Secondary Logo

Share this article on:

Providing Choice in Exercise Influences Food Intake at the Subsequent Meal

BEER, NATALYA J.; DIMMOCK, JAMES A.; JACKSON, BEN; GUELFI, KYM J.

Medicine & Science in Sports & Exercise: October 2017 - Volume 49 - Issue 10 - p 2110–2118
doi: 10.1249/MSS.0000000000001330
APPLIED SCIENCES

The benefits of regular exercise for health are well established; however, certain behaviors after exercise, such as unhealthy or excessive food consumption, can counteract some of these benefits.

Purpose To investigate the effect of autonomy support (through the provision of choice) in exercise—relative to a no-choice condition with matched energy expenditure—on appetite and subsequent energy intake.

Methods Fifty-eight men and women (body mass index, 22.9 ± 2.3 kg·m−2; peak oxygen consumption, 52.7 ± 6.4 mL·kg−1·min−1) completed one familiarization session and one experimental trial, in which they were randomized to either a choice or no-choice exercise condition using a between-subjects yoked design. Ad libitum energy intake from a laboratory test meal was assessed after exercise, together with perceptions of mood, perceived choice, enjoyment, and value.

Results Despite similar ratings of perceived appetite across conditions (P > 0.05), energy intake was significantly higher after exercise performed under the no-choice condition (2456 ± 1410 kJ) compared with the choice condition (1668 ± 1215 kJ; P = 0.026; d = 0.60). In particular, the proportion of energy intake from unhealthy foods was significantly greater after exercise in the no-choice condition (1412 ± 1304 kJ) compared with the choice condition (790 ± 861 kJ; P = 0.037, d = 0.56). Participants in the choice condition also reported higher perceptions of choice (P < 0.001), enjoyment (P = 0.008), and value (P = 0.009) relating to the exercise session, whereas there were no between-condition differences in mood (P > 0.05).

Conclusions A lack of choice in exercise is associated with greater energy intake from “unhealthy” foods in recovery. This finding highlights the importance of facilitating an autonomy supportive environment during exercise prescription and instruction.

School of Human Sciences, The University of Western Australia, Perth, Western Australia, AUSTRALIA

Address for correspondence: Kym Guelfi, Ph.D., School of Human Sciences (M408), The University of Western Australia, 35 Stirling Highway, Crawley WA, 6009, Australia; E-mail: kym.guelfi@uwa.edu.au.

Submitted for publication March 2017.

Accepted for publication May 2017.

It is well established that regular exercise is associated with numerous health benefits (40); however, certain behaviors after exercise, such as unhealthy snacking or excessive food consumption, can counteract some of these benefits (3). Indeed, research indicates that postexercise food consumption is highly variable across individuals and situations (36). Accordingly, research is needed to explore what factors influence the extent to which individuals alter their food consumption in the aftermath of an exercise session and to determine how the nature of an exercise session might be modified to benefit subsequent food choice.

The effect of an acute bout of exercise on subsequent appetite and energy intake appears to be influenced by numerous factors. For instance, a growing body of evidence suggests that the physiological demands of exercise in relation to the mode (1), duration (11), and intensity (37) may influence appetite regulation in recovery. Furthermore, there is some evidence to suggest that psychological factors associated with exercise may also influence postexercise food intake (12,41). More specifically, recent conceptual work has highlighted the potential for exercise motivation to influence postexercise eating behavior (9). These researchers suggested that postexercise eating behavior may be influenced by the extent to which exercise is experienced as autonomous (i.e., characterized by a sense of value, alignment with one’s identity, and/or enjoyment) or controlled (i.e., characterized by feelings of internal or external pressures). Their rationale was based partly on evidence linking the experience of autonomy or control to different responses in subsequent activities requiring self-control (27,34). However, the authors discussed their rationale more broadly in terms of three overlapping categories—the facilitation of conscious reflective licensing, nonconscious impulsive processes, and/or as a result of physiological responses (9).

Support for the notion that motivation for exercise may influence subsequent food consumption comes from a study by Werle and colleagues (41). These researchers found that participants were more likely to seek snacks after reading about “tiring” physical activity as opposed to reading about a “fun” physical activity. The notion of “fun” is directly associated with intrinsic motivation—the most autonomous form of motivation (7)—whereas the concept of “tiring” activity is more closely aligned with controlled motivation (32). In other work, Fenzl and colleagues (12) found that individuals who self-imposed physical activity (i.e., exhibited controlled motivation) were more prone to consume a food reward postexercise compared with individuals who possessed more self-determined regulation (i.e., autonomous motivation) for exercise. Although this finding suggests a potential role for exercise motivation to influence postexercise energy intake, no researchers have directly manipulated the social conditions that have been found to influence exercise motivation with the goal of examining the subsequent effect on appetite and energy intake. Nevertheless, evidence from other contexts suggests that autonomy in a task (which is integral to the formation and sustainment of autonomous motivation) can be manipulated, and that such manipulations can influence food consumption (23). Specifically, Magaraggia and colleagues (23) found that students snacked on more glucose-rich food (jelly beans) during a learning task that had been framed in controlling conditions (i.e., when students were given no choice over the task), compared with students completing the same learning task under more autonomy supportive conditions in which choice over learning content was offered. Given that the nature of an exercise session may influence postexercise food intake, the aim of this study was to investigate whether manipulating choice during an acute bout of exercise would influence appetite and subsequent energy intake relative to participation in a controlling (no choice) session. It was hypothesized that individuals exercising under conditions with limited choice would consume more total energy at the postexercise meal and make “unhealthier” food choices (i.e., consume significantly more “unhealthy” food) compared with individuals who exercise under autonomy supportive conditions of choice.

