Elsevier

Appetite

Volume 49, Issue 1, July 2007, Pages 66-73
Appetite

Research report
The influence of trait and induced state impulsivity on food intake in normal-weight healthy women

https://doi.org/10.1016/j.appet.2006.11.008Get rights and content

Abstract

Previous research has linked overeating and overweight/obesity to impulsivity. To find out whether impulsivity causes overeating and hence overweight and obesity, we attempted to prime the concept of impulsivity in healthy participants. In a within-subjects design one sample participated in two conditions. In both conditions participants did a priming task that either hinted subtly at the concept of impulsivity or that was neutral in content. Each time the priming task was followed by a bogus taste test. Trait impulsivity was measured by means of a behavioural task and self-report. Firstly, we hypothesized that participants would eat more during the taste test after they had been primed with the concept “impulsivity” compared to after the control session. Secondly, we expected that a more impulsive personality would predict a heightened food intake. Thirdly, we expected that impulsivity would predict food intake better than restraint. The results showed that both the self-report measure of impulsivity and the behavioural task predicted food intake. Restraint did not significantly predict food intake. Primed impulsivity did not increase food intake, possibly because the priming effect did not last long enough.

Introduction

The World Health Organization (WHO) has declared obesity a global epidemic (WHO, 2003). This epidemic brings with it a rise in chronic diseases like diabetes and cancers (Hu, 2003). Quality of life from a psychological and social point of view is in many cases also seriously affected (Karlsson, Taft, Sjöström, Torgerson, & Sullivan, 2003). Besides the physical, psychological and social consequences for the individual, there are the financial consequences for society. In the United States an estimated 9% of medical expenditures are attributable to overweight and obesity (Flegal, 2005). It is clear that something needs to be done to stop and even reverse this obesity epidemic. For this reason much research has been done on causal and maintaining factors of obesity. Genetic, biological, psychological and sociocultural factors have been suggested (Drewnowski, 1991).

One such psychological factor is impulsivity. Generally impulsivity is defined as the tendency to think, control and plan insufficiently. In most cases this results in an inaccurate or maladaptive response (Solanto et al., 2001). Impulsivity is considered a multidimensional construct (Wingrove & Bond, 1997) for two reasons. First, correlations between self-report measures and behavioural measures of impulsivity are generally weak. This could mean that self-report impulsiveness questionnaires measure a different aspect of impulsivity than behavioural impulsivity tasks do. Second, even within the behavioural tasks different operationalizations and explanatory models of impulsivity are used and these often intercorrelate poorly.

Roughly, one could say that there are three main aspects of impulsivity. The first is impulsiveness, defined as “acting on the spur of the moment without being aware of any risk involved” (Eysenck, Eating, & Pearson, 1984, p. 315) and measured by self-report. The second is response inhibition, which is predominantly measured by behavioural tasks. Barkley (1997) suggests that inhibitory control is necessary for executive functions such as self-regulation. If response inhibition is disturbed, then the very first self-regulatory act of inhibiting the prepotent response already fails, with all its consequences. The third aspect of impulsivity is sensitivity to reward, which is measured both by behavioural tasks and by self-report questionnaires. People who are sensitive to reward detect more rewarding stimuli and are more likely to approach these rewarding stimuli (Avila, 2001). It should be noted that impulsiveness and response inhibition, but not reward sensitivity, are the focus of the current study.

One can imagine that all three aspects of impulsivity contribute to the obesity epidemic (Davis, Levitan, Smith, Tweed, & Curtis, 2006). First, when one often acts on the spur of the moment, momentary craving might be more important than future goals of losing weight. Second, when one is confronted with palatable food, the prepotent response is to eat it. In times when food was scarce this was an adaptive response (Blundell & Gillett, 2001). However, it is feasible that in today's obesogenic environment not being able to inhibit one's prepotent responses contributes significantly to the problem of obesity. Third, sensitivity to reward might lead people make the wrong food choices: they might prefer foods that are sweet and fat because palatable food has a greater rewarding value than bland food (Davis et al., 2007).

Research has begun to demonstrate a link between overeating and impulsivity, measured with self-report, with response inhibition tasks and with sensitivity to reward tasks. Nasser, Gluck, and Geliebter (2004) have shown that Binge Eating Disorder (BED) patients score significantly higher on a self-report measure of general impulsiveness compared to controls. They also found positive significant correlations (∼0.50) between participants’ impulsivity score and the BED criteria “Loss of control during a binge” and “Eating when not physically hungry”. Nederkoorn, Braet, Van Eijs, Tanghe, and Jansen (2006) found that obese children were impaired in general response inhibition as measured by the stop-signal paradigm compared to control participants. In other words, even on a very basic motoric level that has nothing to do with food the obese children were less able to inhibit their responses. Moreover, impulsivity turned out to be an obstacle in the treatment of the obese children: the children that were worst at inhibiting responses, lost less weight (Nederkoorn, Braet, et al., 2006; Nederkoorn, Jansen, Mulkens, & Jansen, in press).

