Principles and Validations of an Artificial Intelligence-Based Recommender System Suggesting Acceptable Food Changes.

BACKGROUND: Along with the popularity of smartphones, artificial intelligence-based personalized suggestions can be seen as promising ways to change eating habits toward more desirable diets. OBJECTIVES: Two issues raised by such technologies were addressed in this study. The first hypothesis tested is a recommender system based on automatically learning simple association rules between dishes of the same meal that would make it possible to identify plausible substitutions for the consumer. The second hypothesis tested is that for an identical set of dietary-swaps suggestions, the more the user is-or thinks to be-involved in the process of identifying the suggestion, the higher is their probability of accepting the suggestion. METHODS: Three studies are presented in this article, first, we present the principles of an algorithm to mine plausible substitutions from a large food consumption database. Second, we evaluate the plausibility of these automatically mined suggestions through the results of online tests conducted for a group of 255 adult participants. Afterward, we investigated the persuasiveness of 3 suggestion methods of such recommendations in a population of 27 healthy adult volunteers through a custom designed smartphone application. RESULTS: The results firstly indicated that a method based on automatic learning of substitution rules between foods performed relatively well identifying plausible swaps suggestions. Regarding the form that should be used to suggest, we found that when users are involved in selecting the most appropriate recommendation for them, the resulting suggestions were more accepted (OR = 3.168; P < 0.0004). CONCLUSIONS: This work indicates that food recommendation algorithms can gain efficiency by taking into account the consumption context and user engagement in the recommendation process. Further research is warranted to identify nutritionally relevant suggestions.

Diet Protein Content and Individual Phenotype Affect Food Intake and Protein Appetence in Rats.

BACKGROUND: A low-protein diet can induce compensatory intake of excess energy. This must be better evaluated to anticipate the obesogenic risk that may result from the dietary recommendations for reducing animal protein consumption. OBJECTIVES: We aimed to further characterize the behavioral and physiological responses to a reduction in dietary protein and to identify the determinants of protein appetite. METHODS: Thirty-two male Wistar rats [4 wk old, (mean ± SEM) 135 ± 32 g body weight] were fed a low-protein (LP; 6% energy value) or normal-protein (NP; 20%) diet for 8 wk. Food intake and body mass were measured during the entire intervention. During self-selection sessions after 4 wk of experimental diets, we evaluated rat food preference between LP, NP, or high-protein (HP; 55%) pellets. At the end of the experiment, we assessed their hedonic response [ultrasonic vocalizations (USVs)] and c-Fos neuronal activation in the olfactory tubercle and nucleus accumbens (NAcc) associated with an LP or HP meal. RESULTS: Rats fed an LP diet had greater food intake (24%), body weight (5%), and visceral adiposity (30%) than NP rats. All LP rats and half of the NP rats showed a nearly exclusive preference for HP pellets during self-selection sessions, whereas the other half of the NP rats showed no preference. This suggests that the appetite for proteins is driven not only by a low protein status but also by individual traits in NP rats. LP or HP meal induced similar USV emission and similar neuronal activation in the NAcc in feed-deprived LP and NP rats, showing no specific response linked to protein appetite. CONCLUSIONS: Protein appetite in rats is driven by low protein status or individual preferences in rats receiving adequate protein amounts. This must be considered and further analyzed, in the context of current recommendations for protein intake reduction.

Local Foods Can Increase Adequacy of Nutrients Other than Iron in Young Urban Egyptian Women: Results from Diet Modeling Analyses.