Back to Top | Article Outline

METHODS

Participants

A total of 58 healthy participants (n men = 38; n women = 20, age: 22 ± 4 yr) met the inclusion criteria and completed this study. The study was approved by the Institutional Human Ethics Committee and written informed consent was obtained from all participants; however, to minimize the potential for biased responses, participants were not informed that their food intake was being assessed in the study. Instead, they were informed that the aim of the study was to investigate the effect of motivation for exercise on the psychological and physiological markers of stress and inflammation in response to an acute bout of exercise. All participants were debriefed as to the true purpose of the study upon completion of data collection.

Back to Top | Article Outline

Experimental Design

Using a between-subject yoked design, each participant was required to attend two separate laboratory sessions. The first visit was a familiarization session, which included baseline assessments of motivational orientations towards exercise, current exercise behaviors, psychological reactance, and eating habits. After this session, participants were pair matched based on sex, age (±5 yr), peak oxygen consumption ((V˙O2peak) ±5 mL·kg−1·min−1), body mass (±5 kg), and height (±10 cm). Participants within each pair were then randomly allocated (using random number generator software) into either a choice or no choice exercise condition, after which participants returned to the laboratory for completion of their assigned experimental session. Participants randomized to the choice condition were provided with choice of exercise mode (bike or treadmill), exercise intensity, duration of exercise (30–60 min), the time of commencement of the session (0600–0900 h), and the type of music played during exercise, as well as being provided with verbal cues indicative of “choice” (such as the ability to change their preferences throughout). Participants assigned to the no-choice condition were also asked to select their preferences for the exercise session, before being explicitly told that, despite their preferences, they would not have choice over their exercise experience and instead would be required to complete their session under the parameters (i.e., mode, intensity, duration, background music, time of commencement) chosen by their matched partner. The effect of this choice manipulation on subsequent perceived appetite, food choices, and overall food intake, together with subjective vitality and experiences of the exercise bout, was assessed. Women were tested in the follicular phase of the menstrual cycle (day 8 ± 3) given the well-established effect of menstrual cycle on appetite and energy intake (10).

Back to Top | Article Outline

Familiarization Session

Participants completed a series of questionnaires to assess current exercise behaviors, motivation toward exercise, psychological reactance, compensatory eating, and dietary restraint. The Godin Leisure Time Exercise Questionnaire (GLTEQ) (16) was used to assess current exercise participation in a typical week, and motivational orientation toward exercise was assessed using the Behavioral Regulation in Exercise Questionnaire-3 (BREQ-3) (24). Briefly, the BREQ-3 comprises 24 items that are rated on a five-point response scale anchored at 0 (not true for me) and 4 (very true for me), and is designed to assess levels of amotivation, external regulation, introjected regulation, identified regulation, integrated regulation, and intrinsic motivation toward exercise. Psychological reactance was measured using the Hong Psychological Reactance Scale (20). This instrument, which was used to measure whether participants’ motives to resist threats to their freedom differed before our choice manipulation, includes 10 items that are rated on a five-point Likert-type scale anchored at 1 (strongly disagree) and 5 (strongly agree). Sample items include: “I become angry when my choice is restricted” and “I normally resist the attempts of others to influence me.” Self-reported motivation for compensatory eating was measured with the 15-item Compensatory Eating Motives Questionnaire (28). Items in this scale are rated on a four-point Likert-type scale anchored at 1 (never) and 4 (always), which represent three underlying compensatory orientations (i.e., reward, relief, recovery). Finally, dietary restraint was assessed using the restraint scale of the Revised Three-Factor Eating Questionnaire (6), with participants excluded if they scored >3.5 to minimize any influence on the measurement of laboratory energy intake.

After completion of the baseline questionnaires, V˙O2peak was measured using a graded exercise test on a treadmill. This test consisted of 3-min stages with the speed of the treadmill progressively increased at each stage, until voluntary exhaustion was reached. During the test, participants wore a HR monitor (Polar, Kempele, Finland) and breathed through a mouthpiece connected to a computerized gas analysis system (Meta 2000; University of Western Australia, Perth, Australia). This system consisted of a ventilometer (Universal ventilation meter; VacuMed, Ventura, CA), which measured the volume of inspired air, as well as oxygen and carbon dioxide analyzers (Ametek Applied Electrochemistry S-3A/1 and CD-3A, Oak Ridge, TN; AE1 Technologies, Pittsburgh, PA), which measured the percentage of oxygen and carbon dioxide in the expired air. The gas analyzers were calibrated before each test using a beta standard reference gas, whereas the ventilometer was calibrated using a 1-L syringe, as per manufacturer’s specification. Individual V˙O2peak was determined from the highest minute average during the test. In addition, the HR and V˙O2 data for the final minute of each 3-min workload for each participant was later plotted to determine individual equations to represent the HR–V˙O2 relationship. This relationship was used to ensure that energy expenditure during exercise in the subsequent experimental session was matched between pairs (detailed later).

