Changes in food reward have been implicated in exercise-induced compensatory eating behaviour. However, the underlying mechanisms of food reward, and the physiological correlates of exercise-induced changes in food reward, are unknown.
Day-to-day food intake involves the coordination of both homeostatic and nonhomeostatic signals in the overall expression of eating behaviour [
Changes in food reward may also play a role in compensatory eating behaviour following exercise [
Forty-six overweight and obese individuals participated in the present study (30 females, BMI =
Participants completed a 12-week supervised aerobic exercise program designed to expend 2500 kcal.wk−1. Body composition, food intake, and fasting metabolic-related hormones (glucose, insulin, and leptin) were measured at baseline, week six, and postintervention. In addition, explicit liking and implicit wanting for a standardised array of high fat and low fat foods were assessed before a fixed-energy meal, at baseline, week six, and postintervention, using a validated computer based task, for example, the Leeds Food Preference Questionnaire [
Participants completed a 12-week aerobic exercise program in which they exercised five days per week, expending 500 kcal per session at 70% of age-predicted maximum heart rate. All exercise sessions were supervised in the research laboratory, and participants could choose from a range of exercise modes (running, cycling, rowing, or stepping). Individual exercise prescriptions were calculated using standard stoichiometric equations [
At baseline, week six, and postintervention, venous blood, body composition, and maximal aerobic capacity were measured in the morning (7–9am) following an overnight fast (10–12 hrs). Baseline measures were taken prior to the start of the intervention, while postintervention measures were taken upon completion of the exercise intervention (a minimum of 48 hrs after the final exercise bout and within one week of finishing the intervention). Body composition was measured using air-displacement plethysmography (BOD POD Body Composition System, Life Measurement, Inc., Concord, USA). After voiding, participants were weighed (to the nearest 0.01 kg) and instructed to sit in the BOD POD. Measurements were taken according to manufacturers’ instructions, with thoracic gas volumes estimated using the manufacturer’s software. In addition, the fat mass index (FMI; fat mass/height2) and the fat-free mass index (FFMI; fat-free mass/height2) were calculated from these body composition data. Maximal aerobic capacity (
Fasting glucose, insulin, and leptin were measured at baseline, week six, and postintervention in a subsample of 32 participants who completed the exercise intervention. Fasting venous blood samples were collected into EDTA-containing Monovette tubes. After collection, blood samples were centrifuged for 10 min at 4°C at 3500 rpm and were immediately pipetted into Eppendorf tubes and stored at −80°C until analysis. Insulin and leptin were analysed using a magnetic bead based multiples kit (Millipore, Billerica, MA, USA). Furthermore, insulin resistance was calculated using the homeostatic model of assessment (HOMA) [
A laboratory-based test meal protocol was used to measure food intake at baseline, week six, and postintervention. At each time point, participants consumed test meals at 4-hour intervals. No exercise was performed on these days. A detailed description of the foods provided can be found elsewhere [
Prior to the lunch test meal, food reward was assessed using the Leeds Food Preference Questionnaire (LFPQ; [
Data are reported as mean ± SEM throughout. Statistical analyses were performed using IBM SPSS for Windows (Chicago, Illinois, Version 21). For food reward measures, mean scores for high fat and low fat categories were computed for implicit wanting and explicit liking outcomes. Mean low fat scores were then subtracted from the mean for high fat scores to provide a composite score representing reward value for high fat relative to low fat food for both liking and wanting. Using this approach, a positive score indicated greater liking or wanting for high fat foods over low fat foods; a negative score indicated greater liking or wanting for low fat foods over high fat foods; and a score of zero indicated an equal liking or wanting for high and low fat foods. Scores on each food reward outcome were calculated at baseline, week six, and postintervention and analysed using one-way repeated measures ANOVAs.
Changes in body composition, metabolic-related hormones, and total daily energy intake were examined using one-way repeated measures ANOVAs. Where appropriate, Greenhouse-Geisser probability levels were used to adjust for sphericity, and Bonferroni adjustments were applied to control for multiple post hoc comparisons. To test for associations between physiological variables and food reward, Pearson partial correlation coefficients were used, controlling for sex. Firstly, cross-sectional models were examined using mean scores on each variable collapsed across the three time points of the exercise intervention (i.e., baseline, week six, and postintervention). Secondly, associations between changes in physiological variables and changes in explicit liking and implicit wanting following the exercise intervention were performed. Change variables were calculated by subtracting baseline scores from postintervention scores. To control for confounding effects of body composition, metabolic hormones were tested both with and without adjustment for adiposity by dividing by percentage body fat.
