case of healthy food options the same happens: children have more immediate access to the relatively small benefits of unhealthy food and have little experience with the delayed benefits of consuming healthy food. Therefore, their estimate of the sizes of the reinforcer or the response dimensions accompanying food are biased in favor of junk food, as was observed in the baseline results of this study before exposure to the associated reinforcers and fading procedure. Consistent with the results of previous research (³⁵,³⁶), overweight might be a consequence of the way people discount future health benefits.

In this regard, considering the impulsivity of children with overweight and obesity as a choice behavior, where the relative value of a high-quality reinforcer is devalued by delay, effort and reinforcement rate, enables the operationalization of part of the problem and thereby quantifies the progress achieved in scenarios with multiple enhancement options. To the extent that insensitivity to delayed consequences contributes to the emergence and maintenance of these behavioral problems, an understanding of the variables that control this temporary reduction is critical for the development of effective treatments (³⁷).