After completing the graded exercise test, participants provided a 35-μL capillary blood sample from the fingertip for the purpose of familiarizing them with the blood collection procedure before the experimental trial. During this blood collection period, participants were asked about their “typical breakfast” and “favorite food treats.” Their responses were recorded and used to guide what foods would be provided in the experimental trial, so as to ensure that the foods presented were liked by, and familiar to, the participants.

Back to Top | Article Outline

Experimental Sessions

Participants arrived at the laboratory between 0600 h and 0900 h after an overnight fast having been instructed to consume 300 mL of water upon waking. The specific time of arrival within this period was chosen by the participants randomized to the choice condition, whereas paired participants in the no-choice condition were instructed to attend at the time chosen by their matched partner. Upon arrival to the laboratory, baseline measures of perceived appetite, mood, subjective vitality, blood glucose, and blood lactate were obtained (for more information, see Experimental Measures section). Next, preferences for the exercise session in relation to the mode (choice of treadmill or stationary cycling), duration (choice of time between 30 and 60 min), and music station (choices included hip hop/rap, jazz, country pop, current hits, etc.; Pandora, Oakland, CA) were recorded.

All participants reported their preferences under the pretense that they would be exercising as per their choices; however, individuals in the no-choice condition were then informed that they had been allocated a “partner” in the choice group, and that they would be performing under the parameters chosen by their partner. All participants were given instructions on how to operate the equipment they had chosen or had been assigned (stationary bicycle or treadmill); however, those who were provided with choice were reminded that they were able to change their choices at any time and exercise at a self-selected intensity, whereas the no choice group were instructed of the specific HR zone they were expected to achieve throughout their specified exercise session which was predetermined to ensure that exercise intensity (and the resulting energy expenditure) was standardized within pairs. More specifically, the HR achieved by each individual in the choice condition (where exercise intensity was self-selected) was used to determine the associated rate of oxygen consumption based on the individualized HR–V˙O2 relationship determined in the initial familiarization session. Then, the HR needed to achieve an equivalent rate of oxygen consumption for each partner randomized to the no-choice exercise condition was calculated using their individualized HR–V˙O2 relationships to ensure that oxygen consumption and energy expenditure was similar between pairs. RPE (4) were collected at 15-min intervals and at the end of each participant’s exercise session. All instructions were delivered using a prepared script to standardize any experimenter–participant interactions during the experiment.

Immediately after the exercise, participants completed a series of questionnaires to assess perceived appetite, mood, subjective vitality, and subjective exercise experience, before a second blood sample was taken. While participants were completing these questionnaires, the experimenter placed a series of breakfast foods on display in a predetermined position, which was standard across participants. Individuals were then informed that the experimenter would leave the room for 20 min before a final blood sample was needed and were invited to help themselves to the food while they waited. After the 20-min period, a final blood sample was taken and participants completed a final set of questionnaires to assess appetite, as well as health and taste ratings of the foods to which they had just been exposed. Finally, participants were probed for suspicion by answering the single item, “In your own words, please describe your interpretation of the purpose of this study.” This suspicion probe item was used to determine whether participants were aware of the true nature of the experiment.

Back to Top | Article Outline

Experimental Measures

Perceived appetite and energy intake

Perception of appetite was assessed using a modified visual analogue scale at baseline (preexercise), immediately postexercise, and after the laboratory test meal. This validated scale (13) takes the form of four straight lines (100 mm in length each) accompanied by a question anchored with words representing extreme states of fullness, hunger, desire to eat, and prospective food consumption at either end (e.g., “how hungry do you feel?” anchored by “not hungry at all” and “as hungry as I have ever felt”). Participants were required to make a vertical mark along the line, which was then measured to assess how they felt at that time. To disguise the purpose of the study to monitor appetite, three “filler” questions were included (e.g., “how tired do you feel”).

In addition to perceived appetite, energy intake was assessed using a laboratory test meal provided 10 min after completing the exercise bout. Participants were informed that the experimenter would leave the room for 20 min, and that they were welcome to help themselves to the food available, which was provided because the participants had not eaten, and as a thank you gesture for completing the experiment. The laboratory test meal consisted of products of known and differing macronutrient composition, including an assortment of typical breakfast foods and treats such as bread, spreads, cereal, milk, and biscuits. All food provided was weighed before participants’ arrival and reweighed after consumption. To minimize the influence of environmental factors on eating behavior (39), the investigator left the nutrition laboratory during consumption, the ambient temperature was controlled, and foods were provided in the same amount and presented in the same position for each trial. Water was not offered to participants during the exercise; however, a standardized bottle of plain drinking water (~1000 mL) was made available during the laboratory test meal. All participants finished eating before the experimenter returned to the room. To determine total energy intake and macronutrient breakdown, the postmeal weight was subtracted from the premeal weight of each food item. The amount of food consumed (g) was multiplied by the amount of kilojoules and macronutrient per gram within the product, as specified by the manufacturer’s nutrition facts label, or by FoodWorks software package (FoodWorks v 4.2.0; Xyris Software, Qld, Australia) where nutrition labels were not available. To classify foods as “healthy” and “unhealthy,” participants rated the foods they were exposed to after the laboratory test meal on a scale anchored at 1 (very unhealthy) and 7 (very healthy). Foods that scored on average below the midpoint of the scale (i.e., 3.5) were classified as “unhealthy” and vice versa for “healthy” foods. Foods that were classified as unhealthy were confectionary, croissants, muffins, jam, and chocolate biscuits (2.66 ± 1.15), and those considered healthy included fruits, whole-meal bread, low-fat milk, low-fat spreads, and Sanitarium Weetbix breakfast cereal (5.37 ± 0.69). These participant-derived classifications were verified by independent ratings obtained from a dietitian who was blind to the study purpose (i.e., the same foods were classified as healthy and unhealthy).