BACKGROUND: Nutrition transition and recent changes in lifestyle in Middle Eastern countries have resulted in the double burden of malnutrition. In Egypt, 88% of urban women are overweight or obese and 50% are iron deficient. Their energy, sugar, and sodium intakes are excessive, while intakes of iron, vitamin D, and folate are insufficient. OBJECTIVE: This study aimed to formulate dietary advice based on locally consumed and affordable foods and determine the need for fortified products to meet the nutrient requirements of urban Egyptian women. METHODS: Food intakes were assessed using a 4-d food diary collected from 130 urban Egyptian women aged 19-30 y. Food prices were collected from modern and traditional markets to calculate diet cost. Population-based linear and goal programming analyses (Optifood tool) were used to identify "limiting nutrients" and to assess whether locally consumed foods (i.e., consumed by >5% of women) could theoretically improve nutrient adequacy at an affordable cost (i.e., less than or equal to the mean diet cost), while meeting recommendations for SFAs, sugars, and sodium. The potential of hypothetical fortified foods for improving intakes of micronutrients was also assessed. RESULTS: Iron was the most limiting nutrient. Daily consumption of fruits, vegetables, milk or yogurt, meat/fish/eggs, and tahini (sesame paste) were likely to improve nutrient adequacy for 11 out of 12 micronutrients modeled. Among fortified foods tested, iron-fortified rice, milk, water, bread, or yogurt increased the minimized iron content of the modeled diet from 40% to >60% of the iron recommendation. CONCLUSIONS: A set of dietary advice based on locally consumed foods, if put into practice, can theoretically meet requirements for most nutrients, except for iron for which adequacy is harder to achieve without fortified products. The acceptability of the dietary changes modeled needs evaluation before promoting them to young Egyptian women.

Social modeling of food choices in real life conditions concerns specific food categories.

The social context of eating has a profound effect on consumption choices. Social modeling, that involves using others' behavior as a guide for appropriate consumption, has been well documented for food intake, but less is known about social modeling of food choices. Moreover, social modeling has mainly been studied in laboratory settings. We conducted an observational study in a self-service canteen to examine whether the food choices of an individual were influenced by the choice of the person ahead in the queue. We recorded food choices of 546 individuals (333 men and 211 women) and those of the person in front of them in the queue along a linear buffet. Starters were sub-categorized into salads, mixed starters (e.g. avocado shrimp mayonnaise), and cold meat starters, and desserts were sub-categorized into fruits, dairy products and pastries. There was a significantly higher probability of taking a starter in general (OR = 1.65, IC = 1.06-2.57, p = 0.03), a salad (OR = 1.78, CI = 1.08-2.93, p = 0.02), a mixed starter (OR = 2.98, CI = 1.42-6.05, p < 0.01), but not a cold meat, if the person ahead in the queue also took one compared to when the person ahead did not take one. No significant modelling was found for desserts which may be because almost all participants took a dessert. These results highlight that social modeling influences food choices, and that this phenomenon can be observed in a real life setting. These data also suggest that some food categories, such as starters, could be more susceptible to social modeling than are others. Finally, we observed modeling both between familiar and unfamiliar participants, which suggests that social norms could be used to promote healthier eating in a range of settings including friendship groups.

Self-declared attitudes and beliefs regarding protein sources are a good prediction of the degree of transition to a low-meat diet in France.

Meat consumption in Western countries is declining and, while the proportion of strict vegetarians remains low, intermediate diets such as flexitarianism have been developing in recent years. Our objectives were to identify the different levels of transition towards low-meat diets, characterize how these diets differ in terms of food intake, and identify whether attitudes and beliefs can explain these degrees of transition. In a representative survey of the French adult population conducted in 2018 (n = 2055), participants declared whether they followed a particular diet and completed a food frequency questionnaire on 29 food sources of protein and a questionnaire on their attitudes and beliefs regarding protein sources. We identified four dietary types based on these declarative data: vegetarians, flexitarians, pro-flexitarians and omnivores. The theory of planned behavior was used to predict meat intake and intentions to reduce meat intake. The sample contained 2.5% vegetarians, 6.3% flexitarians, 18.2% pro-flexitarians and 72.9% omnivores. The diet groups displayed specific dietary profiles and attitudinal scores. Compared with omnivores, pro-flexitarians consumed less red meat, more vegetables and legumes and were much more in agreement about the environmental impacts of meat. Compared with pro-flexitarians, flexitarians consumed less red meat and processed meat, and agreed much more about the health impacts of meat. Finally, versus flexitarians, vegetarians consumed almost no meat but far more legumes, nuts and seeds, and were much more sensitive to animal welfare issues. Attitudes, social norms and perceived behavioral control (PBC) predicted intentions to reduce meat consumption but attitude was the most important predictor. Intentions and PBC were both predictive of meat consumption. The dietary type related to the level of meat intake could be predicted by self-declared attitudes and beliefs regarding protein sources.

