. 2020 Apr 15;9(2):63–80. doi: 10.1007/s13679-020-00372-3

Table 1.

Description of the observational, acute, and chronic training studies examining the impact of physical activity on food reward

Reference	Participant characteristics	Exercise/physical activity details	Food reward method	Food reward results	Associations with appetite outcomes
Observational studies
Beaulieu et al. 2017 [34] UK	Sex: females and males BMI status: < 29.9 kg/m² Active: n = 20 (50% males) Age: 30 (10) years BMI: 22.6 (1.9) kg/m² Inactive: n = 19 (42% males) Age: 30 (9) years BMI: 23.1 (2.7) kg/m²	PA level assessment: IPAQ and PA monitor (SenseWear) Active: ≥ 4 exercise sessions/week (MVPA = 182 (67) min/day) Inactive: ≤ 1 exercise session/week (MVPA = 103 (37) min/day) Exercise session: ≥ 40 min MVPA	- LFPQ: liking and implicit wanting bias for fat - Setting: laboratory - State: pre and post high-fat or high-carbohydrate lunch (ad libitum)	- No difference in food reward (liking and wanting) between groups. - Difference between liking/wanting: none	- Food intake: energy intake greater in high-fat relative to high-carbohydrate in both groups. - Eating behavior traits: tendency for restraint to be higher in HiPA compared with LoPA - Body composition: HiPA had lower body fat percentage than LoPA
Beaulieu et al. 2017 [35] UK	Sex: females and males BMI status: < 29.9 kg/m² HiMVPA: n = 12 (33% males) Age: 29 (10) years BMI: 22.4 (2.1) kg/m² ModMVPA: n = 11 (27% males) Age: 26 (3) years BMI: 22.7 (2.2) kg/m² LoMVPA: n = 11 (27% males) Age: 30 (11) years BMI: 23.1 (2.9) kg/m²	PA level assessment: Tertiles of daily MVPA measured by PA monitor (SenseWear) HiMVPA: 174 (39) min/day ModMVPA: 121 (15) min/day LoMVPA: 83 (16) min/day	- LFPQ: liking and implicit wanting bias for fat - Setting: laboratory - State: pre and post preloads—HE (~ 700 kcal) and LE (~ 260 kcal) relative to water control (0 kcal)	- No difference in food reward (liking and wanting) between groups - Difference between liking/wanting: none	- Food intake: greater reduction in liking and wanting after HE preload relative to LE preload in all groups. ModMVPA and HiMVPA reduced ad libitum energy intake at the lunch meal following consumption of the HE compared with the LE preload, while the LoMVPA group did not. No effect of MVPA group on energy intake at dinner or evening snack box. - Eating behavior traits: no differences between groups - Body composition: no differences between groups
Drummen et al. 2019 [36•] Netherlands	Sex: females and males BMI status: > 25 kg/m² n = 39 (56% males) Age: 53 (11) years BMI: 32.3 (3.7) kg/m² Participants had impaired fasting glucose and/or glucose tolerance	PA level assessment: Baecke questionnaire (work, sport, and leisure-time physical activity) [37] Work = 2.6 (0.8) range 1.3–4.3 Sport = 2.6 (0.7) range 1.0–4.0 Leisure = 2.9 (0.7) range 1.5–4.4 Scores ranging from 1 (low level of PA) and 5 (high level of PA)	- fMRI: BOLD signals to high-energy food, low-energy food and non-food images - Setting: laboratory - State: after overnight fast	- Inverse association between leisure-time PA and food compared with non-food brain activation in the right thalamus, left middle cingulate gyrus, right precuneus, left putamen, and left angular gyrus. - Positive association between work-related physical activity and brain activation (disappeared after adjusting for BMI and age) - No association with sport-related physical activity. - Difference between liking/wanting: NR	- Food intake: NR - Eating behavior traits: positive association between brain activation and disinhibition. No association with restraint or susceptibility to hunger. - Body composition: no association
Horner et al. 2016 [38] Australia	Sex: males BMI status: 18–40 kg/m² Active: n = 22 Age: 29 (8) years BMI: 24.5 (2.6) kg/m² Inactive: n = 22 Age: 31 (9) years BMI: 27.4 (4.2) kg/m²	PA level assessment: Self-report and PA monitor (ActiGraph) Active: ≥ 4 exercise sessions/week (PA = 709 (239) kcal/day) Inactive: ≤ 1 exercise session/week (PA = 525 (185) kcal/day) Exercise session: ≥ 40 min MVPA	- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: post fixed breakfast (400 kcal) and pre lunch 5 h later	- Fed: Active had lower liking for high-fat foods, low-fat sweet foods and for foods overall compared with inactive. Active had greater wanting for low-fat savory foods. - Hungry: No differences in liking between active and inactive. Active had greater wanting for low-fat savory foods. - Fed to hungry: Active had greater increase in liking for all food categories combined than inactive. - Difference between liking/wanting: Faster gastric emptying associated with liking for savory foods and slower gastric emptying associated with greater implicit wanting for high-fat foods.	- Gastric emptying: Inverse association with post-prandial changes in liking low-fat savory foods. Positive association with liking taste bias in hungry state (i.e., faster gastric emptying associated with greater liking for savory foods). No association with liking fat bias nor implicit wanting taste bias. Positive association with implicit wanting fat bias in fed and hungry states. Effects independent of body fat.
Killgore et al. 2013 [39] USA	Sex: females and males BMI status: 19.8–34.8 kg/m² n = 37 (59% males) Age: 30 (8) years BMI: 24.5 (3.7) kg/m²	PA level assessment: Self-reported habitual PA in min/week (typical days/week × duration/day) PA = 151 (160) min/week (range 0–540)	- fMRI: response to high-energy food, low-energy food and non-food images - Subjective food preferences: desire to eat depicted food item at that moment (VAS) - Setting: laboratory - State: 1 h fasted (pre scan food intake 323 (245) kcal)	- Inverse association between habitual PA and fMRI responses (medial orbitofrontal cortex and left insula) to high-energy foods relative to low-energy foods - Inverse association between habitual PA and preference for high-energy savory foods	- Subjective food preference: fMRI responses positively associated with preference for high-energy savory foods (not for high-energy sweet foods). No association between fMRI responses to high-energy relative to low-energy foods and preference for low-energy foods.
Luo et al. 2018 [40•] USA	Sex: females and males BMI status: NR n = 40 (48% males) Lean individuals: n = 22 (46% males) Age: 21 (2) years BMI: 22.6 (1.9) kg/m² Individuals with obesity: n = 18 (50% males) Age: 22 (2) years BMI: 35.2 (4.0) kg/m²	PA level assessment: Self-reported from 3 to 5 24-h recalls over 2 months (mean daily minutes of MVPA i.e., activities ≥ 3 METs) Lean individuals: MVPA = 125 (84) min/day Individuals with obesity: MVPA = 134 (114) min/day	- fMRI: responses to high-energy food and non-food images - Setting: laboratory - State: 9–11 am after overnight fast, task performed 20–30 min after 75 g glucose ingestion	- Inverse association between MVPA and brain response to food cues in middle insula and left postcentral gyrus. - Individuals with obesity: inverse association between MVPA and brain responses. - Lean individuals: non-significant inverse association between MVPA and brain responses. - Association between MVPA and brain responses stronger in males than females.	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Oustric et al. 2018 [41] UK	Sex: females BMI status: 18.5–45.0 kg/m² n = 156 Age: 53 (11) years BMI: 32.3 (3.7) kg/m² Pooled data from 6 studies	PA level assessment: Quintiles of daily MVPA measured by PA monitor (SenseWear) Q1: 25 (8) min/day Q2: 53 (9) min/day Q3: 83 (9) min/day Q4: 120 (12) min/day Q5: 197 (62) min/day	- LFPQ: liking and implicit wanting bias for fat/taste - Setting: laboratory - State: hungry (3–10 h fast)	- MVPA inversely associated with liking and implicit wanting fat bias. - MVPA positively associated with liking taste bias. - Q5 greater liking and wanting for low-fat foods, while Q1-Q3 greater liking and wanting for high-fat foods. - Difference between liking/wanting: none	- Food intake: NR - Eating behavior traits: No association between PA and food cravings. Craving for sweet food (Control of Eating Questionnaire; CoEQ) positively associated with explicit liking and implicit wanting for sweet foods on the LFPQ. Craving for savory foods (CoEQ) associated with LFPQ explicit wanting for savory foods. Craving control (CoEQ) negatively associated with implicit wanting for high-fat foods. - Body composition: Fat mass index (FMI) but not waist circumference (WC) was inversely associated with explicit liking and implicit wanting for sweet relative to savory foods. WC was positively associated with all liking and implicit wanting for high-fat relative to low-fat foods. FMI was also associated with implicit wanting for high-fat foods.