Back to Top | Article Outline

Blood glucose and lactate

Capillary blood (35 μL) was sampled at baseline, immediately postexercise, and after the laboratory test meal from a fingertip using a sterile lancet (Unistick 2 Normal; Owen Mumford, Oxford, UK). Blood glucose and lactate concentrations were measured for the baseline and postexercise samples using a blood gas analyzer (Radiometer, Copenhagen, Denmark). The purpose of the final blood sample was to keep participants in the laboratory for 20 min postexercise without arousing suspicion about the laboratory test meal.

Back to Top | Article Outline

Mood

Mood was assessed at baseline and immediately postexercise using the Profile of Mood States Adolescent Inventory (POMS-A) (38), which assesses six dimensions of mood: anger, confusion, depression, fatigue, tension, and vigor. Using the stem, “How do you feel right now?” participants responded to 24 statements on a response scale ranging from 0 (not at all) to 4 (extremely).

Back to Top | Article Outline

Vitality

The Subjective Vitality Scale (35) was used to assess the degree to which participants felt energized/vitalized before and after the exercise bout. Participants completed an instrument consisting of seven items on a seven-point Likert scale ranging from 1 (not at all true) to 7 (very true), with items such as “At this time, I have energy and spirit,” and “I feel energized right now.”

Back to Top | Article Outline

Perceptions of the exercise task

The Intrinsic Motivation Inventory (IMI) (33) was completed immediately postexercise to assess participants’ perceptions of the exercise experience. The 45-item questionnaire assesses participants’ enjoyment (seven items, e.g., “I enjoyed this very much”), perceived competence (six items, e.g., “I thought I was pretty good at this exercise”), effort (five items, e.g., “I put a lot of effort into this exercise”), choice (seven items, e.g., “I did this because I wanted to”), perceived value (7 items, e.g., “I believe doing this exercise could be of some value to me”), felt tension (five items, e.g., “I was anxious while exercising”) and relatedness (eight items, e.g., “I feel close to the experimenter”). A seven-point response scale was employed, anchored at 1 (not true at all) to 7 (very true).

Back to Top | Article Outline

Statistical Analyses

Statistical analyses were conducted using SPSS version 23.0 software package for Windows, with statistical significance being accepted at an alpha level of P < 0.05. To assess whether the background characteristics of participants randomized to the two conditions differed before the experimental manipulation, a multivariate analysis of variance (MANOVA) was conducted to compare age, body mass, height, body mass index (BMI), V˙O2peak, and baseline activity levels (i.e., GLTEQ scores). In a separate MANOVA, self-reported motivation toward exercise (i.e., across the six BREQ-3 subscales), restrained eating, compensatory eating motives, and psychological reactance were compared between conditions. Likewise, physiological responses to exercise (i.e., intensity, duration) were compared between conditions using one-way MANOVA. Blood variables were compared using two-way (condition × time) MANOVA. The effect of condition on subjective vitality was assessed using two-way (condition × time) ANOVA. Perceptions of the exercise session (i.e., IMI scores) were compared using MANOVA, whereas mood was assessed using two-way (condition × time) MANOVA. Perceived appetite was assessed using two-way (condition × time) MANOVA. The effect of choice on overall energy intake, as well as the intake from healthy and unhealthy sources, was assessed using two-way (sex × condition) ANOVAs due to the propensity for men to consume more calories than women. Further, Cohen d effect sizes (d) were calculated for pairwise comparisons of energy intake.

Back to Top | Article Outline

RESULTS

Participant characteristics

The descriptive characteristics of the sample are displayed in Table 1. Participants were well matched between groups, with a nonsignificant multivariate effect for condition on age, body mass, height, BMI, current physical activity levels, and fitness F(6, 51) = 0.10, P = 0.996, η2 p = 0.61. Likewise, a nonsignificant multivariate effect was observed for between-condition differences on restrained eating, compensatory eating motives, psychological reactance, and self-reported motivation toward exercise (i.e., across the six BREQ-3 subscales), F(9, 48) = 0.80, P = 0.616, η2 p = 0.13.

TABLE 1

TABLE 1

Back to Top | Article Outline

Physiological characteristics of exercise

The physiological characteristics of the exercise are shown in Table 2. Participants’ exercise sessions were well matched across the two conditions, with MANOVA revealing a nonsignificant multivariate effect of condition on HR, energy expenditure, RPE, and exercise duration, F(4, 52) = 0.96, P = 0.683, η2 p = 0.04. Likewise, there was no difference in the blood lactate or blood glucose response between groups, with no significant multivariate main effect for condition, F(2, 44) = 0.10, P = 0.90, η2 p = 0.01, and a nonsignificant interaction at the multivariate level between condition and time, F(2, 44) = 0.91, P = 0.409, η2 p = 0.04. However, there was a significant multivariate effect for time, F(2, 44) = 25.64, P < 0.001, η2 p = 51.28, with an increase in blood lactate from preexercise to postexercise, F(1, 45) = 48.05, P < 0.001, η2 p = 0.52. Similarly, the environmental temperature and humidity in the laboratory were well controlled for all sessions (18.7°C ± 0.9°C and 46.1% ± 5.8%, respectively), with MANOVA revealing no significant between-group mean difference, F(2, 55) = 0.41, P = 0.665, η2 p = 0.01.