Differences in BOLD responses in brain reward network reflect the tendency to assimilate a surprising flavor stimulus to an expected stimulus.

External information can modify the subjective value of a tasted stimulus, but little is known about neural mechanisms underlying these behavioral modifications. This study used flavored drinks to produce variable degrees of discrepancy between expected and received flavor. During a learning session, 43 healthy young men learned 4 symbol-flavor associations. In a separate session, associations were presented again during an fMRI scan, but half of the trials introduced discrepancy with previously learned associations. Liking ratings of drinks were collected and were analyzed using a linear model to define the degree to which discrepant symbols affected liking ratings of the subjects during the fMRI session. Based on these results, a GLM analysis of fMRI data was conducted to determine neural correlates of observed behavior. Groups of subjects were composed based on their behavior in response to discrepant symbols, and comparison of brain activity between groups showed that activation in the PCC and the caudate nucleus was more potent in those subjects in which liking was not affected by discrepant symbols. These activations were not found in subjects who assimilated unexpected flavors to flavors preceeded by discrepant symbols. Instead, these subjects showed differences in the activity in the parietal operculum. The activity of reward network appears to be related to assimilation of received flavor to expected flavor in response to symbol-flavor discrepancy.

The Protein Status of Rats Affects the Rewarding Value of Meals Due to their Protein Content.

Background: Protein status is controlled by the brain, which modulates feeding behavior to prevent protein deficiency. Objective: This study tested in rats whether protein status modulates feeding behavior through brain reward pathways. Methods: Experiments were conducted in male Wistar rats (mean ± SD weight; 230 ± 16 g). In experiment 1, rats adapted for 2 wk to a low-protein (LP; 6% of energy) or a normal-protein (NP; 14% of energy) diet were offered a choice between 3 cups containing high-protein (HP; 50% of energy), NP, or LP feed; their intake was measured for 24 h. In 2 other experiments, the rats were adapted for 2 wk to NP and either HP or LP diets and received, after overnight feed deprivation, a calibrated HP, NP, or LP meal daily. After the meal, on the last day, rats were killed and body composition and blood protein, triglycerides, gut neuropeptides, and hormones were determined. In the brain, neuropeptide mRNAs in the hypothalamus and c-Fos protein and opioid and dopaminergic receptor mRNAs in the nucleus accumbens (NAcc) were measured. Results: Rats fed an LP compared with an NP diet had 7% lower body weight, significantly higher protein intake in a choice experiment (mean ± SD: 30.5% ± 0.05% compared with 20.5% ± 0.05% of energy), higher feed-deprived blood ghrelin, lower postmeal blood leptin, and higher neuropeptide Y (Npy) and corticotropin-releasing hormone (Crh) mRNA expression in the hypothalamus. In contrast to NP, rats fed an LP diet showed postmeal c-Fos protein expression in the NAcc, which was significantly different between meals, with LP < NP < HP. In contrast, in rats adapted to an HP diet compared with an NP diet, energy intake was lower; and in the NAcc, meal-induced c-Fos protein expression was 20% lower, and mRNA expression was 17% higher for dopamine receptor 2 (Drd2) receptors and 38% lower for κ opioid receptor (Oprk1) receptors. Conclusion: A protein-restricted diet induced a reward system-driven appetite for protein, whereas a protein-rich diet reduced the meal-induced activation of reward pathways and lowered energy intake in male rats.

Expected satiation alone does not predict actual intake of desserts.

The degree to which consumers expect foods to satisfy hunger, referred to as expected satiation, has been reported to predict food intake. Yet this relationship has not been established precisely, at a quantitative level. We sought to explore this relationship in detail by determining whether expected satiation predicts the actual intake of semi-solid desserts. Two separate experiments were performed: the first used variations of a given food (eight apple purées), while the second involved a panel of different foods within a given category (eight desserts). Both experiments studied the consumption of two products assigned to volunteers based on their individual liking and expected satiation ratings, given ad libitum at the end of a standardised meal. A linear model was used to find predictors of food intake and included expected satiation scores, palatability scores, BMI, age, sex, TFEQ-R, TFEQ-D, water consumption during the meal, reported frequency of eating desserts, and reported frequency of consuming tested products as explanatory variables. Expected satiation was a significant predictor of actual food intake in both experiments (apple purée: F(1,97) = 18.60, P < .001; desserts: F(1,106) = 9.05, P < .01), along with other parameters such as product palatability and the volunteers' age, sex and food restriction (variation explained by the model/expected satiation in the experiments: 57%/23% and 36%/17%, respectively). However, we found a significant gap between expected and actual consumption of desserts, on group and on individual level. Our results confirm the importance of expected satiation as a predictor of subsequent food intake, but highlight the need to study individual consumption behaviour and preferences in order to fully understand the role of expected satiation.