Acute exercise studies
Alkahtani et al. 2014 [42] Australia	Sex: males BMI status: > 25 kg/m² PA level: sedentary (criteria NR) n = 12 Age: 29 (4) years BMI: 29.1 (2.4) kg/m²	- Intensity: moderate-intensity interval training (MIIT; alternating between ± 20% FAT_max) vs. high-intensity interval training (HIIT; 85% VO_2peak) - Type: cycle ergometer - Duration: MIIT 5-min stages at ±20% FAT_max for 30 min, HIIT 15-s intervals and 15-s recovery (workload matched to MIIT; ~ 18 min) - Timing: morning - Control condition: no; MIIT vs. HIIT	- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: before and after exercise (after an overnight fast)	- Decrease in wanting and increase in liking for all the food categories independent of the intensity - Difference between liking/wanting: liking increased while wanting decreased but without a control this response might be due to the effect of time rather than exercise	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Alkahtani et al. 2019 [43] Saudi Arabia	Sex: males BMI status: NR PA level: moderately active (2–5 h structured aerobic exercise/week) n = 14 (8 for food reward data) Age: 24 (6) years BMI: 23.4 (3.3) kg/m²	- Intensity: moderate (60% VO_2max) interspersed with low (30% VO_2max) - Type: contraction type eccentric (downhill running at − 12% inclination) vs. concentric (flat running) - Duration: 5 stages of 8 min at 60%VO_2max/2 min at 30%VO_2max - Timing: morning - Control condition: yes; no exercise	- LFPQ: liking and implicit wanting bias for fat/taste - Setting: laboratory - State: before, after exercise and 24 h after exercise (before an ad libitum lunch)	- No change in food reward after exercise - Difference between liking/wanting: greater liking of savory foods over sweet foods in downhill running than front running	- Food intake: no change - Eating behavior traits: NR - Body composition: NR
Crabtree et al. 2014 [44] UK	Sex: males BMI status: 21.8–26.6 kg/m² PA level: moderately active (2 h/week) n = 16 Age: 23 (3) years BMI: 24.2 (2.4) kg/m²	- Intensity: high (70% VO_2max) - Type: treadmill run - Duration: 60 min - Timing: morning - Control condition: yes; no exercise	- fMRI: BOLD signals to high- and low-energy food cues compared with non-food pictures - Setting: laboratory - State: fasted	- Decreased activation in the pallidum for high-energy food and increase for low-energy food after exercise - Difference between liking/wanting: NR	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Evero et al. 2012 [45] USA	Sex: females and males BMI status: < 25 kg/m² PA level: habitually active (> 3 h/week) n = 30 (57% males) Age: 22 (4) years BMI: 23.6 (2.2) kg/m²	- Intensity: high (83% HR_max) - Type: cycle ergometer - Duration: 60 min - Timing: morning - Control condition: yes; no exercise	- fMRI: BOLD signals to high- and low-energy food cues compared with neutral control - Setting: laboratory - State: fMRI was performed after exercise after an 8–12 h overnight fast	- Exercise reduced the neuronal response to food cues in brain regions related with food reward (i.e., insula, putamen, rolandic operculum) - Difference between liking/wanting: decrease in regions related to liking and wanting but no behavioral measures	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Farah et al. 2012 [46] UK	Sex: females and males BMI status: no limits PA level: NR n = 27 (52% males) Female: n = 13 Age: 26 (3) years BMI: BMI: 22.8 (3.1) kg/m² Male: n = 14 Age: 36 (10) years BMI: 26.1 (3.3) kg/m²	- Intensity: moderate (6 METs) - Type: treadmill walk - Duration: 60 min - Timing: morning - Control condition: yes; no exercise	- VAS: liking - Setting: laboratory - State: immediately, 60, 120, and 180 min after exercise (overnight fasted)	- No change in liking - Difference between liking/wanting: wanting not measured	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Finlayson et al. 2009 [47] UK	Sex: females BMI status: < 25 kg/m² PA level: 2.4 (1.2) engagements/week n = 24 Age: 24 (6) years BMI: 22.3 (2.9) kg/m²	- Intensity: moderate (70% HR_max) - Type: cycle ergometer - Duration: 50 min - Timing: morning - Control condition: yes, no exercise	- LFPQ: relative preference (food choice), liking and implicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: before and after exercise (2 h after a fixed breakfast, kcal NR) and after an ad libitum lunch 30-min post exercise	- Increase in implicit wanting after exercise in those who compensated or ate more at the ad libitum lunch in response to exercise - Difference between liking/wanting: Changes in implicit wanting but not liking	- Food intake: After exercise some individuals increased their energy intake (compensators) and had enhanced implicit wanting - Eating behavior traits: NR - Body composition: NR
Martins et al. 2015 [48] Norway	Sex: females and males BMI status: > 25 kg/m² PA level: sedentary (criteria NR) n = 12 (42% males) Age: 33 (10) years BMI: 32.3 (2.7) kg/m²	- Intensity: HIIT and ½ HIIT (all out; average ~ 85% HR_max), continuous (70% HR_max) - Type: HIIT vs. ½ HIIT vs. continuous cycling - Duration: HIIT (8-s intervals and 12-s recovery for 250 kcal; ~ 18 min), ½ HIIT (8-s intervals and 12-s recovery for 125 kcal; ~ 9 min), continuous exercise (250 kcal; ~ 27 min) - Timing: morning - Control condition: yes; no exercise	- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat - Setting: laboratory - State: before and after exercise/ before an ad libitum lunch (standardized breakfast of 600 kcal consumed 1 h before exercise/rest)	- No change in food reward - Difference between liking/wanting: no differences	- Food intake: no change - Eating behavior traits: NR - Body composition: NR
McNeil et al. 2015 [49] Canada	Sex: females and males BMI status: < 25 kg/m² PA level: inactive (<150 min/week) n = 16 (50% males) Age: 22 (3) years BMI: 22.8 (1.8) kg/m²	- Intensity: high (aerobic 70% VO_2peak, resistance 70% 1-repetition maximum) - Type: aerobic vs. resistance - Duration: aerobic ~ 24 min, resistance ~ 86 min (matched for energy expenditure at 4 kcal/kg; ~ 275 kcal) - Timing: morning - Control condition: yes; no exercise	- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste - Setting: laboratory - State: pre and post ad libitum lunch 30 min after exercise (standardized breakfast of 534 kcal consumed ~ 1.5 h before exercise)	- Decrease in the relative preference for high-fat relative to low-fat foods after both exercise - Decrease in liking for high-fat foods following resistance, but not aerobic - Difference between liking/wanting: change in food choice and liking, but not implicit wanting	- Food intake: no change - Eating behavior traits: NR - Body composition: no difference in bodyweight
Miguet et al. 2018 [50] France	Sex: females and males (adolescents) BMI status: > 29.9 kg/m² PA level: inactive (<2 h/week) n = 33 (36% males) Age: 13 (1) years BMI: 35.0 (4.3) kg/m²	- Intensity: high (70%, 75%, 80%, 85%, and 90% HR_max) - Type: high-intensity interval training cycling - Duration: 5 × 2-min increasing intensity intervals followed by 30-s recovery (15 min) - Timing: morning - Control condition: yes; no exercise	- LFPQ: relative preference, liking and implicit wanting bias for fat/taste - Setting: laboratory - State: pre and post ad libitum lunch 30 min after exercise (standardized breakfast of 500 kcal consumed 2.5 h before exercise)	- Decrease in implicit wanting for sweet in the exercise condition vs. increase in the control - Difference between liking/wanting: implicit wanting but not liking is decreasing	- Food intake: decrease in energy intake at lunch and dinner - Eating behavior traits: NR - Body composition: NR