TABLE 2

TABLE 2

Back to Top | Article Outline

Perception of the exercise session

Participants’ subjective experience of exercise, assessed by the IMI, is presented in Table 3. A MANOVA revealed a significant multivariate effect for condition, F(7, 50) = 9.56, P < 0.001, η2 p = 0.57, which was accounted for by differences in enjoyment, F(1, 56) = 7.54, P = 0.008, η2 p = 0.12), perceived value, F(1, 56) = 7.39, P = 0.009, η2 p = 0.12, perceived choice, F(1, 56) = 50.90, P < 0.001, η2 p = 0.48, and effort F(1, 56) = 5.96, P = 0.018, η2 p = 0.10, whereby those in the choice condition scored significantly higher than those in the no-choice condition on all variables. No differences emerged on competence, F(1, 56) = 1.74, P = 0.193, η2 p = 0.03, tension, F(1, 56) = 0.11, P = 0.739, η2 p = 0.11, or relatedness, F(1, 56) = 2.00, P = 0.162, η2 p = 0.035. With respect to subjective vitality, there was no significant main effect for condition, F(1, 56) = 2.12, P = 0.151, η2 p = 0.04, or condition–time interaction F(1, 56) = 1.33, P = 0.254, η2 p = 0.02; however, there was a main effect for time F(1, 56) = 23.98, P < 0.001, η2 p = 0.30, such that subjective vitality increased postexercise (5.2 ± 0.9) from baseline (4.7 ± 1.0), independent of condition.

TABLE 3

TABLE 3

Back to Top | Article Outline

Mood

The effect of exercise under choice and no-choice conditions on mood is shown in Table 4. There was no significant multivariate main effect for condition, F(6, 51) = 2.05, P = 0.076, η2 p = 0.19, and a nonsignificant multivariate condition–time interaction, F(6, 51) = 0.43, P = 0.857, η2 p = 0.05. However, there was a significant multivariate main effect for time, F(6, 51) = 8.61, P < 0.001, η2 p = 51.63, with a significant decrease in confusion, F(1, 56) = 9.22, P = 0.004, η2 p = 0.14, depression, F(1, 56) = 6.93, P = 0.011, η2 p = 0.11, fatigue, F(1, 56) = 7.95, P = 0.007, η2 p = 0.124, and tension, F(1, 56) = 16.03, P < 0.001, η2 p = 0.223, and an increase in vigor, F(1, 56) = 34.30, P < 0.001, η2 p = 0.38, after the completion of exercise, independent of condition.

TABLE 4

TABLE 4

Back to Top | Article Outline

Appetite and energy intake

For self-reported hunger, fullness, desire to eat, and prospective food consumption, we observed no significant multivariate effect for condition, F(4, 53) = 1.90, P = 0.125, η2 p = 0.13, or time, F(1, 53) = 1.88, P = 0.128, η2 p = 0.12, as well as no significant multivariate condition–time interaction, F(4, 53) = 1.62, P = 0.182, η2 p = 0.11 (results not shown). With respect to the laboratory test meal, 2 (sex) × 2 (condition) ANOVAs were undertaken due to the propensity for men to consume more calories than women. For overall energy intake, there was no significant condition–sex interaction, F(1, 54) = 0.04, P = 0.844, η2 p = 0.001; however, significant main effects emerged for condition, F(1, 54) = 5.49, P = 0.023, d = 0.60, and sex F(1, 54) = 8.03, P = 0.006, d = 0.80. More specifically, energy intake was significantly higher in the no-choice condition compared to the choice condition (no choice 2456 ± 1410 kJ; choice 1668 ± 1215 kJ; Fig. 1), and men had greater energy intake compared with women (men 2400 ± 1444 kJ; women, 1421 ± 1363 kJ).

FIGURE 1

FIGURE 1

When considering energy intake separately from “healthy” and “unhealthy” foods, there was no significant main effects for condition, F(1, 54) = 0.63, P = 0.431, d = 0.21, or sex, F(1, 54) = 0.94, P = 0.337, d = 0.28, as well as no significant condition–sex interaction, F(1, 54) = 0.025, P = 0.874, η2 p < 0.001, for healthy food consumption. For unhealthy food intake, there was no significant condition–sex interaction, F(1, 54) = 0.01, P = 0.912, η2 p < 0.001; however, significant main effects emerged for condition, F(1, 54) = 4.65, P = 0.036, d = 0.56, and sex, F(1, 54) = 6.77, P = 0.012, d = 0.75, whereby “unhealthy” energy intake was significantly greater in the no-choice condition (1412 ± 1304 kJ) compared with the choice condition (790 ± 861 kJ). Water consumption did not differ significantly between participants in the choice (177 ± 215 mL) and no choice (230 ± 312 mL; t (56) = −0.64, P = 0.527) conditions.