Behavioural measures of child's eating temperament and their link with BMI.

Rothbart's model of temperament, defined as individual differences in reactivity and self-regulation, has a strong heuristic value with applications in a wide variety of children's outcomes. Our objective was to test Rothbart's model applied to children's food behaviours and BMI outcome through behavioural measures. Our hypotheses, according to Rothbart's model, were as follows: (i) self-regulation in eating modulates appetite reactivity; (ii) appetite reactivity increases the risk of excess BMI, whereas self-regulation in eating limits this risk. One hundred and four children aged between 7 and 12 years completed four behavioural tasks to assess scores for two components of appetite reactivity (i.e. appetite arousal and appetite persistence) and two components of self-regulation in eating (i.e. self-regulation in eating without hunger and self-regulation in eating speed). Their heights and weights were measured in order to calculate their BMI-for-age. T-tests and regression analysis were used to verify our hypotheses. None of the scores of self-regulation in eating was directly associated with BMI but we observed a significant impact of self-regulation in eating without hunger on appetite arousal (p-value = 0.04), together with a modest but significant association between appetite persistence and BMI (p-value = 0.02). We can thus conclude that our behavioural measures could be used for the determination of the child's eating temperament. Further studies are needed to investigate how to use these measures to improve the treatment of overweight in children.

In-store marketing of inexpensive foods with good nutritional quality in disadvantaged neighborhoods: increased awareness, understanding, and purchasing.

BACKGROUND: Consumers often do not understand nutrition labels or do not perceive their usefulness. In addition, price can be a barrier to healthy food choices, especially for socio-economically disadvantaged individuals. METHOD: A 6-month intervention combined shelf labeling and marketing strategies (signage, prime placement, taste testing) to draw attention to inexpensive foods with good nutritional quality in two stores located in a disadvantaged neighborhood in Marseille (France). The inexpensive foods with good nutritional quality were identified based on their nutrient profile and their price. Their contribution to customers' spending on food was assessed in the two intervention stores and in two control stores during the intervention, as well as in the year preceding the intervention (n = 6625). Exit survey (n = 259) and in-depth survey (n = 116) were used to assess customers' awareness of and perceived usefulness of the program, knowledge of nutrition, understanding of the labeling system, as well as placement-, taste- and preparation-related attractiveness of promoted products. Matched purchasing data were used to assess the contribution of promoted products to total food spending for each customer who participated in the in-depth survey. RESULTS: The contribution of inexpensive foods with good nutritional quality to customers' total food spending increased between 2013 and 2014 for both the control stores and the intervention stores. This increase was significantly higher in the intervention stores than in the control stores for fruits and vegetables (p = 0.001) and for starches (p = 0.011). The exit survey revealed that 31 % of customers had seen the intervention materials; this percentage increased significantly at the end of the intervention (p < 0.001). The in-depth survey showed that customers who had seen the intervention materials scored significantly higher on quizzes assessing nutrition knowledge (p < 0.001) and understanding of the labeling system (p = 0.024). CONCLUSION: A social marketing intervention aimed at increasing the visibility and attractiveness of inexpensive foods with good nutritional quality may improve food purchasing behaviors in disadvantaged neighborhoods.

Assimilation and contrast are on the same scale of food anticipated-experienced pleasure divergence.