Thackray et al. unpublished UK	Sex: females and males BMI status: 18.5–29.9 kg/m² PA level: habitually active n = 32 Age: 23 (2) years BMI: 23.9 (2.6) kg/m²	- Intensity: self-determined moderate-to-high intensity (RPE of 15 “hard”) - Type: swimming vs. cycling - Duration: 6 × 8-min intervals with 2-min recovery - Timing: morning - Control condition: yes, no exercise	- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste - Setting: laboratory - State: post-exercise (3 h after fixed breakfast of 650 kcal for males, 525 kcal for females, and before ad libitum lunch meal)	- Tendency for a main effect of trial for implicit wanting fat bias (post hoc: cycling < control, cycling < swimming). - No impact of swimming or cycling on other food reward parameters. - Difference between liking/wanting: Changes in implicit wanting but not liking	- Food intake: increase in ad libitum energy intake after swimming but not after cycling - Eating behavior traits: NR - Body composition: NR
Thivel et al. 2019 [51] France	Sex: females and males BMI status: < 25 kg/m² PA level: moderately active (150–240 min/week) n = 19 (52% males) Age: 21 (1) years BMI: 22.3 (2.9) kg/m²	- Intensity: low 50% VO_2max, high 75% VO_2max - Type: cycling - Duration: low intensity 45 min, high intensity 30 min - Timing: morning - Control condition: yes; no exercise	- LFPQ: relative preference, liking and implicit wanting bias for fat/taste - Setting: laboratory - State: pre and post fixed lunch 30 min after exercise (females 750 kcal and males 900 kcal; standardized breakfast of 500 kcal consumed 3 h before exercise).	- No change in food reward - Difference between liking/wanting: no differences	- Food intake: (self-reported) no change - Eating behavior traits: NR - Body composition: NR
Saanijoki et al. 2018 [52] Finland	Sex: males BMI status: 19.9–26.9 kg/m² PA level: NR n = 24 Age: 27 (5) years BMI: 23.5 (1.6) kg/m²	- Intensity: moderate (74% HR_max) - Type: aerobic cycling - Duration: 60 min - Timing: NR - Control condition: yes; no exercise	- fMRI: BOLD signals to palatable and non-palatable foods compared with neutral control (cars) - Setting: laboratory - State: post-exercise fasted for 3 h before the scans	- No effects of exercise vs rest on neuronal responses. - Individual variability in the BOLD signals after exercise might be explained by changes in the brain opioid system. Participants who showed most increases in endogenous opioid release also had highest anticipatory fMRI reward responses following the exercise - Difference between liking/wanting: not assessed	- Food intake: NR - Eating behavior traits: NR - Body composition: NR
Chronic exercise training studies
Alkahtani et al. 2014 [53] Australia	Sex: males BMI status: ≥ 25 kg/m² PA level: inactive (criteria NR) n = 10 Age: 29 (4) years Moderate-intensity interval training (MIIT): BMI baseline = 30.7 (3.5) kg/m² BMI post = 30.8 (3.5) kg/m² High-intensity interval training (HIIT): BMI baseline = 30.9 (3.2) kg/m² BMI post = 30.9 (3.2) kg/m²	- Frequency: 3 days/week for 4 week (cross-over with 6-week washout) - Intensity: MIIT ± 20% workload at 45% VO_2peak, HIIT 90% VO_2peak - Type: MIIT vs. HIIT - Duration: 30–45 min (MIIT 5-min stages alternating between ± 20% workload, HIIT 30-s intervals and 30-s recovery) - Timing: NR - Supervision: yes - Control group: no; MIIT vs. HIIT	- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: pre and post 45-min cycling at 45% VO_2max - Measurement time points: week 0 and week 4 in each intervention	- Tendency for explicit liking for high-fat non-sweet foods after acute exercise to increase with MIIT and decrease with HIIT. - No changes in wanting. - Difference between liking/wanting: yes; changes in liking but not wanting	- Food intake: No effects of training on food intake or energy intake. Tendency for fat intake (g) and % energy from fat to increase after MIIT. - Eating behavior traits: NR - Body composition: NR
Beaulieu et al. 2019 [54] UK	Sex: females and males BMI status: 26.0–38.0 kg/m² PA level: inactive (≤ 2 h/week) Exercisers: n = 46 (35% males) Age: 43 (8) years BMI baseline = 30.5 (3.8) kg/m² BMI post = 29.9 (4.0) kg/m² Controls: n = 15 (40% males) Age: 41 (11) years BMI baseline = 31.4 (3.7) kg/m² BMI post = 31.8 (3.9) kg/m²	- Frequency: 5 days/week for 12 weeks - Intensity: 70% HR_max - Type: aerobic (treadmill, rower, cycle ergometer, and elliptical) - Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min) - Timing: NR - Supervision: yes - Control group: yes; no exercise	- LFPQ: liking and implicit wanting bias for fat - Setting: laboratory - State: pre and post fixed lunch (high-fat or high-CHO; 800 kcal) - Measurement time points: week 0 and week 12	- Decrease in wanting after training - No change in liking - Difference between liking/wanting: yes; changes in wanting but not liking	- Food intake: reduction in high fat ad libitum dinner intake but no change in daily high-fat energy intake [55] - Eating behavior traits: decrease in disinhibition and binge eating score - Body composition: reduction in body weight and percentage body fat, but not associated with changes in wanting
Cornier et al. 2012 [56] USA	Sex: females and males BMI status: > 25 kg/m² PA level: NR n = 12 (42% males) Age: 38 (10) years BMI baseline = 33.3 (4.3) kg/m² BMI post = NR	- Frequency: 5 days/week for 24 weeks - Intensity: up to 75% VO_2max - Type: treadmill - Duration: up to 500 kcal/day (40–60 min/day) - Timing: NR - Supervision: yes - Control group: no	- fMRI: responses to food vs non-food cues - Setting: laboratory - State: fasted without exercise for 24 h (chronic exercise) and fasted within 30 min of acute exercise (chronic+acute; 500 kcal 60–75% VO_2max for 40-60 min) - Measurement time points: week 0 and week 24	- Chronic exercise: reduction in neuronal responses observed in the bilateral parietal cortices, left insula and visual cortex. - Chronic + acute exercise: intermediate attenuation of the response to visual food cues in brain region important in food regulation compared with chronic exercise and baseline - Difference between liking/wanting: NR	- Food intake: self-reported energy intake lower after training compared with baseline but no change in macronutrient intake. No association with changes in neuronal responses. - Eating behavior traits: no change in dietary restraint or disinhibition - Body composition: reduction in body fat percentage. Changes in anterior insula responses positively associated with changes in body mass and fat mass.
Finlayson et al. 2011 [57] UK	Sex: females and males BMI status: 26.0–38.0 kg/m² PA level: inactive (≤2 h/week) Responders (non-compensators): n = 20 (43% males) Age: 41 (9) years BMI baseline = 32.3 (4.3) kg/m² BMI post = 30.9 (4.3) kg/m² Non-Responders (compensators): n = 14 (50% males) Age: 37 (12) years BMI baseline = 29.7 (2.2) kg/m² BMI post = 29.3 (2.5) kg/m²	- Frequency: 5 days/week for 12 weeks - Intensity: 70% HR_max - Type: aerobic (treadmill, rower, cycle ergometer, and elliptical) - Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min) - Timing: NR - Supervision: yes - Control group: no	- LFPQ: relative preference (food choice), liking, explicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: pre and post acute exercise (scheduled session) - Measurement time points: week 0 and week 12	- Increase in liking after exercise in non-responders compared with responders at baseline and week 12. - Increase in explicit wanting for high-fat sweet foods in non-responders. - Increase in relative preference for high-fat sweet food in non-responders. - Difference between liking/wanting: implicit wanting NR	- Food intake: NR - Eating behavior traits: NR - Body composition: greater fat mass loss in responders