Back to Top | Article Outline

DISCUSSION

The primary aim of this study was to investigate whether manipulating a key component of autonomy support (i.e., choice) during an acute bout of exercise would influence subsequent (total, healthy, and unhealthy) energy intake. Analyses demonstrated that overall energy intake was significantly altered by manipulating choice in exercise for both men and women. More specifically, energy intake from “unhealthy” food sources—and accordingly, overall energy intake—was higher after exercise with limited choice. This finding suggests that the facilitation of an autonomy supportive environment, particularly relating to the provision of choice, may be an important consideration for exercise prescription and instruction, and may contribute to a decreased likelihood of unhealthy food consumption in the aftermath of an exercise session.

This study is the first to show that choice in exercise influences food intake at the subsequent meal, and in doing so, is the first investigation to include both healthy and unhealthy food options. The inclusion of different food “types” is noteworthy because it not only allowed us to examine between-condition differences on overall energy intake but it also enabled us to distinguish between effects on overall energy intake and specific food preferences. Although the former may be most important from a weight management perspective, the latter is also important from a health perspective. In the present study, total energy intake from the laboratory test meal was not only statistically different between conditions but it may also be argued that a difference of 788 kJ at this single meal may be clinically meaningful given the evidence in the literature to suggest that additional energy intake of 125 kJ·d−1 can lead to a small, consistent degree of positive energy balance, resulting in continuous gradual weight gain (19). More importantly though, the greater caloric intake from unhealthy food sources may have implications for a number of lifestyle diseases associated with the consumption of unhealthy, energy dense foods, such as cardiovascular disease and type 2 diabetes (30). With respect to perceived appetite, the lack of difference between conditions was not consistent with the significant difference in overall energy intake; however, researchers have previously reported that feelings of hunger do not necessarily reflect actual energy intake (25). Indeed, it is possible that appetite-related differences may have emerged if preferences for certain types of foods, or cravings for sweet or fatty foods, were assessed.

A number of mechanisms may assist in explaining the effect of choice in exercise on subsequent energy intake. Dimmock and colleagues (9) suggested that there are at least three (potentially overlapping) pathways through which controlled exercise motives/experiences may increase the likelihood of consuming pleasurable but unhealthy food during recovery: through reflective cognitive processes, impulsive cognitive processes, and/or as a result of physiological responses. It is difficult to determine exactly which mechanism (or mechanisms) may have been responsible for the effect we observed, given that our aim was not to identify the processes underpinning the effect (and so our data do not provide evidence relating to all possible mechanisms). However, with respect to possible reflective processes, the increased consumption of unhealthy food seen in the no-choice condition may be contributed to, at least in part, the activation of compensatory health beliefs—such as licensing. Licensing represents the belief that unhealthy behaviors (e.g., unhealthy food choices) are justified after engagement in a healthy behavior (e.g., exercise). Of relevance, Rabia and colleagues (31) suggested that compensatory health behaviors are more likely to be present when individuals experience controlled motivation for a task. Unfortunately, this mechanism could not be directly assessed in the current study because it may have alerted participants to the covert measurement of food intake. However, given that individuals in the controlled condition reported experiencing significantly lower levels of perceived choice, the potential contribution of this mechanism should be explored in future research.

Although reflective justifications for food consumption (such as compensatory beliefs) develop via conscious processes, there is also a body of literature that focuses on nonconscious processes and their influence on eating behavior. According to the strength model of self-control (2), exertion of self-control (which is considered a limited resource) results in a subsequent state of “ego-depletion,” which consequently reduces the ability to further override natural or habitual responses (such as consuming pleasurable but unhealthy foods). However, it is important to note that there were no differences in self-reported vitality (a measure of ego depletion) between exercisers in the choice and no-choice conditions in the present study. This nonsignificant difference was somewhat surprising given the well-established link between autonomous motivation for exercise and vitality (29). Meanwhile, ratings of effort based on the IMI were higher in the choice condition compared with the no-choice condition (although RPE was similar during exercise). This suggests that greater perceived effort was unlikely to have contributed to the increase in energy intake observed here. Although it remains unclear whether reflective justifications for food consumption (such as licensing) facilitated the increased consumption of unhealthy foods, these findings suggest that nonconscious processes (such as ego-depletion) were unlikely to contribute to the results seen in our study.

Another possible mechanism through which autonomy in exercise may influence postexercise consumption is via changes to mood. Indeed, there is some evidence to suggest that motivation for exercise may influence mood (21,22), which in turn may influence food intake (14). However, this explanation appears unlikely to have played a role in the present study given that there were no differences in mood between the two conditions. The discrepancy between these results and those in the literature may be related to the large majority of negative mood items on our mood questionnaire (20/24)—thus, it is possible that positive moods not captured by the POMS-A might have been influenced by our choice manipulation.

