Human eating patterns are heterogeneous and erratic [1
], and meal timing is variously influenced by socio-cultural [2
] and hereditary [3
] factors. An accumulating body of research indicates that the timing of food intake can influence energy balance and metabolism, with evening eating associated with greater energy intake and BMI compared to morning eating [4
]. Indeed, scientific bodies now recognize that patterns of daily energy intake, including meal timing and frequency, influence the management of body weight and cardiometabolic risk [7
]. Dietary interventions targeting food timing have been proposed for the treatment of obesity and obesity-related diseases, but evidence is limited [8
]. Some studies have reported that greater food intake in the morning compared to the evening leads to more favorable weight loss outcomes, independent of total energy intake [9
]. While the mechanisms for these effects are yet to be fully elucidated, they could include an altered metabolic response to meals [11
], satiety [13
], food choice and macronutrient composition [1
] or other energy balance-related behaviors such as physical activity [12
]. Indeed, observational evidence suggests the satiety value of food may be greater earlier in the day compared to later [13
There is individual variability in diurnal rhythms for preferred timing of sleep, physical activity and eating, defined as chronotype. Chronotype differences can be measured through, for example, Horne and Ostberg’s Morningness–Eveningness Questionnaire [15
]. Evening (late) chronotype is associated with higher BMI [16
], binge eating behaviors [17
] and greater evening energy intake [18
], specifically from fewer fruits and vegetables [4
], and more fast food and alcohol [19
]. In contrast, morning (early) chronotype is positively associated with cognitive restraint, and inversely associated with disinhibition and susceptibility to hunger [20
]. Interestingly, there is evidence suggesting that the relationship between meal timing and BMI is influenced by chronotype [16
]. This suggests a potential interaction between meal timing and chronotype on appetite and eating behavior.
A very limited number of studies have examined diurnal or circadian variations in appetite under controlled conditions. Indeed, most of these studies have relied on self-reported food intake obtained from, e.g., food diaries. Measurement of appetite across the diurnal cycle is methodologically challenging as the appropriateness of components and size of the meals at different times of the day may influence outcomes. Therefore, there is a need to standardize the methodology across the day to assess diurnal variations in appetite and food reward.
Food reward includes the perception of liking food, but also involves other motivational drivers of food choice and intake, such as wanting [21
]. Liking and wanting influence the strength of satiety, direction of food preferences and control over food intake [22
]. They also play a role in expressions of overeating such as binge eating [23
]. To our knowledge, however, no studies have assessed diurnal variations in food reward in humans. The Leeds Food Preference Questionnaire (LFPQ) methodology provides a framework, based on responses to an array of food images, to interpret the impact of interventions on liking and wanting as separate and distinct processes; however, an array of images that is equally suitable for different times of day has yet to be developed [24
Therefore, the aim of this study was to examine whether meal timing and chronotype affect appetite and reward responses to food. Importantly, as a preliminary step, we sought to validate a test meal and LFPQ array of food images that were time-of-day appropriate for both early and late meal timing in young British adults.
With increasing evidence suggesting that meal timing may affect appetite and body weight control, experimental studies are required to assess outcomes and investigate underlying mechanisms using valid and reliable methodologies. The current study examined the impact of meal timing and chronotype under controlled laboratory conditions. We firstly validated a test meal and an array of food images that were time-of-day appropriate, then we assessed appetite and food reward responses in early (8–10 a.m.) and late (4–6 p.m.) meal timing sessions. Clear diurnal patterns of appetite and food reward—both lowest in the morning—as well as chronotype differences were observed, with the impact of meal timing and chronotype appearing to be an additive effect.
We found lower overall appetite in AM compared to PM, consistent with previous observational studies [13
]. Forced desynchrony studies suggest the presence of an endogenous circadian rhythm in appetite, with greater evening hunger independent of wake time and energy intake [38
]. This circadian pattern in appetite may stem from a greater metabolic response to meals in the morning (e.g., thermic effect of food, glucose tolerance, gastric emptying) [11
], potentially promoting stronger satiety response [13
]. Furthermore, there is evidence of circadian rhythms in acylated ghrelin secretion paralleling those of hunger [40
], which may also explain the differences in appetite observed in the current study. Evidence regarding circadian patterns in satiety-related peptides such as glucagon-like peptide-1, peptide YY or cholecystokinin remains to be demonstrated. In the current study, there was no difference in the degree of appetite suppression between both meal timing conditions and chronotypes, suggesting a similar satiety response to food. Interestingly, the early chronotypes did appear to have slightly greater suppression of appetite in response to the test meal (exploratory post hoc t-test revealed a significant difference at 60 min post-meal, p
< 0.05). It is possible that a larger meal (>300 kcal) might have shown differential effects on satiety responsiveness in the morning, as the energy content of the test meal used in the current study was more typical of a breakfast meal than an evening meal [13
]. Nevertheless, overall appetite was lower in AM. Furthermore, test meal size was not calibrated according to BMI in this study, which may have improved the sensitivity of our design.