Finlayson et al. unpublished UK	Sex: females and males BMI status: 26.0–38.0 kg/m² PA level: inactive (≤ 2 h/week) Non-compensators: n = 15 (33% males) Age: 42 (8) years BMI baseline = 30.7 (4.9) kg/m² BMI post = 29.1 (5.0) kg/m² Compensators: n = 15 (33% males) Age: 41 (9) years BMI baseline = 31.8 (3.7) kg/m² BMI post = 32.1 (3.9) kg/m² Controls: n = 15 (33% males) Age: 41 (11) years BMI baseline = 31.4 (3.7) kg/m² BMI post = 31.8 (3.9) kg/m²	- Frequency: 5 days/week for 12 weeks - Intensity: 70% HR_max - Type: aerobic (treadmill, rower, cycle ergometer, and elliptical) - Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min) - Timing: NR - Supervision: yes - Control group: yes, no exercise	- LFPQ: liking and implicit wanting bias for fat - Setting: laboratory - State: pre and post fixed lunch (high-fat or high-CHO; 800 kcal) - Measurement time points: week 0 and week 12	- Non-compensators showed a smaller liking and implicit wanting for high-fat food. - At baseline, compensators showed a strong liking and wanting for high-fat food whereas non-compensators showed no difference between high-fat and low-fat food. - Greater baseline reward for high-fat food in compensators reduced following the exercise intervention. - In the non-compensators, small increase in liking for high-fat food after exercise training, but a simultaneous decrease in wanting for high-fat food. - Difference between liking/wanting: yes, non-compensators increased liking but decreased wanting for high-fat food post-intervention.	- Food intake: tendency for non-compensators to decrease ad libitum dinner meal size from baseline to post-intervention that was not seen in compensators. - Eating behavior traits: NR - Body composition: significant reduction in BMI, body mass, fat mass and WC in non-compensators, whereas no changes in compensators, and increase in body mass and WC.
Martin et al. 2019 [58] USA	Sex: females and males BMI status: 25–45 kg/m² PA level: inactive (< 20 min < 3 days/week) 8 kcal/kg body weight/week (KKW): n = 59 (27% males) Age: 48 (11) years BMI baseline = 31.4 (4.6) kg/m² BMI post = NR (~ 31.3 kg/m²) 20 KKW: n = 51 (29% males) Age: 49 (12) years BMI baseline = 30.6 (4.4) kg/m² BMI post = NR (~ 30.0 kg/m²) Control: n = 61 (26% males) Age: 50 (1) years BMI baseline = 32.3 (4.8) kg/m² BMI post = NR (~ 32.2 kg/m²) Pooled exercisers (n = 110) divided (median split) into compensators/non-compensators based on actual and predicted weight loss.	- Frequency: 3–5 days/week (self-selected) for 24 weeks - Intensity: 65–85% VO_2peak (self-selected) - Type: treadmill - Duration: 8 KKW ~ 35 min/session (~ 700 kcal/week) vs. 20 KKW ~ 55 min/session (~ 1760 kcal/week) - Timing: NR - Supervision: yes - Control group: yes; no exercise	- Food Preference Questionnaire [59] preferences for food classified as alongside 2 components: fat (2 factors: High Fat and Low Fat) and carbohydrate (3 factors: High Simple Sugar, High Complex CHO, and Low CHO/High Protein) - Setting: laboratory - State: NR - Measurement time points: week 0 and week 24	- 8 KKW group increased preference for high fat/high CHO foods whereas 20 KKW group decreased. - Compensators decreased preferences for high CHO, low fat, and low fat/high CHO foods whereas non-compensators increased. - Difference between liking/wanting: NR	- Food intake: adjusted doubly-labeled water energy intake increased in both exercise groups (relative to control group). No changes in test meal energy intake. - Eating behavior traits: greater reduction in disinhibition in non-compensators relative to compensators. - Body composition: difference in body mass and body fat percentage loss between 20 KKW and control.
Martins et al. 2017 [60] Norway	Sex: females and males BMI status: ≥ 30 kg/m² PA level: inactive (<1 day/week MVPA, <20 min/day < 3 days/week light PA) HIIT: n = 16 (13 completers; 40% males) Age: 34 (8) years BMI baseline = 33.2 (3.5) kg/m² BMI post = NR ½HIIT: n = 16 (9 completers; 80% males) Age: 34 (7) years BMI baseline = 32.4 (2.9) kg/m² BMI post = NR Continuous: n = 14 (13 completers; 60% males) Age: 33 (10) years BMI baseline = 33.3 (2.4) kg/m² BMI post = NR	- Frequency: 3 days/week for 12 weeks - Intensity: HIIT and ½ HIIT (85–90% HR_max), continuous (70% HR_max) - Type: HIIT vs. ½ HIIT vs. continuous cycling - Duration: HIIT (8-s intervals and 12-s recovery for 250 kcal; ~ 20 min), ½ HIIT (8-s intervals and 12-s recovery for 125 kcal; ~ 10 min), continuous exercise (250 kcal; ~ 32 min). - Timing: NR - Supervision: yes - Control group: no; HIIT vs. ½ HIIT vs. continuous	- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste - Setting: laboratory - State: pre and post fixed breakfast (600 kcal) - Measurement time points: week 0 and week 12	- No effect of exercise on liking or wanting - Difference between liking/wanting: no	- Food intake: no change [61] - Eating behavior traits: NR - Body composition: reduction in body weight and trunk and leg % fat mass [61]
Miguet et al. under review France	Sex: females and males (adolescents) BMI status: ≥ 95th percentile for sex and age PA level: inactive (<2 h/week) n = 30 (23% males) Age: 13 (1) years BMI baseline = 35.7 (4.5) kg/m² BMI post = 30.9 (5.0) kg/m²	- Frequency: 4 days/week for 10 months - Intensity: NR - Type: various (aerobic, strength, aquatic and leisure-time activities) - Duration: 60 min - Timing: NR - Supervision: yes - Control group: no As part of an inpatient multidisciplinary weight loss program	- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA - Setting: laboratory - State: pre and post lunch (ad libitum) - Measurement time points: baseline, 5 months, 10 months	- Hungry: Increase in liking at 5 months followed by decrease at 10 months (similar to baseline values). No change in wanting. - Fed: Decrease in liking at 5 and 10 months. No change in wanting. - Hungry to fed: Decrease in liking at 5 and 10 months, but not baseline. No change in wanting. - Difference between liking/wanting: yes; decrease in liking, no change in wanting	- Food intake: decrease in lunch EI at 10 months - Eating behavior traits: decrease in uncontrolled eating and emotional eating at 5 and 10 months - Body composition: decrease in percentage fat mass at 5 and 10 months
Riou et al. 2019 [62] Canada	Sex: females (premenopausal) BMI status: > 27 kg/m² PA level: inactive (< 150 min/week) Low intensity: n = 11 week Moderate intensity: n = 10 Age: 31 (11) years BMI baseline = 35.1 (6.2) kg/m² BMI post = NR (~ 35.5 kg/m²)	- Frequency: 5days/week for 12–14 weeks - Intensity: LOW (40% VO_2reserve) vs. moderate (MOD; 60% VO_2reserve) - Type: aerobic (treadmill or cycle ergometer) - Duration: to 300 kcal (LOW ~ 62 min, MOD ~ 46 min) - Timing: NR - Supervision: 3/5 days - Control group: no; LOW vs. MOD	- LFPQ: liking and implicit wanting bias for fat/taste - Setting: laboratory - State: post breakfast/pre exercise (ad libitum breakfast on first session and quantities replicated on subsequent sessions; LOW ~ 648 kcal, MOD ~ 746 kcal) and post rest (week 4) or exercise (scheduled session; weeks 1 and 12–14) - Measurement time points: week 4, week 1, and week 12–14 (in line with menstrual cycle)	- Decrease in wanting for fat after training. - Increase in liking for savory foods pre to post-exercise at week 1 but decrease at week 12–14. - Difference between liking/wanting: yes; changes in wanting but not liking for fat with training	- Food intake: no change - Eating behavior traits: increase in susceptibility to hunger - Body composition: group by time interaction for body weight and fat mass showing increase in MOD and decrease in LOW
Thivel et al. 2019 [63] France	Sex: females and males (adolescents) BMI status: > 90th percentile for sex and age PA level: NR Eccentric: n = 12 (50% males) Age: 14 (1) years BMI baseline = 34.8 (5.5) kg/m² BMI post = 29.0 (4.5) kg/m² Concentric: n = 12 (50% males) Age: 13 (1) years BMI baseline = 31.8 (3.8) kg/m² BMI post = 27.6 (4.0) kg/m²	- Frequency: 3 days/week for 12 weeks (+2 h/week physical education) - Intensity: 50% up to 70% VO_2peak - Type: eccentric vs. concentric cycling - Duration: 30 up to 45 min - Timing: NR - Supervision: yes - Control group: no; eccentric vs. concentric Phase 2 (week 12–24) of a 24-week inpatient multidisciplinary weight loss program	- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/sweet - Setting: laboratory - State: fasted - Measurement time points: baseline, week 12 and week 24	- Eccentric group: increase in preference for high-fat foods and savory foods (decrease in sweet bias). Increase in implicit wanting for savory foods (decrease in sweet bias). - Group by time interaction showed concentric group increased preference and implicit wanting for sweet foods while eccentric decreased. - Difference between liking/wanting: yes; changes in implicit wanting for sweet foods, no change in liking.	- Food intake: total daily energy intake increased in both groups from baseline to week 24, but only increased in concentric group from week 12 to 24. - Eating behavior traits: NR - Body composition: greater percentage body fat loss after eccentric vs. concentric exercise from week 12 to 24