Aside from the psychological processes that may link exercise experiences to food intake, it is also possible that exercising in an autonomy-supportive environment may have consequences for the physiological regulation of appetite (9). For instance, there is some evidence to suggest that engaging in tasks requiring self-control (e.g., exercising under controlling conditions) may result in lower levels of blood glucose, which in turn may have implications for appetite (26). In the present study, we saw no difference in blood glucose between conditions, suggesting that this was unlikely to account for the relationship between autonomy and food intake observed here. Other factors that may influence postexercise energy intake include the energy expenditure of the exercise and the characteristics of the individual (15,18). Importantly, these factors were well controlled for in the present study with strict pair-matching of participants on key background and physiological variables (such as sex, BMI and V˙O2peak), and standardization of energy expenditure within pairs. Nonetheless, there may be additional factors that influence postexercise intake, such as habitual exercise patterns, particularly in relation to the time of day that exercise is performed which were not considered here. Regardless, further research is required to determine the mechanisms through which autonomous motivation for exercise affects subsequent energy intake, and whether these mechanisms operate alone or in combination with each other. Furthermore, researchers are encouraged to investigate whether autonomy in exercise influences the physiological regulation of appetite, given preliminary evidence in the literature to suggest that the circulating concentrations of appetite-regulating peptides may be altered by psychological mindset alone (5).

The inclusion of the IMI in the present study confirmed that our manipulation of choice was successful, with participants in the no-choice condition reporting significantly lower perceptions of choice than those randomized to the choice condition. We also observed significantly greater ratings of enjoyment, perceived value, and perceived effort for those in the choice group; a finding which may have implications for long-term exercise adherence (17). An interesting extension to this work would be to include a condition in which participants are forewarned about the potential challenges to their enjoyment/interest/value perceptions (i.e., degree of autonomous motivation) before exercising under controlling conditions to see whether the predisposition towards “unhealthy” food consumption in the present study can be dampened. Dimmock and colleagues (8), for example, demonstrated that inoculation messages provided before a monotonous and controlling exercise session (i.e., one in which autonomous motivation may be “challenged”) may help protect desirable exercise experiences in the face of motivational challenges. Thus, by protecting more positive forms of motivation during exercise, this may attenuate the desire to consume hedonically pleasurable food in the aftermath of a controlling exercise session.

It is important to acknowledge that the assessment of energy intake in the present study did not extend past the immediate postexercise meal given that this was where the greatest impact was expected to be observed. However, examining the persistence of these effects beyond the postexercise meal is important to better understand how autonomy in exercise affects free living energy intake in the long term. Another potential limitation of the present study is that it is not clear whether a specific aspect of choice relating to the exercise bout (i.e., duration, mode, intensity) had the most impact on subsequent food intake. Given the lack of research in this area to strengthen our manipulation of autonomy (or lack thereof), choice was manipulated through a combination of mode of exercise, time of commencement of exercise, exercise duration, and music accompaniment. Future research could isolate these aspects of choice (and others) to determine whether some aspects are more important than others in facilitating an autonomy supportive environment, and whether less comprehensive manipulations of choice may yield similar results to those in the present study.

In summary, we have demonstrated, for the first time that individuals who experience choice during exercise, relative to those who have no-choice, consume less unhealthy food during recovery, which in turn assists in moderating overall energy intake. Given the successful manipulation of autonomy support through relatively simple offerings of choice (timing and mode of exercise, duration, intensity, and background music), the findings from this study have important implications for exercise prescription and instruction by underscoring the importance of facilitating an autonomy-supportive exercise environment.

No funding to declare. The authors would like to acknowledge the assistance of Tasmiah Masih (accredited dietitian) for her independent health ratings of the foods provided in the laboratory test meal. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

The results of this study do not constitute endorsement by ACSM.