As can be seen in Table
Body composition and metabolic values during the 12-week exercise intervention (
Baseline | Week six | Postintervention | Delta Δ |
|
|
---|---|---|---|---|---|
Body mass (kg) | 88.21 (2.04) | 87.39 (2.00) | 86.49 (2.04) | −1.72 (0.41) | 0.000* |
Fat mass (kg) | 35.71 (1.34) | 34.48 (1.35) | 33.49 (1.43) | −2.23 (0.38) | 0.000* |
Fat mass index (kg/m2) | 12.61 (0.52) | 12.17 (0.53) | 11.85 (0.56) | −0.76 (0.14) | 0.000* |
Body fat (%) | 40.33 (1.13) | 39.24 (1.16) | 38.43 (1.22) | −1.90 (0.32) | 0.000* |
Fat-free mass (kg) | 52.48 (1.43) | 52.91 (1.41) | 53.00 (1.39) | 0.52 (0.17) | 0.081 |
Fat-free mass index (kg/m2) | 18.25 (0.31) | 18.40 (0.30) | 18.41 (0.30) | 0.17 (0.62) | 0.009* |
VO2peak (mL·kg−1·min−1) | 33.33 (1.17) | 37.45 (1.08) | 39.16 (0.09) | 5.83 (0.95) | 0.000* |
Fasting glucose (mmol·L−1) | 4.93 (0.15) | 4.88 (0.17) | 4.73 (0.19) | −0.20 (0.25) | 0.415 |
Fasting insulin (ng·L−1) | 1034.32 (106.24) | 918.77 (105.33) | 991.34 (113.24) | −42.98 (82.94) | 0.230 |
HOMA index | 3.18 (0.31) | 2.92 (0.31) | 3.02 (0.33) | −0.16 (0.25) | 0.554 |
Fasting leptin (ng·L−1) | 38318.80 (4832.26) | 369923.92 (4612.41) | 32102.87 (5333.58) | −6215.93 (3076.37) | 0.023* |
Delta Δ: baseline to postintervention change. VO2peak: maximal aerobic capacity. HOMA: homeostatic model of assessment. *Significant difference between baseline and postintervention (
Table
Changes in food intake, explicit liking, and implicit wanting for high fat versus low fat foods during the 12-week exercise intervention (
Baseline | Week six | Postintervention | Delta Δ |
|
|
---|---|---|---|---|---|
Total daily EI (kcal·day−1) | 2949.29 (79.15) | 2877.24 (92.77) | 2892.81 (88.11) | −56.48 (60.15) | 0.438 |
Explicit liking |
−0.20 (2.25) | −1.08 (2.16) | −0.85 (2.02) | −0.65 ( 1.72) | 0.919 |
Implicit wanting |
1.10 (4.18) | −2.56 (4.47) | −3.17 (3.98) | −4.27 (2.58) | 0.114 |
EI: energy intake; Delta Δ: baseline to postintervention change. Positive appeal bias score = preference for high fat foods > low fat foods. Negative appeal bias score = preference for low foods > high fat foods.
As can be seen in Table
Pearson partial correlation coefficients (controlling for sex) between food reward and the cross-sectional and exercise-induced changes in body composition and fasting metabolic-related hormones.