Reference

Participant characteristics

Exercise/physical activity details

Food reward method

Food reward results

Associations with appetite outcomes

Observational studies

Beaulieu et al. 2017 [34]

Sex: females and males

BMI status: < 29.9 kg/m²

Active: n = 20 (50% males)

Age: 30 (10) years

BMI: 22.6 (1.9) kg/m²

Inactive: n = 19 (42% males)

Age: 30 (9) years

BMI: 23.1 (2.7) kg/m²

PA level assessment: IPAQ and PA monitor (SenseWear)

Active: ≥ 4 exercise sessions/week (MVPA = 182 (67) min/day)

Inactive: ≤ 1 exercise session/week (MVPA = 103 (37) min/day)

Exercise session: ≥ 40 min MVPA

- LFPQ: liking and implicit wanting bias for fat

- Setting: laboratory

- State: pre and post high-fat or high-carbohydrate lunch (ad libitum)

- No difference in food reward (liking and wanting) between groups.

- Difference between liking/wanting: none

- Food intake: energy intake greater in high-fat relative to high-carbohydrate in both groups.

- Eating behavior traits: tendency for restraint to be higher in HiPA compared with LoPA

- Body composition: HiPA had lower body fat percentage than LoPA

Beaulieu et al. 2017 [35]

Sex: females and males

BMI status: < 29.9 kg/m²

HiMVPA: n = 12 (33% males)

Age: 29 (10) years

BMI: 22.4 (2.1) kg/m²

ModMVPA: n = 11 (27% males)

Age: 26 (3) years

BMI: 22.7 (2.2) kg/m²

LoMVPA: n = 11 (27% males)

Age: 30 (11) years

BMI: 23.1 (2.9) kg/m²

PA level assessment: Tertiles of daily MVPA measured by PA monitor (SenseWear)

HiMVPA: 174 (39) min/day

ModMVPA: 121 (15) min/day

LoMVPA: 83 (16) min/day

- LFPQ: liking and implicit wanting bias for fat

- Setting: laboratory

- State: pre and post preloads—HE (~ 700 kcal) and LE (~ 260 kcal) relative to water control (0 kcal)

- No difference in food reward (liking and wanting) between groups

- Difference between liking/wanting: none

- Food intake: greater reduction in liking and wanting after HE preload relative to LE preload in all groups. ModMVPA and HiMVPA reduced ad libitum energy intake at the lunch meal following consumption of the HE compared with the LE preload, while the LoMVPA group did not. No effect of MVPA group on energy intake at dinner or evening snack box.

- Eating behavior traits: no differences between groups

- Body composition: no differences between groups

Drummen et al. 2019 [36•]

Netherlands

Sex: females and males

BMI status: > 25 kg/m²

n = 39 (56% males)

Age: 53 (11) years

BMI: 32.3 (3.7) kg/m²

Participants had impaired fasting glucose and/or glucose tolerance

PA level assessment: Baecke questionnaire (work, sport, and leisure-time physical activity) [37]

Work = 2.6 (0.8) range 1.3–4.3

Sport = 2.6 (0.7) range 1.0–4.0

Leisure = 2.9 (0.7) range 1.5–4.4

Scores ranging from 1 (low level of PA) and 5 (high level of PA)

- fMRI: BOLD signals to high-energy food, low-energy food and non-food images

- Setting: laboratory

- State: after overnight fast

- Inverse association between leisure-time PA and food compared with non-food brain activation in the right thalamus, left middle cingulate gyrus, right precuneus, left putamen, and left angular gyrus.

- Positive association between work-related physical activity and brain activation (disappeared after adjusting for BMI and age)

- No association with sport-related physical activity.

- Difference between liking/wanting: NR

- Food intake: NR

- Eating behavior traits: positive association between brain activation and disinhibition. No association with restraint or susceptibility to hunger.

- Body composition: no association

Horner et al. 2016 [38]

Australia

Sex: males

BMI status: 18–40 kg/m²

Active: n = 22

Age: 29 (8) years

BMI: 24.5 (2.6) kg/m²

Inactive: n = 22

Age: 31 (9) years

BMI: 27.4 (4.2) kg/m²

PA level assessment: Self-report and PA monitor (ActiGraph)

Active: ≥ 4 exercise sessions/week (PA = 709 (239) kcal/day)

Inactive: ≤ 1 exercise session/week (PA = 525 (185) kcal/day)

Exercise session: ≥ 40 min MVPA

- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: post fixed breakfast (400 kcal) and pre lunch 5 h later

- Fed: Active had lower liking for high-fat foods, low-fat sweet foods and for foods overall compared with inactive. Active had greater wanting for low-fat savory foods.

- Hungry: No differences in liking between active and inactive. Active had greater wanting for low-fat savory foods.

- Fed to hungry: Active had greater increase in liking for all food categories combined than inactive.

- Difference between liking/wanting: Faster gastric emptying associated with liking for savory foods and slower gastric emptying associated with greater implicit wanting for high-fat foods.

- Gastric emptying: Inverse association with post-prandial changes in liking low-fat savory foods. Positive association with liking taste bias in hungry state (i.e., faster gastric emptying associated with greater liking for savory foods). No association with liking fat bias nor implicit wanting taste bias. Positive association with implicit wanting fat bias in fed and hungry states. Effects independent of body fat.

Killgore et al. 2013 [39]

USA

Sex: females and males

BMI status: 19.8–34.8 kg/m²

n = 37 (59% males)

Age: 30 (8) years

BMI: 24.5 (3.7) kg/m²

PA level assessment: Self-reported habitual PA in min/week (typical days/week × duration/day)

PA = 151 (160) min/week (range 0–540)

- fMRI: response to high-energy food, low-energy food and non-food images

- Subjective food preferences: desire to eat depicted food item at that moment (VAS)

- Setting: laboratory

- State: 1 h fasted (pre scan food intake 323 (245) kcal)

- Inverse association between habitual PA and fMRI responses (medial orbitofrontal cortex and left insula) to high-energy foods relative to low-energy foods

- Inverse association between habitual PA and preference for high-energy savory foods

- Subjective food preference: fMRI responses positively associated with preference for high-energy savory foods (not for high-energy sweet foods). No association between fMRI responses to high-energy relative to low-energy foods and preference for low-energy foods.

Luo et al. 2018 [40•]

USA

Sex: females and males

BMI status: NR

n = 40 (48% males)

Lean individuals: n = 22 (46% males)

Age: 21 (2) years

BMI: 22.6 (1.9) kg/m²

Individuals with obesity: n = 18 (50% males)

Age: 22 (2) years

BMI: 35.2 (4.0) kg/m²

PA level assessment: Self-reported from 3 to 5 24-h recalls over 2 months (mean daily minutes of MVPA i.e., activities ≥ 3 METs)

Lean individuals: MVPA = 125 (84) min/day

Individuals with obesity:

MVPA = 134 (114) min/day

- fMRI: responses to high-energy food and non-food images

- Setting: laboratory

- State: 9–11 am after overnight fast, task performed 20–30 min after 75 g glucose ingestion

- Inverse association between MVPA and brain response to food cues in middle insula and left postcentral gyrus.

- Individuals with obesity: inverse association between MVPA and brain responses.

- Lean individuals: non-significant inverse association between MVPA and brain responses.

- Association between MVPA and brain responses stronger in males than females.

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Oustric et al. 2018 [41]

Sex: females

BMI status: 18.5–45.0 kg/m²

n = 156

Age: 53 (11) years

BMI: 32.3 (3.7) kg/m²

Pooled data from 6 studies

PA level assessment: Quintiles of daily MVPA measured by PA monitor (SenseWear)

Q1: 25 (8) min/day

Q2: 53 (9) min/day

Q3: 83 (9) min/day

Q4: 120 (12) min/day

Q5: 197 (62) min/day

- LFPQ: liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: hungry (3–10 h fast)

- MVPA inversely associated with liking and implicit wanting fat bias.

- MVPA positively associated with liking taste bias.

- Q5 greater liking and wanting for low-fat foods, while Q1-Q3 greater liking and wanting for high-fat foods.

- Difference between liking/wanting: none

- Food intake: NR

- Eating behavior traits: No association between PA and food cravings. Craving for sweet food (Control of Eating Questionnaire; CoEQ) positively associated with explicit liking and implicit wanting for sweet foods on the LFPQ. Craving for savory foods (CoEQ) associated with LFPQ explicit wanting for savory foods. Craving control (CoEQ) negatively associated with implicit wanting for high-fat foods.

- Body composition: Fat mass index (FMI) but not waist circumference (WC) was inversely associated with explicit liking and implicit wanting for sweet relative to savory foods. WC was positively associated with all liking and implicit wanting for high-fat relative to low-fat foods. FMI was also associated with implicit wanting for high-fat foods.