Consumption of a product is preceded by an anticipation of its qualities by the consumer, which can itself modify the consumption experience. Improved knowledge of anticipation would allow better manipulation of it, for example to enhance the acceptance of healthier foods. According to the Assimilation-Contrast theory, the size of anticipation-reality divergence determines how anticipation influences consumers' satisfaction. For small divergences, experienced pleasure is the same as the anticipated pleasure (Assimilation); for large ones, the effect of surprise provokes an even larger discordance with that which was anticipated (Contrast). Few studies have attempted to observe both effects simultaneously, or to consider the anticipation-reality divergence quantitatively rather than qualitatively; these were the study's objectives. A range of 10 flavored drinks was developed to vary progressively in intensity. Ninety healthy young men consumed samples during two separate sessions. In session 1, hedonic and sensory scores of all drinks were recorded during blind tasting. In session 2, three drinks were chosen as references for taste intensity, and associated with neutral symbols that served as labels. Subjects then consumed 36 drink samples, each one bearing a label. For half of the samples drinks did not correspond to labels, creating a range of anticipation-reality divergence. By predicting session 2 scores using linear modeling with session 1 blind ratings as input, it was confirmed that both Assimilation and Contrast effects on hedonic ratings were present (Assimilation (t(89) = 5.645, p < 0.0001) and Contrast (t(89) = 3.186, p = 0.002 or t(89) = 2.494, p = 0.015, depending on the drink-label combination)). This study was the first to position Assimilation and Contrast within a quantitative context using controlled divergence variation rather than products from distinct categories.

Environmental enrichment alters splenic immune cell composition and enhances secondary influenza vaccine responses in mice.

Chronic stress has deleterious effects on immune function, which can lead to adverse health outcomes. However, studies investigating the impact of stress reduction interventions on immunity in clinical research have yielded divergent results, potentially stemming from differences in study design and genetic heterogeneity, among other clinical research challenges. To test the hypothesis that reducing glucocorticoid levels enhances certain immune functions, we administered influenza vaccine once (prime) or twice (boost) to mice housed in either standard control caging or environmental enrichment (EE) caging. We have shown that this approach reduces mouse corticosterone production. Compared with controls, EE mice had significantly lower levels of fecal corticosterone metabolites (FCMs) and increased splenic B and T lymphocyte numbers. Corticosterone levels were negatively associated with the numbers of CD19(+) (r(2) = 0.43, p = 0.0017), CD4(+) (r(2) = 0.28, p = 0.0154) and CD8(+) cells (r(2) = 0.20, p = 0.0503). Vaccinated mice showed nonsignificant differences in immunoglobulin G (IgG) titer between caging groups, although EE mice tended to exhibit larger increases in titer from prime to boost than controls; the interaction between the caging group (control versus EE) and vaccine group (prime versus boost) showed a strong statistical trend (cage-group*vaccine-group, F = 4.27, p = 0.0555), suggesting that there may be distinct effects of EE caging on primary versus secondary IgG vaccine responses. Vaccine-stimulated splenocytes from boosted EE mice had a significantly greater frequency of interleukin 5 (IL-5)-secreting cells than boosted controls (mean difference 7.7, IL-5 spot-forming units/10(6) splenocytes, 95% confidence interval 0.24-135.1, p = 0.0493) and showed a greater increase in the frequency of IL-5-secreting cells from prime to boost. Our results suggest that corticosterone reduction via EE caging was associated with enhanced secondary vaccine responses, but had little effect on primary responses in mice. These findings help identify differences in primary and secondary vaccine responses in relationship to stress mediators that may be relevant in clinical studies.

The effect of social norms on vegetarian choices is moderated by intentions to follow a vegetarian diet in the future: Evidence from a laboratory and field study.

Social norms could be a tool in dietary transition toward more sustainable diets, but the results of social norms interventions aimed at encouraging the selection of plant-based foods to date have been inconsistent. One reason for this might be because there are important moderating factors that have yet to be investigated. Here we examine social modeling of vegetarian food choices and test whether modeling is dependent upon individual intentions to follow a vegetarian diet in the future in two different settings. In a laboratory study of 37 women, participants with low intentions to become a vegetarian consumed fewer plant-based foods in the presence of a vegetarian confederate, compared to eating alone. In an observational study of 1,037 patrons of a workplace restaurant, participants with a higher score of on vegetarian intentions had a greater likelihood of taking a vegetarian main course or starter, and a vegetarian social norm was associated with a greater likelihood of a vegetarian choice for the main course but not for the starter. These data suggest that participants with low intentions to follow a vegetarian diet may exhibit reactance against an explicit vegetarian norm in an unfamiliar context (as in Study 1) but that general norm following regardless of dietary intentions be more likely when it is conveyed implicitly in a familiar context (as in Study 2).