Nederkoorn, Braet, et al. (2006) also found that the obese children were more sensitive to reward during a behavioural task compared to control children. Other research has shown that obese children have difficulties with delay of gratification tasks only when the incentive is edible (Bonato & Boland, 1983). This suggests that food could be especially rewarding for the obese. This has indeed been found. In a study by Saelens and Epstein (1996) obese participants chose to work for food instead of sedentary activities more often than controls. In sum, overweight and obese people seem more sensitive to reward and less adequate at the inhibition of prepotent responses, especially when it comes to resisting palatable food.

Recently it has been shown that even in healthy, lean participants impulsivity is of importance when it comes to food. Healthy individuals who are more sensitive to reward according to a self-report questionnaire turned out to have more pronounced neural responses to images of appetizing food (Beaver et al., 2006). This could indicate that for high-impulsive people it is harder to resist food than for low-impulsive people. Indeed, Guerrieri, Nederkoorn, and Jansen (2007) found that high-impulsive women ate more during a bogus taste test than their low-impulsive peers when impulsivity was measured by self-report, but not by the stop-signal paradigm. The high-impulsive women also scored significantly higher on the Eating Disorders Examination Questionnaire (EDE-Q; Fairburn & Beglin, 1994). This indicates that the high-impulsives exhibited more “eating disordered” cognitions and behaviours compared to the low-impulsives, although the clinical cut-off scores for eating disorders were not reached.

In conclusion, impulsivity in all its forms has repeatedly been linked to overweight and obesity. Although there are exceptions (see Guerrieri et al., 2007) impulsivity seems to be linked to reactions to palatable food, to “eating disordered” cognitions and to overeating in healthy participants. The relationship between increased impulsivity and the problem of overeating is quite robust since it generally persists even when impulsivity is measured in different ways (self-report versus behavioural tasks; response inhibition versus sensitivity to reward). However, based on previous research one cannot draw conclusions on causality. The studies that have been conducted are mostly of a correlational or quasi-experimental nature. This makes it impossible to rule out a third variable that influences impulsivity as well as overeating. If one wishes to conclude that impulsivity causes a heightened food intake, one needs to manipulate impulsivity experimentally. One should randomly assign healthy subjects to one of two groups: an experimental group in which impulsivity is manipulated experimentally versus a control group. If the experimental group shows a heightened food intake during a bogus taste test, then one could rightfully conclude that increased impulsivity caused the heightened food intake. We could find only one study in which similar methodology was used. Rotenberg et al. (2005) primed “lack of control” thoughts and these thoughts did indeed lead to greater food intake compared to priming “control” thoughts. This suggests that impulsivity, which shows overlap with the construct of “lack of control”, might indeed promote overeating in a sample of healthy participants. However, no neutral condition was included, so it is not clear whether “control” cognitions led to less food intake, whether “lack of control” cognitions led to increased food intake, or both.

Because of the importance of this type of experimental research, which is the only way to investigate the causal relation between impulsivity and overeating, we tried to generalize the results of Rotenberg et al. (2005). We included a neutral control condition to avoid the issues in the Rotenberg study. We attempted to manipulate impulse control experimentally by means of a priming task that differed from the task that Rotenberg et al. (2005) used. Priming is a frequently used method in social cognition research. It is applied to enhance the cognitive availability of a psychological construct. Research has shown that priming procedures affect people's judgments and behaviour (Bargh, Gollwitzer, Lee-Chai, Barndollar, & Trötschel, 2001; Higgins, Rholes, & Jones, 1977). In the current study priming was used to manipulate the accessibility of the impulsivity construct in the hope that this priming would affect eating behaviour. In a within-subjects design one sample participated in two conditions. In both conditions participants did priming task that either hinted subtly at the concept of impulsivity or that was neutral in content. After the priming task participants did a bogus taste test in both conditions. Trait impulsivity was measured by means of a behavioural task and self-report.

We hypothesized that participants would eat more during the taste test after they had been primed with the concept “impulsivity” compared to after the control session. In light of the findings of Guerrieri et al. (2007) it was also expected that a more impulsive personality would predict a heightened food intake. It is possible that trait impulsivity interacts with the priming task. In other words, high-impulsive people could react differently to the priming of impulsivity compared to low-impulsive people. However, we had no specific expectations concerning the presence of this interaction or its specific form.