Acute exercise studies

Alkahtani et al. 2014 [42]

Australia

Sex: males

BMI status: > 25 kg/m²

PA level: sedentary (criteria NR)

n = 12

Age: 29 (4) years

BMI: 29.1 (2.4) kg/m²

- Intensity: moderate-intensity interval training (MIIT; alternating between ± 20% FAT_max) vs. high-intensity interval training (HIIT; 85% VO_2peak)

- Type: cycle ergometer

- Duration: MIIT 5-min stages at ±20% FAT_max for 30 min, HIIT 15-s intervals and 15-s recovery (workload matched to MIIT; ~ 18 min)

- Timing: morning

- Control condition: no; MIIT vs. HIIT

- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: before and after exercise (after an overnight fast)

- Decrease in wanting and increase in liking for all the food categories independent of the intensity

- Difference between liking/wanting: liking increased while wanting decreased but without a control this response might be due to the effect of time rather than exercise

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Alkahtani et al. 2019 [43]

Saudi Arabia

Sex: males

BMI status: NR

PA level: moderately active (2–5 h structured aerobic exercise/week)

n = 14 (8 for food reward data)

Age: 24 (6) years

BMI: 23.4 (3.3) kg/m²

- Intensity: moderate (60% VO_2max) interspersed with low (30% VO_2max)

- Type: contraction type eccentric (downhill running at − 12% inclination) vs. concentric (flat running)

- Duration: 5 stages of 8 min at 60%VO_2max/2 min at 30%VO_2max

- Timing: morning

- Control condition: yes; no exercise

- LFPQ: liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: before, after exercise and 24 h after exercise (before an ad libitum lunch)

- No change in food reward after exercise

- Difference between liking/wanting: greater liking of savory foods over sweet foods in downhill running than front running

- Food intake: no change

- Eating behavior traits: NR

- Body composition: NR

Crabtree et al. 2014 [44]

Sex: males

BMI status: 21.8–26.6 kg/m²

PA level: moderately active (2 h/week)

n = 16

Age: 23 (3) years

BMI: 24.2 (2.4) kg/m²

- Intensity: high (70% VO_2max)

- Type: treadmill run

- Duration: 60 min

- Timing: morning

- Control condition: yes; no exercise

- fMRI: BOLD signals to high- and low-energy food cues compared with non-food pictures

- Setting: laboratory

- State: fasted

- Decreased activation in the pallidum for high-energy food and increase for low-energy food after exercise

- Difference between liking/wanting: NR

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Evero et al. 2012 [45]

USA

Sex: females and males

BMI status: < 25 kg/m²

PA level: habitually active (> 3 h/week)

n = 30 (57% males)

Age: 22 (4) years

BMI: 23.6 (2.2) kg/m²

- Intensity: high (83% HR_max)

- Type: cycle ergometer

- Duration: 60 min

- Timing: morning

- Control condition: yes; no exercise

- fMRI: BOLD signals to high- and low-energy food cues compared with neutral control

- Setting: laboratory

- State: fMRI was performed after exercise after an 8–12 h overnight fast

- Exercise reduced the neuronal response to food cues in brain regions related with food reward (i.e., insula, putamen, rolandic operculum)

- Difference between liking/wanting: decrease in regions related to liking and wanting but no behavioral measures

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Farah et al. 2012 [46]

Sex: females and males

BMI status: no limits

PA level: NR

n = 27 (52% males)

Female: n = 13

Age: 26 (3) years

BMI: BMI: 22.8 (3.1) kg/m²

Male: n = 14

Age: 36 (10) years

BMI: 26.1 (3.3) kg/m²

- Intensity: moderate (6 METs)

- Type: treadmill walk

- Duration: 60 min

- Timing: morning

- Control condition: yes; no exercise

- VAS: liking

- Setting: laboratory

- State: immediately, 60, 120, and 180 min after exercise (overnight fasted)

- No change in liking

- Difference between liking/wanting: wanting not measured

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Finlayson et al. 2009 [47] UK

Sex: females

BMI status: < 25 kg/m²

PA level: 2.4 (1.2) engagements/week

n = 24

Age: 24 (6) years

BMI: 22.3 (2.9) kg/m²

- Intensity: moderate (70% HR_max)

- Type: cycle ergometer

- Duration: 50 min

- Timing: morning

- Control condition: yes, no exercise

- LFPQ: relative preference (food choice), liking and implicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: before and after exercise (2 h after a fixed breakfast, kcal NR) and after an ad libitum lunch 30-min post exercise

- Increase in implicit wanting after exercise in those who compensated or ate more at the ad libitum lunch in response to exercise

- Difference between liking/wanting: Changes in implicit wanting but not liking

- Food intake: After exercise some individuals increased their energy intake (compensators) and had enhanced implicit wanting

- Eating behavior traits: NR

- Body composition: NR

Martins et al. 2015 [48]

Norway

Sex: females and males

BMI status: > 25 kg/m²

PA level: sedentary (criteria NR)

n = 12 (42% males)

Age: 33 (10) years

BMI: 32.3 (2.7) kg/m²

- Intensity: HIIT and ½ HIIT (all out; average ~ 85% HR_max), continuous (70% HR_max)

- Type: HIIT vs. ½ HIIT vs. continuous cycling

- Duration: HIIT (8-s intervals and 12-s recovery for 250 kcal; ~ 18 min), ½ HIIT (8-s intervals and 12-s recovery for 125 kcal; ~ 9 min), continuous exercise (250 kcal; ~ 27 min)

- Timing: morning

- Control condition: yes; no exercise

- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat

- Setting: laboratory

- State: before and after exercise/ before an ad libitum lunch

(standardized breakfast of 600 kcal consumed 1 h before exercise/rest)

- No change in food reward

- Difference between liking/wanting: no differences

- Food intake: no change

- Eating behavior traits: NR

- Body composition: NR

McNeil et al. 2015 [49]

Canada

Sex: females and males

BMI status: < 25 kg/m²

PA level: inactive (<150 min/week)

n = 16 (50% males)

Age: 22 (3) years

BMI: 22.8 (1.8) kg/m²

- Intensity: high (aerobic 70% VO_2peak, resistance 70% 1-repetition maximum)

- Type: aerobic vs. resistance

- Duration: aerobic ~ 24 min, resistance ~ 86 min (matched for energy expenditure at 4 kcal/kg; ~ 275 kcal)

- Timing: morning

- Control condition: yes; no exercise

- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: pre and post ad libitum lunch 30 min after exercise (standardized breakfast of 534 kcal consumed ~ 1.5 h before exercise)

- Decrease in the relative preference for high-fat relative to low-fat foods after both exercise

- Decrease in liking for high-fat foods following resistance, but not aerobic

- Difference between liking/wanting: change in food choice and liking, but not implicit wanting

- Food intake: no change

- Eating behavior traits: NR

- Body composition: no difference in bodyweight

Miguet et al. 2018 [50]

France

Sex: females and males (adolescents)

BMI status: > 29.9 kg/m²

PA level: inactive (<2 h/week)

n = 33 (36% males)

Age: 13 (1) years

BMI: 35.0 (4.3) kg/m²

- Intensity: high (70%, 75%, 80%, 85%, and 90% HR_max)

- Type: high-intensity interval training cycling

- Duration: 5 × 2-min increasing intensity intervals followed by 30-s recovery (15 min)

- Timing: morning

- Control condition: yes; no exercise

- LFPQ: relative preference, liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: pre and post ad libitum lunch 30 min after exercise (standardized breakfast of 500 kcal consumed 2.5 h before exercise)

- Decrease in implicit wanting for sweet in the exercise condition vs. increase in the control

- Difference between liking/wanting: implicit wanting but not liking is decreasing

- Food intake: decrease in energy intake at lunch and dinner

- Eating behavior traits: NR

- Body composition: NR

Thackray et al. unpublished

Sex: females and males

BMI status: 18.5–29.9 kg/m²

PA level: habitually active

n = 32

Age: 23 (2) years

BMI: 23.9 (2.6) kg/m²

- Intensity: self-determined moderate-to-high intensity (RPE of 15 “hard”)

- Type: swimming vs. cycling

- Duration: 6 × 8-min intervals with 2-min recovery

- Timing: morning

- Control condition: yes, no exercise

- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: post-exercise (3 h after fixed breakfast of 650 kcal for males, 525 kcal for females, and before ad libitum lunch meal)

- Tendency for a main effect of trial for implicit wanting fat bias (post hoc: cycling < control, cycling < swimming).

- No impact of swimming or cycling on other food reward parameters.

- Difference between liking/wanting: Changes in implicit wanting but not liking

- Food intake: increase in ad libitum energy intake after swimming but not after cycling

- Eating behavior traits: NR

- Body composition: NR

Thivel et al. 2019 [51]

France

Sex: females and males

BMI status: < 25 kg/m²

PA level: moderately active (150–240 min/week)

n = 19 (52% males)

Age: 21 (1) years

BMI: 22.3 (2.9) kg/m²

- Intensity: low 50% VO_2max, high 75% VO_2max

- Type: cycling

- Duration: low intensity 45 min, high intensity 30 min

- Timing: morning

- Control condition: yes; no exercise

- LFPQ: relative preference, liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: pre and post fixed lunch 30 min after exercise (females 750 kcal and males 900 kcal; standardized breakfast of 500 kcal consumed 3 h before exercise).