Rats chirp with their mouth full: During an experimental meal, adult male Wistar rats emitted flat ultrasonic vocalisations upon feeding.

Rats produce ultrasonic vocalisation (USVs) that are classified into different types, based on their average frequency. In pups 40 kHz USVs are produced upon social isolation, and in adults USVs can be associated with affective states and specific behavioural patterns (i.e., appetitive 50 kHz vocalisations of frequency range 30-100 kHz, or aversive 20 kHz vocalisations of frequency range 18-30 kHz). Generally, USVs of frequency around 50 kHz are linked to activation of brain reward pathways, during anticipation or experience of rewarding stimuli. Previous studies have described several subtypes of 50 kHz USVs, according to their acoustic properties. We asked whether USV production might be relevant to feeding behaviour. We recorded USVs from 14-week old adult rats during the satisfaction of a physiological need: refeeding following mild food deprivation (17 h overnight fast). We analysed a 10 min consummatory phase, preceded by a 10 min anticipatory phase, as a control for the experimental meal. Following identification of USV subtypes, we applied frequentist and Bayesian (Monte Carlo shuffling) statistical analyses to investigate the relationship between USV emission and rat behaviour. We found that it was not total USV quantity that varied in response to food consumption, but the subtype of USV produced. Most importantly we found that rats who feed tend to produce flat USVs of a frequency around 40 kHz. Beyond the previous reports of circumstantial association feeding-flat USVs, our observation directly correlate vocalisation and ingestive behaviour. Our study highlights that, in addition to quantification of the production rate, study of USV subtypes might inform us further on rat consummatory behaviour. Since this vocalisation behaviour can have a communicative purpose, those findings also illustrate nutrition studies might benefit from considering the possible social dimension of feeding behaviour.

The calm mouse: an animal model of stress reduction.

Chronic stress is associated with negative health outcomes and is linked with neuroendocrine changes, deleterious effects on innate and adaptive immunity, and central nervous system neuropathology. Although stress management is commonly advocated clinically, there is insufficient mechanistic understanding of how decreasing stress affects disease pathogenesis. Therefore, we have developed a "calm mouse model" with caging enhancements designed to reduce murine stress. Male BALB/c mice were divided into four groups: control (Cntl), standard caging; calm (Calm), large caging to reduce animal density, a cardboard nest box for shelter, paper nesting material to promote innate nesting behavior, and a polycarbonate tube to mimic tunneling; control exercise (Cntl Ex), standard caging with a running wheel, known to reduce stress; and calm exercise (Calm Ex), calm caging with a running wheel. Calm, Cntl Ex and Calm Ex animals exhibited significantly less corticosterone production than Cntl animals. We also observed changes in spleen mass, and in vitro splenocyte studies demonstrated that Calm Ex animals had innate and adaptive immune responses that were more sensitive to acute handling stress than those in Cntl. Calm animals gained greater body mass than Cntl, although they had similar food intake, and we also observed changes in body composition, using magnetic resonance imaging. Together, our results suggest that the Calm mouse model represents a promising approach to studying the biological effects of stress reduction in the context of health and in conjunction with existing disease models.

Dietary fibers solubilized in water or an oil emulsion induce satiation through CCK-mediated vagal signaling in mice.

This study focused on the fate of the satiating potency of dietary fibers when solubilized in a fat-containing medium. Fourteen percent of either guar gum (GG) or fructo-oligosaccharide (FOS) or a mixture of the 2 (GG-FOS, 5% GG and 9% FOS) were solubilized in water or an oil emulsion (18-21% rapeseed oil in water, v:v) and administered by gavage to mice before their food intake was monitored. When compared with water (control), only GG-FOS solubilized in water or in the oil emulsion reduced daily energy intake by 21.1 and 14.1%, respectively. To further describe this effect, the meal pattern was characterized and showed that GG-FOS increased satiation without affecting satiety by diminishing the size and duration of meals for up to 9 h after administration independently of the solubilization medium. The peripheral blockade of gut peptide receptors showed that these effects were dependent on the peripheral signaling of cholecystokinin but not of glucagon-like peptide 1, suggesting that anorectic signals emerge from the upper intestine rather than from distal segments. Measurements of neuronal activation in the nucleus of solitary tract supported the hypothesis of vagal satiation signaling because a 3-fold increase in c-Fos protein expression was observed in that nucleus after the administration of GG-FOS, independently of the solubilization medium. Taken together, these data suggest that a mixture of GG and FOS can maintain its appetite suppressant effect in fatty media. Adding these dietary fibers to fat-containing foods might therefore be useful in managing food intake.