Back to Top | Article Outline

REFERENCES

1. Balaguera-Cortes L, Wallman KE, Fairchild TJ, Guelfi KJ. Energy intake and appetite-related hormones following acute aerobic and resistance exercise. Appl Physiol Nutr Metab. 2011;36(6):958–66.
2. Baumeister RF, Vohs KD, Tice DM. The strength model of self-control. Curr Dir Psychol Sci. 2007;16(6):351–5.
3. Blundell JE, Gibbons C, Caudwell P, Finlayson G, Hopkins M. Appetite control and energy balance: impact of exercise. Obes Rev. 2015;16(1 Suppl):67–76.
4. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14(5):377–81.
5. Crum AJ, Corbin WR, Brownell KD, Salovey P. Mind over milkshakes: mindsets, not just nutrients, determine ghrelin response. Health Psychol. 2011;30(4):424–9.
6. de Lauzon B, Romon M, Deschamps V, et al. The Three-Factor Eating Questionnaire-R18 is able to distinguish among different eating patterns in a general population. J Nutr. 2004;134(9):2372–80.
7. Deci EL, Ryan RM. The general causality orientations scale: Self-determination in personality. J Res Pers. 1985;19(2):109–34.
8. Dimmock JA, Gagné M, Proud L, Howle TC, Rebar AL, Jackson B. An exercise in resistance: inoculation messaging as a strategy for protecting motivation during a monotonous and controlling exercise class. J Sport Exerc Psychol. 2016;38(6):567–578.
9. Dimmock JA, Guelfi KJ, West JS, Masih T, Jackson B. Does motivation for exercise influence post-exercise snacking behavior? Nutrients. 2015;7(6):4804–16.
10. Dye L, Blundell J. Menstrual cycle and appetite control: implications for weight regulation. Hum Reprod. 1997;12(6):1142–51.
11. Erdmann J, Tahbaz R, Lippl F, Wagenpfeil S, Schusdziarra V. Plasma ghrelin levels during exercise—effects of intensity and duration. Regul Pept. 2007;143(1-3):127–35.
12. Fenzl N, Bartsch K, Koenigstorfer J. Labeling exercise fat-burning increases post-exercise food consumption in self-imposed exercisers. Appetite. 2014;81:1–7.
13. Flint A, Raben A, Blundell JE, Astrup A. Reproducibility, power and validity of visual analogue scales in assessment of appetite sensations in single test meal studies. Int J Obes Relat Metab Disord. 2000;24(1):38–48.
14. Garg N, Wansink B, Inman JJ. The influence of incidental affect on consumers’ food intake. J Marketing. 2007;71(1):194–206.
15. George VA, Morganstein A. Effect of moderate intensity exercise on acute energy intake in normal and overweight females. Appetite. 2003;40(1):43–6.
16. Godin G, Shephard R. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci. 1985;10(3):141–6.
17. Hagberg LA, Lindahl B, Nyberg L, Hellénius ML. Importance of enjoyment when promoting physical exercise. Scand J Med Sci Sports. 2009;19(5):740–7.
18. Hagobian TA, Sharoff CG, Stephens BR, et al. Effects of exercise on energy-regulating hormones and appetite in men and women. Am J Physiol Regul Integr Comp Physiol. 2009;296(2):R233–42.
19. Hall KD, Sacks G, Chandramohan D, et al. Quantification of the effect of energy imbalance on bodyweight. Lancet. 2011;378(9793):826–37.
20. Hong S-M, Faedda S. Refinement of the Hong psychological reactance scale. Educ Psychol Meas. 1996;56(1):173–82.
21. Lennox SS, Bedell JR, Stone AA. The effect of exercise on normal mood. J Psychosom Res. 1990;34(6):629–36.
22. Lutz R, Lochbaum M, Turnbow K. The role of relative autonomy in post-exercise affect responding. J Sport Behav. 2003;26(2):137.
23. Magaraggia C, Dimmock JA, Jackson B. The effect of learning climate on snack consumption and ego depletion among undergraduate students. Appetite. 2013;69:174–9.
24. Markland D, Tobin V. A modification to the behavioural regulation in exercise questionnaire to include an assessment of amotivation. J Sport Exerc Psychol. 2004;26(2):191–6.
25. Mattes R. Hunger ratings are not a valid proxy measure of reported food intake in humans. Appetite. 1990;15(2):103–13.
26. Melanson KJ, Westerterp-Plantenga MS, Saris WH, Smith FJ, Campfield LA. Blood glucose patterns and appetite in time-blinded humans: carbohydrate versus fat. Am J Physiol. 1999;277(2):R337–45.
27. Moller AC, Deci EL, Ryan RM. Choice and ego-depletion: the moderating role of autonomy. Pers Soc Psychol Bull. 2006;32(8):1024–36.
28. Moshier SJ, Landau AJ, Hearon BA, et al. The development of a novel measure to assess motives for compensatory eating in response to exercise: the CEMQ. Behav Med. 2016;42(2):93–104.
29. Nix GA, Ryan RM, Manly JB, Deci EL. Revitalization through self-regulation: the effects of autonomous and controlled motivation on happiness and vitality. J Exp Soc Psychol. 1999;35(3):266–84.
30. Parillo M, Riccardi G. Diet composition and the risk of type 2 diabetes: epidemiological and clinical evidence. Br J Nutr. 2004;92(1):7–19.
31. Rabia M, Knäuper B, Miquelon P. The eternal quest for optimal balance between maximizing pleasure and minimizing harm: the compensatory health beliefs model. Br J Health Psychol. 2006;11(Pt 1):139–53.
32. Raedeke TD. The relationship between enjoyment and affective responses to exercise. J Appl Sport Psychol. 2007;19(1):105–15.
33. Ryan RM. Control and information in the intrapersonal sphere: an extension of cognitive evaluation theory. J Pers Soc Psychol. 1982;43(3):450.
34. Ryan RM, Deci EL. From ego depletion to vitality: theory and findings concerning the facilitation of energy available to the self. Soc Personal Psychol Compass. 2008;2(2):700–17.
35. Ryan RM, Frederick CM. On energy, personality, and health: subjective vitality as a dynamic reflection of well-being. J Pers. 1997;65(3):529–65.
36. Schubert MM, Desbrow B, Sabapathy S, Leveritt M. Acute exercise and subsequent energy intake. A meta-analysis. Appetite. 2013;63:92–104.
37. Sim AY, Wallman KE, Fairchild TJ, Guelfi KJ. High-intensity intermittent exercise attenuates ad-libitum energy intake. Int J Obes (Lond). 2014;38(3):417–22.
38. Terry PC, Lane AM, Fogarty GJ. Construct validity of the Profile of Mood States—adolescents for use with adults. Psychol Sport Exerc. 2003;4(2):125–39.
39. Wansink B. Environmental factors that increase the food intake and consumption volume of unknowing consumers. Annu Rev Nutr. 2004;24:455–79.
40. Warburton DE, Nicol CW, Bredin SS. Health benefits of physical activity: the evidence. CMAJ. 2006;174(6):801–9.
41. Werle CO, Wansink B, Payne CR. Just thinking about exercise makes me serve more food. Physical activity and calorie compensation. Appetite. 2011;56(2):332–5.
Keywords:

APPETITE; ENERGY INTAKE; EXERCISE MOTIVATION; COMPENSATION

© 2017 American College of Sports Medicine