- No change in food reward

- Difference between liking/wanting: no differences

- Food intake: (self-reported) no change

- Eating behavior traits: NR

- Body composition: NR

Saanijoki et al. 2018 [52]

Finland

Sex: males

BMI status: 19.9–26.9 kg/m²

PA level: NR

n = 24

Age: 27 (5) years

BMI: 23.5 (1.6) kg/m²

- Intensity: moderate (74% HR_max)

- Type: aerobic cycling

- Duration: 60 min

- Timing: NR

- Control condition: yes; no exercise

- fMRI: BOLD signals to palatable and non-palatable foods compared with neutral control (cars)

- Setting: laboratory

- State: post-exercise fasted for 3 h before the scans

- No effects of exercise vs rest on neuronal responses.

- Individual variability in the BOLD signals after exercise might be explained by changes in the brain opioid system. Participants who showed most increases in endogenous opioid release also had highest anticipatory fMRI reward responses following the exercise

- Difference between liking/wanting: not assessed

- Food intake: NR

- Eating behavior traits: NR

- Body composition: NR

Chronic exercise training studies

Alkahtani et al. 2014 [53]

Australia

Sex: males

BMI status: ≥ 25 kg/m²

PA level: inactive (criteria NR)

n = 10

Age: 29 (4) years Moderate-intensity interval training (MIIT):

BMI baseline = 30.7 (3.5) kg/m²

BMI post = 30.8 (3.5) kg/m²

High-intensity interval training (HIIT):

BMI baseline = 30.9 (3.2) kg/m²

BMI post = 30.9 (3.2) kg/m²

- Frequency: 3 days/week for 4 week (cross-over with 6-week washout)

- Intensity: MIIT ± 20% workload at 45% VO_2peak, HIIT 90% VO_2peak

- Type: MIIT vs. HIIT

- Duration: 30–45 min (MIIT 5-min stages alternating between ± 20% workload, HIIT 30-s intervals and 30-s recovery)

- Timing: NR

- Supervision: yes

- Control group: no; MIIT vs. HIIT

- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: pre and post 45-min cycling at 45% VO_2max

- Measurement time points: week 0 and week 4 in each intervention

- Tendency for explicit liking for high-fat non-sweet foods after acute exercise to increase with MIIT and decrease with HIIT.

- No changes in wanting.

- Difference between liking/wanting: yes; changes in liking but not wanting

- Food intake: No effects of training on food intake or energy intake. Tendency for fat intake (g) and % energy from fat to increase after MIIT.

- Eating behavior traits: NR

- Body composition: NR

Beaulieu et al. 2019 [54]

Sex: females and males

BMI status: 26.0–38.0 kg/m²

PA level: inactive (≤ 2 h/week)

Exercisers: n = 46 (35% males)

Age: 43 (8) years

BMI baseline = 30.5 (3.8) kg/m²

BMI post = 29.9 (4.0) kg/m²

Controls: n = 15 (40% males)

Age: 41 (11) years

BMI baseline = 31.4 (3.7) kg/m²

BMI post = 31.8 (3.9) kg/m²

- Frequency: 5 days/week for 12 weeks

- Intensity: 70% HR_max

- Type: aerobic (treadmill, rower, cycle ergometer, and elliptical)

- Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min)

- Timing: NR

- Supervision: yes

- Control group: yes; no exercise

- LFPQ: liking and implicit wanting bias for fat

- Setting: laboratory

- State: pre and post fixed lunch (high-fat or high-CHO; 800 kcal)

- Measurement time points: week 0 and week 12

- Decrease in wanting after training

- No change in liking

- Difference between liking/wanting: yes; changes in wanting but not liking

- Food intake: reduction in high fat ad libitum dinner intake but no change in daily high-fat energy intake [55]

- Eating behavior traits: decrease in disinhibition and binge eating score

- Body composition: reduction in body weight and percentage body fat, but not associated with changes in wanting

Cornier et al. 2012 [56]

USA

Sex: females and males

BMI status: > 25 kg/m²

PA level: NR

n = 12 (42% males)

Age: 38 (10) years

BMI baseline = 33.3 (4.3) kg/m²

BMI post = NR

- Frequency: 5 days/week for 24 weeks

- Intensity: up to 75% VO_2max

- Type: treadmill

- Duration: up to 500 kcal/day (40–60 min/day)

- Timing: NR

- Supervision: yes

- Control group: no

- fMRI: responses to food vs non-food cues

- Setting: laboratory

- State: fasted without exercise for 24 h (chronic exercise) and fasted within 30 min of acute exercise (chronic+acute; 500 kcal 60–75% VO_2max for 40-60 min)

- Measurement time points: week 0 and week 24

- Chronic exercise: reduction in neuronal responses observed in the bilateral parietal cortices, left insula and visual cortex.

- Chronic + acute exercise: intermediate attenuation of the response to visual food cues in brain region important in food regulation compared with chronic exercise and baseline

- Difference between liking/wanting: NR

- Food intake: self-reported energy intake lower after training compared with baseline but no change in macronutrient intake. No association with changes in neuronal responses.

- Eating behavior traits: no change in dietary restraint or disinhibition

- Body composition: reduction in body fat percentage. Changes in anterior insula responses positively associated with changes in body mass and fat mass.

Finlayson et al. 2011 [57]

Sex: females and males

BMI status: 26.0–38.0 kg/m²

PA level: inactive (≤2 h/week)

Responders (non-compensators): n = 20 (43% males)

Age: 41 (9) years

BMI baseline = 32.3 (4.3) kg/m²

BMI post = 30.9 (4.3) kg/m²

Non-Responders (compensators): n = 14 (50% males)

Age: 37 (12) years

BMI baseline = 29.7 (2.2) kg/m²

BMI post = 29.3 (2.5) kg/m²

- Frequency: 5 days/week for 12 weeks

- Intensity: 70% HR_max

- Type: aerobic (treadmill, rower, cycle ergometer, and elliptical)

- Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min)

- Timing: NR

- Supervision: yes

- Control group: no

- LFPQ: relative preference (food choice), liking, explicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: pre and post acute exercise (scheduled session)

- Measurement time points: week 0 and week 12

- Increase in liking after exercise in non-responders compared with responders at baseline and week 12.

- Increase in explicit wanting for high-fat sweet foods in non-responders.

- Increase in relative preference for high-fat sweet food in non-responders.

- Difference between liking/wanting: implicit wanting NR

- Food intake: NR

- Eating behavior traits: NR

- Body composition: greater fat mass loss in responders

Finlayson et al. unpublished

Sex: females and males

BMI status: 26.0–38.0 kg/m²

PA level: inactive (≤ 2 h/week)

Non-compensators: n = 15 (33% males)

Age: 42 (8) years

BMI baseline = 30.7 (4.9) kg/m²

BMI post = 29.1 (5.0) kg/m²

Compensators: n = 15 (33% males)

Age: 41 (9) years

BMI baseline = 31.8 (3.7) kg/m²

BMI post = 32.1 (3.9) kg/m²

Controls: n = 15 (33% males)

Age: 41 (11) years

BMI baseline = 31.4 (3.7) kg/m²

BMI post = 31.8 (3.9) kg/m²

- Frequency: 5 days/week for 12 weeks

- Intensity: 70% HR_max

- Type: aerobic (treadmill, rower, cycle ergometer, and elliptical)

- Duration: 500 kcal (males ~ 40–45 min, females ~ 60 min)

- Timing: NR

- Supervision: yes

- Control group: yes, no exercise

- LFPQ: liking and implicit wanting bias for fat

- Setting: laboratory

- State: pre and post fixed lunch (high-fat or high-CHO; 800 kcal)

- Measurement time points: week 0 and week 12

- Non-compensators showed a smaller liking and implicit wanting for high-fat food.

- At baseline, compensators showed a strong liking and wanting for high-fat food whereas non-compensators showed no difference between high-fat and low-fat food.

- Greater baseline reward for high-fat food in compensators reduced following the exercise intervention.

- In the non-compensators, small increase in liking for high-fat food after exercise training, but a simultaneous decrease in wanting for high-fat food.

- Difference between liking/wanting: yes, non-compensators increased liking but decreased wanting for high-fat food post-intervention.

- Food intake: tendency for non-compensators to decrease ad libitum dinner meal size from baseline to post-intervention that was not seen in compensators.

- Eating behavior traits: NR

- Body composition: significant reduction in BMI, body mass, fat mass and WC in non-compensators, whereas no changes in compensators, and increase in body mass and WC.