While appetite suppression in response to the test meal was similar across all conditions and groups, the test meal was rated more filling in AM compared to PM, and by early compared to late chronotypes. The lowest values observed were from the early chronotype in the AM condition, similar to their appetite ratings, suggesting an additive effect of meal timing and chronotype on perceived test meal fillingness. An uncoupling between subjective appetite response to food and measured satiety ratio (duration of after-meal interval divided by meal size from 7-day food diaries) has been previously reported in a large free-living sample (n
= 867, age = 36 ± 14 years and BMI = 24.5 ± 4.3 kg/m2
]. Future studies utilizing more objective and controlled measures of satiety, such as a preload-test meal protocol, may help clarify these findings.
This is the first study to assess diurnal rhythms in food reward. We used the diurnal LFPQ with food image categories validated to be appropriate for early and late consumption to examine behavioral responses in explicit liking and implicit desire for high-fat relative to low-fat foods. Both liking and wanting scores were lower in AM relative to PM. Some small studies have examined diurnal and circadian rhythms of non-food reward in healthy young adults (with no information on weight status). One study assessed the influence of time-of-day on general liking and wanting, including taste–smell, among five other components, over seven free-living days using VAS on a smartphone software [42
]. Peaks in liking and wanting were achieved at 6–7 p.m. Byrne & Murray [43
] showed a peak in wanting (measured with the automatic Balloon Analogue Risk Task and International Affective Picture System arousal response) at 2 p.m. relative to 10 a.m. and 7 p.m. in males, while no diurnal variation in liking (measured with the International Affective Picture System pleasure response) was observed. Another study found time-of-day effects in neural (fMRI) response to a monetary reward task in the ventral striatum, with greater responses shown in later (~5 p.m.) relative to earlier (~10 a.m.) scans [44
]. Lastly, a forced desynchrony study showed a circadian rhythm in positive affect and reward activation, operationalized as heart rate (calculated from ECG) and reaction time during the Fowles motor task (performed every 2–3 h per 28-h day), which coincided with core body temperature across circadian phases, peaking at 180°–240° (approx. 5–9 p.m.) [45
]. While these non-food reward studies suggest circadian peaks in reward activation in the late afternoon/early evening, this may not directly compare to food reward and eating behaviors, as greater later evening and night-time eating has been associated with greater daily energy intake [13
] and BMI [4
]. Clearly, more work investigating extended time periods is needed to understand diurnal and circadian rhythms in food reward and their impact on food intake and susceptibility to overconsumption.
In terms of the impact of chronotype on food reward, in the whole sample, MEQ score was inversely associated with wanting but not liking. Interestingly, wanting scores were greatest in the late chronotypes in PM. While not directly comparable to the current study, a small study found that individuals with obesity and evening hyperphagia (consumption of ≥25% daily energy intake after evening meal) showed different mid-day pre- and post-meal neural response to food cues compared to matched-control participants [46
]. This suggests that circadian eating patterns may be associated with altered diurnal neural responses to food. However, as that study only measured brain activity at one time-of-day in a specific sample, larger studies in populations with different cultural practices and objective measures of food intake are needed to clarify the role of food reward on food intake across the day, and if this is dependent upon chronotype.
We found an inverse association between MEQ score and BMI, similar to past studies showing earlier chronotypes associated with lower BMI [16
]. However, unlike a previous larger study [20
], we found no association between MEQ score and TFEQ factors, which may be due to our relatively smaller sample size. Furthermore, we found no association between MEQ score and self-reported daily energy intake from dietary record. However, as prior studies have found chronotype differences in daily energy intake patterns (i.e., proportion of energy intake in morning vs. evening) and not necessarily in daily energy intake [18
], classifying energy intake patterns according to clock or circadian time may have been required to see differences in energy intake between chronotypes in the current study.
While this experimental study is one of the first to simultaneously assess effects of meal timing on appetite and food reward and associations with chronotype, several limitations should be acknowledged and addressed in future research. While participants were requested to fast ≥3 h prior to testing, some participants may have incurred a longer fasting period in the early meal timing condition than the late meal timing condition (i.e., if they fasted overnight), but unfortunately the actual duration of the fasting period was not measured. MEQ scores in our sample shifted relatively toward the morning/intermediate range of the scale and the median split applied to the data did not result in an evening chronotype group comparable to the definition proposed by the original authors (i.e., scores < 41 as indicative of ‘evening types’) [14
]. Meal timing was based on clock time and not individual circadian time. Menstrual cycle was not controlled for, which may have influenced the results. Finally, this study was conducted in young British adults with a relatively healthy BMI, which limits the generalization of the findings and of the specific foods validated for this methodology. Future studies outside of the UK should culturally adapt and validate food images for the diurnal assessment of food reward using the Leeds Food Preference Questionnaire in their specific population [34