Body composition and VO2peak | Metabolic hormones | ||||
---|---|---|---|---|---|
Liking | Wanting | Liking | Wanting | ||
BM |
|
|
Glucose | 0.019 | 0.060 |
ΔBM | −0.251 | 0.116 | ΔGlucose | −0.014 | 0.077 |
FM |
|
|
Adjusted Glucose | −0.267 | −0.336 |
ΔFM | −0.196 | 0.004 | ΔAdjusted Glucose | −0.039 | 0.061 |
FMI | 0.265 |
|
Insulin | −0.236 | 0.311 |
ΔFMI | −0.223 | −0.016 | ΔInsulin | −0.206 | −0.216 |
BF% | 0.212 |
|
Adjusted Insulin | 0.155 | 0.194 |
ΔBF% | −0.210 | −0.101 | ΔAdjusted Insulin | −0.213 | −0.178 |
FFM |
|
0.230 | Leptin |
|
|
ΔFFM | −0.138 | 0.265 | ΔLeptin |
|
−0.110 |
FFMI | 0.213 | 0.184 | Adjusted Leptin |
|
|
ΔFFMI | −0.121 | −0.094 | ΔAdjusted Leptin |
|
−0.159 |
VO2peak | −0.224 | −0.231 | HOMA | 0.213 | 0.008 |
ΔVO2peak | −0.178 | −0.179 | ΔHOMA | −0.123 | −0.151 |
Adjusted HOMA | 0.090 | 0.139 | |||
ΔAdjusted HOMA | −0.124 | −0.152 |
VO2peak: maximal aerobic capacity; FM: fat mass; FMI: fat mass index; FFM: fat-free mass; FFMI: fat-free mass index; %BF: percentage body fat; HOMA: homeostatic model of assessment. Delta Δ: baseline to postintervention change.
No associations existed between changes in food reward and changes in body composition following the intervention (Table
Scatter plot illustrating the relationship between the change in fasting leptin (adjusted for percentage body fat) following the exercise intervention and the change in appeal bias scores for liking for high fat foods (
This study examined whether components of body composition and metabolic-related hormones were associated with food reward in overweight and obese individuals during 12 weeks of aerobic exercise. Cross-sectional analyses disclosed associations between body composition (fat mass and fat-free mass), fasting leptin, and food reward. Furthermore, a novel relationship was also disclosed between the change in fasting leptin and the change in explicit liking for high fat foods following the exercise intervention. Specifically, a decline in fasting leptin was associated with an increased liking for high fat foods relative to low fat foods following the intervention. This relationship was independent of changes in fat mass and suggests that leptin may have a key role in mediating changes in food reward during exercise-induced weight loss.
The 12-week exercise intervention resulted in significant (but modest) reductions in body mass, fat mass, and percentage body fat, while fat-free mass was preserved at baseline levels. In addition, significant improvements in
Recent evidence has demonstrated the importance of distinguishing explicit perceptions of liking from behavioural operations of wanting, with these components of food reward considered to be separable risk factors in overconsumption and weight gain [
It has been suggested that obese individuals display a loss of hedonic control over eating when exposed to highly palatable foods compared to lean individuals [
Interestingly, the present study also disclosed a novel relationship between the changes in fasting leptin and explicit liking for high fat foods following the exercise intervention, with a decline in fasting leptin associated with an increase in liking for high fat foods relative to low fat foods. This relationship is consistent with the proposed role of leptin in food reward, in which leptin is thought to tonically inhibit brain reward pathways [
The present findings are in keeping with the idea that leptin is primarily a “starvation” signal rather than a “satiety” signal [
It has previously been reported that the change in leptin (independent of fat mass) during weight loss was negatively associated with the changes in subjective appetite [
The present study has some limitations that deserve comment. When interpreting the findings of the present study, it is important to note that a nonexercise control condition was not included. However, the observed improvements in body composition,
Through the concurrent measurement of physiological and behavioural components of energy balance, this study has disclosed novel relationships between food reward, body composition, and metabolic-related hormones in overweight and obese individuals. Cross-sectional relationships were found between measures of explicit liking and both fat mass and fat-free mass. However, only fat mass was found to be associated with implicit wanting, suggesting that aspects of body composition may differentially affect the separate components of food reward. Independent of adiposity, a positive relationship between fasting leptin and liking and wanting for high fat food was demonstrated. Furthermore, a decline in fasting leptin following the exercise intervention was found to be associated with an increase in liking for high fat relative to low fat foods. Taken together, these findings suggest a dynamic role for fasting leptin as a regulatory signal of food reward during exercise-induced weight loss.
The authors declare that there is no conflict of interests.
This research was supported by BBSRC Grant nos. BBS/B/05079 and BB/G005524/1 (DRINC), EU FP7 Full4Health (#266408), Uppsala University and the Stockholm County Council (ALF).