Martin et al. 2019 [58]

USA

Sex: females and males

BMI status: 25–45 kg/m²

PA level: inactive (< 20 min < 3 days/week)

8 kcal/kg body weight/week (KKW): n = 59 (27% males)

Age: 48 (11) years

BMI baseline = 31.4 (4.6) kg/m²

BMI post = NR (~ 31.3 kg/m²)

20 KKW: n = 51 (29% males)

Age: 49 (12) years

BMI baseline = 30.6 (4.4) kg/m²

BMI post = NR (~ 30.0 kg/m²)

Control: n = 61 (26% males)

Age: 50 (1) years

BMI baseline = 32.3 (4.8) kg/m²

BMI post = NR (~ 32.2 kg/m²)

Pooled exercisers (n = 110) divided (median split) into compensators/non-compensators based on actual and predicted weight loss.

- Frequency: 3–5 days/week (self-selected) for 24 weeks

- Intensity: 65–85% VO_2peak (self-selected)

- Type: treadmill

- Duration: 8 KKW ~ 35 min/session (~ 700 kcal/week) vs. 20 KKW ~ 55 min/session (~ 1760 kcal/week)

- Timing: NR

- Supervision: yes

- Control group: yes; no exercise

- Food Preference Questionnaire [59] preferences for food classified as alongside 2 components: fat (2 factors: High Fat and Low Fat) and carbohydrate (3 factors: High Simple Sugar, High Complex CHO, and Low CHO/High Protein)

- Setting: laboratory

- State: NR

- Measurement time points: week 0 and week 24

- 8 KKW group increased preference for high fat/high CHO foods whereas 20 KKW group decreased.

- Compensators decreased preferences for high CHO, low fat, and low fat/high CHO foods whereas non-compensators increased.

- Difference between liking/wanting: NR

- Food intake: adjusted doubly-labeled water energy intake increased in both exercise groups (relative to control group). No changes in test meal energy intake.

- Eating behavior traits: greater reduction in disinhibition in non-compensators relative to compensators.

- Body composition: difference in body mass and body fat percentage loss between 20 KKW and control.

Martins et al. 2017 [60]

Norway

Sex: females and males

BMI status: ≥ 30 kg/m²

PA level: inactive (<1 day/week MVPA, <20 min/day < 3 days/week light PA)

HIIT: n = 16 (13 completers; 40% males)

Age: 34 (8) years

BMI baseline = 33.2 (3.5) kg/m²

BMI post = NR

½HIIT: n = 16 (9 completers; 80% males)

Age: 34 (7) years

BMI baseline = 32.4 (2.9) kg/m²

BMI post = NR

Continuous: n = 14 (13 completers; 60% males)

Age: 33 (10) years

BMI baseline = 33.3 (2.4) kg/m²

BMI post = NR

- Frequency: 3 days/week for 12 weeks

- Intensity: HIIT and ½ HIIT (85–90% HR_max), continuous (70% HR_max)

- Type: HIIT vs. ½ HIIT vs. continuous cycling

- Duration: HIIT (8-s intervals and 12-s recovery for 250 kcal; ~ 20 min), ½ HIIT (8-s intervals and 12-s recovery for 125 kcal; ~ 10 min), continuous exercise (250 kcal; ~ 32 min).

- Timing: NR

- Supervision: yes

- Control group: no; HIIT vs. ½ HIIT vs. continuous

- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: pre and post fixed breakfast (600 kcal)

- Measurement time points: week 0 and week 12

- No effect of exercise on liking or wanting

- Difference between liking/wanting: no

- Food intake: no change [61]

- Eating behavior traits: NR

- Body composition: reduction in body weight and trunk and leg % fat mass [61]

Miguet et al. under review

France

Sex: females and males (adolescents)

BMI status: ≥ 95th percentile for sex and age

PA level: inactive (<2 h/week)

n = 30 (23% males)

Age: 13 (1) years

BMI baseline = 35.7 (4.5) kg/m²

BMI post = 30.9 (5.0) kg/m²

- Frequency: 4 days/week for 10 months

- Intensity: NR

- Type: various (aerobic, strength, aquatic and leisure-time activities)

- Duration: 60 min

- Timing: NR

- Supervision: yes

- Control group: no

As part of an inpatient multidisciplinary weight loss program

- LFPQ: liking and implicit wanting for HFSW, LFSW, HFSA, LFSA

- Setting: laboratory

- State: pre and post lunch (ad libitum)

- Measurement time points: baseline, 5 months, 10 months

- Hungry: Increase in liking at 5 months followed by decrease at 10 months (similar to baseline values). No change in wanting.

- Fed: Decrease in liking at 5 and 10 months. No change in wanting.

- Hungry to fed: Decrease in liking at 5 and 10 months, but not baseline. No change in wanting.

- Difference between liking/wanting: yes; decrease in liking, no change in wanting

- Food intake: decrease in lunch EI at 10 months

- Eating behavior traits: decrease in uncontrolled eating and emotional eating at 5 and 10 months

- Body composition: decrease in percentage fat mass at 5 and 10 months

Riou et al. 2019 [62]

Canada

Sex: females (premenopausal)

BMI status: > 27 kg/m²

PA level: inactive (< 150 min/week)

Low intensity: n = 11

week

Moderate intensity: n = 10

Age: 31 (11) years

BMI baseline = 35.1 (6.2) kg/m²

BMI post = NR (~ 35.5 kg/m²)

- Frequency: 5days/week for 12–14 weeks

- Intensity: LOW (40% VO_2reserve) vs. moderate (MOD; 60% VO_2reserve)

- Type: aerobic (treadmill or cycle ergometer)

- Duration: to 300 kcal (LOW ~ 62 min, MOD ~ 46 min)

- Timing: NR

- Supervision: 3/5 days

- Control group: no; LOW vs. MOD

- LFPQ: liking and implicit wanting bias for fat/taste

- Setting: laboratory

- State: post breakfast/pre exercise (ad libitum breakfast on first session and quantities replicated on subsequent sessions; LOW ~ 648 kcal, MOD ~ 746 kcal) and post rest (week 4) or exercise (scheduled session; weeks 1 and 12–14)

- Measurement time points: week 4, week 1, and week 12–14 (in line with menstrual cycle)

- Decrease in wanting for fat after training.

- Increase in liking for savory foods pre to post-exercise at week 1 but decrease at week 12–14.

- Difference between liking/wanting: yes; changes in wanting but not liking for fat with training

- Food intake: no change

- Eating behavior traits: increase in susceptibility to hunger

- Body composition: group by time interaction for body weight and fat mass showing increase in MOD and decrease in LOW

Thivel et al. 2019 [63]

France

Sex: females and males (adolescents)

BMI status: > 90th percentile for sex and age

PA level: NR

Eccentric: n = 12 (50% males)

Age: 14 (1) years

BMI baseline = 34.8 (5.5) kg/m²

BMI post = 29.0 (4.5) kg/m²

Concentric: n = 12 (50% males)

Age: 13 (1) years

BMI baseline = 31.8 (3.8) kg/m²

BMI post = 27.6 (4.0) kg/m²

- Frequency: 3 days/week for 12 weeks (+2 h/week physical education)

- Intensity: 50% up to 70% VO_2peak

- Type: eccentric vs. concentric cycling

- Duration: 30 up to 45 min

- Timing: NR

- Supervision: yes

- Control group: no; eccentric vs. concentric

Phase 2 (week 12–24) of a 24-week inpatient multidisciplinary weight loss program

- LFPQ: relative preference (food choice), liking and implicit wanting bias for fat/sweet

- Setting: laboratory

- State: fasted

- Measurement time points: baseline, week 12 and week 24

- Eccentric group: increase in preference for high-fat foods and savory foods (decrease in sweet bias). Increase in implicit wanting for savory foods (decrease in sweet bias).

- Group by time interaction showed concentric group increased preference and implicit wanting for sweet foods while eccentric decreased.

- Difference between liking/wanting: yes; changes in implicit wanting for sweet foods, no change in liking.

- Food intake: total daily energy intake increased in both groups from baseline to week 24, but only increased in concentric group from week 12 to 24.

- Eating behavior traits: NR

- Body composition: greater percentage body fat loss after eccentric vs. concentric exercise from week 12 to 24

Data are means (SD). CoEQ, Control of Eating Questionnaire; FMI, fat mass index; fMRI, functional magnetic resonance imaging; LFPQ, Leeds Food Preference Questionnaire; LFSA, low-fat savory; LFSW, low-fat sweet; HFSA, high-fat savory; HFSW, high-fat sweet; HIIT, high-intensity interval training; MET, metabolic equivalent of a task; MVPA, moderate-to-vigorous physical activity; NR, not reported; PA, physical activity; VAS, visual analogue scales; WC, waist circumference