Peripheral and central mechanisms involved in the control of food intake by dietary amino acids and proteins.

The present review summarises current knowledge and recent findings on the modulation of appetite by dietary protein, via both peripheral and central mechanisms. Of the three macronutrients, proteins are recognised as the strongest inhibitor of food intake. The well-recognised poor palatability of proteins is not the principal mechanism explaining the decrease in high-protein (HP) diet intake. Consumption of a HP diet does not induce conditioned food aversion, but rather experience-enhanced satiety. Amino acid consumption is detected by multiple and redundant mechanisms originating from visceral (during digestion) and metabolic (inter-prandial period) sources, recorded both directly and indirectly (mainly vagus-mediated) by the central nervous system (CNS). Peripherally, the satiating effect of dietary proteins appears to be mediated by anorexigenic gut peptides, principally cholecystokinin, glucagon-like peptide-1 and peptide YY. In the CNS, HP diets trigger the activation of noradrenergic and adrenergic neurons in the nucleus of the solitary tract and melanocortin neurons in the arcuate nucleus. Additionally, there is evidence that circulating leucine levels may modulate food intake. Leucine is associated with neural mechanisms involving mammalian target of rapamycin (mTOR) and AMP-activated protein kinase (AMPK), energy sensors active in the control of energy intake, at least in the arcuate nucleus of the hypothalamus. In addition, HP diets inhibit the activation of opioid and GABAergic neurons in the nucleus accumbens, and thus inhibit food intake by reducing the hedonic response to food, presumably because of their low palatability. Future studies should concentrate on studying the adaptation of different neural circuits following the ingestion of protein diets.

Dopamine Modulates the Processing of Food Odour in the Ventral Striatum.

Food odour is a potent stimulus of food intake. Odour coding in the brain occurs in synergy or competition with other sensory information and internal signals. For eliciting feeding behaviour, food odour coding has to gain signification through enrichment with additional labelling in the brain. Since the ventral striatum, at the crossroads of olfactory and reward pathways, receives a rich dopaminergic innervation, we hypothesized that dopamine plays a role in food odour information processing in the ventral striatum. Using single neurones recordings in anesthetised rats, we show that some ventral striatum neurones respond to food odour. This neuronal network displays a variety of responses (excitation, inhibition, rhythmic activity in phase with respiration). The localization of recorded neurones in a 3-dimensional brain model suggests the spatial segregation of this food-odour responsive population. Using local field potentials recordings, we found that the neural population response to food odour was characterized by an increase of power in the beta-band frequency. This response was modulated by dopamine, as evidenced by its depression following administration of the dopaminergic D1 and D2 antagonists SCH23390 and raclopride. Our results suggest that dopamine improves food odour processing in the ventral striatum.

Associations between Perceived Social Eating Norms and Initiation and Maintenance of Changes in Dietary Habits during the First COVID-19 Lockdown in France.

Changes in dietary habits of the French population have been reported during the national lockdown that was enforced due to the COVID-19 pandemic. This study investigated whether perceived social eating norms were associated with the initiation and maintenance of dietary changes that took place as a result of lockdown. An online study collected information on (1) changes in consumption implemented during the lockdown and the maintenance of these changes, and (2) perceptions about changes in consumption implemented during lockdown by household members, relatives out of home, and the general population. The changes in consumption were classified as foods to increase or to decrease according to French national recommendations. The perception of changes to dietary habits by household members and relatives out of home was related to the changes made by individuals for each of the food categories (all p < 0.05) but not to the perception of changes made the general population. Increased consumption of foods to increase was more likely to be maintained when there was a positive perception of the changes made by household members (p = 0.03). These results highlight the influence of the perception of social eating norms, especially by household members and relatives, on the implementation of dietary changes during lockdown and suggest that social eating norms can have a lasting influence.