Restraint (Herman & Polivy, 1980) or the extent to which people diet (mostly without success) and worry about their weight, is another important variable to consider when looking at eating behaviour. It has been shown repeatedly that, compared to unrestrained eaters, restrained eaters consume more food after having consumed a high-caloric preload. This is called “counterregulation” (Herman & Polivy, 1980). This counterregulation is thought to occur because the consumption of the preload acts as a disinhibitor: it removes inhibitions and eliminates restraint. It turned out that it is not even necessary to consume food in order to disinhibit restrained eaters. Mere exposure to food stimuli is sufficient to break down dietary restraint (Jansen & van den Hout, 1991; Rogers and Hill, 1989). In other studies participants who were identified as restrained eaters were found to eat more than unrestrained eaters when they were given ad libitum access to palatable food, not preceded by a preload or cue exposure (e.g., Jansen, 1996). Moreover, Nederkoorn, Van Eijs, and Jansen (2004) found that restrained eaters were worse at prepotent response inhibition, i. e. more impulsive, compared to controls. In combining these findings we can conclude that the heightened food intake in restrained eaters may be a consequence of their increased impulsivity. Hence, we hypothesized that in the current study impulsivity would predict food intake better than restraint.

In sum, we had three hypotheses: (a) participants will eat more during the taste test after they have been primed with the concept “impulsivity” compared to after the control session, (b) a more impulsive personality will predict a heightened food intake, (c) impulsivity will predict food intake better than restraint. Moreover, we explored whether trait impulsivity would interact with the priming of impulsivity.

Section snippets

Participants

Forty-two normal-weight female students were recruited to participate in a study on “taste perception”. Four participants were excluded because they missed one or more sessions. This left us with 38 participants (mean age: 19.3±1.3 years; mean BMI: 22.2±2.9). They received course credit or a monetary reward for their participation. All participants were debriefed by e-mail.

Materials

The Scrambled Sentences Task (Srull & Wyer, 1979) is a priming method that originates from social psychology. To our

Pre-existing differences and neophobia effect

Paired-samples t-tests indicated that there were no pre-existing differences in hunger and mood between the neutral and the experimental session (see Table 1). We did find a neophobia effect: a within-subjects ANOVA indicated that at least one session differed from the remaining sessions as far as food intake was concerned, F(2, 74)=8.28, p<0.01. Post-hoc analyses with Bonferroni correction (α divided by 3: α=0.017) indicated that the food intake during session 1 was significantly lower

Discussion

In accordance with our previous study (Guerrieri et al., 2007) we found that trait impulsivity predicts heightened food intake in normal-weight healthy women. This was the case for both measures of impulsivity: self-report and behavioural. The low correlation between both sorts of measures has been found in other studies (Wingrove & Bond, 1997) and it might indicate that self-report and behavioural measures of impulsivity measure different aspects of impulsivity in motor, cognitive, social and

References (43)

  • C. Nederkoorn et al.

    Restrained eaters act on impulse

    Personality and Individual Differences

    (2004)
  • C. Nederkoorn et al.

    Why obese children cannot resist food: The role of impulsivity

    Eating Behaviors

    (2006)
  • J. Overduin et al.

    Conditioned insulin and blood sugar responses in humans in relation to binge eating

    Physiology and Behavior

    (1997)
  • A. Roefs et al.

    The effects of information about fat content on food consumption in overweight/obese and lean people

    Appetite

    (2004)
  • P.J. Rogers et al.

    Breakdown of dietary restraint following mere exposure to food stimuli: Interrelationships between restraint, hunger, salivation, and food intake

    Addictive Behaviors

    (1989)
  • K.J. Rotenberg et al.

    Effects of priming thoughts about control on anxiety and food intake as moderated by dietary restraint

    Appetite

    (2005)
  • B.E. Saelens et al.

    Reinforcing value of food in obese and non-obese women

    Appetite

    (1996)
  • J. Wingrove et al.

    Impulsivity: a state as well as a trait variable. Does mood awareness explain low correlations between trait and behavioural measures of impulsivity?

    Personality and Individual Differences

    (1997)
  • A.N. Agranad-Meged et al.

    Childhood obesity and attention deficit/hyperactivity disorder: A newly described comorbidity in obese hospitalized children

    International Journal of Eating Disorders

    (2005)
  • T. Araya et al.

    Reducing prejudice through priming of control-related words

    Experimental Psychology

    (2002)
  • C. Avila

    Distinguishing BIS-mediated and BAS-mediated disinhibition mechanisms: A comparison of disinhibition models of Gray (1981, 1987) and of Patterson and Newman (1993)

    Journal of Personality and Social Psychology

    (2001)
  • Cited by (169)

    View all citing articles on Scopus
    View full text