No description available.

Wellness is a domain of growing interest in computing. Many interventions are designed to modify people’s behavior to make them healthy. But often, they ignore people’s beliefs and socioeconomic context so no sustained change are achieved. This thesis, therefore, focuses on rethinking how we can leverage people’s food beliefs to design healthy eating systems that can achieve sustained change. The thesis begins with a literature review on why healthy eating is complex and how it depends on various contexts including people’s food beliefs. Then we reflect upon current design notion of healthy eating applications and acknowledge the current designs can not address all the complex of healthy eating. Thus, we should explore the alternative design notion, which tries to leverage what people think about food. A study on diet-related topics and social interaction patterns on forums led us to the idea that crowd’s wisdom on food’s healthiness could help people to better understand food and potentially encourage them to change eating behavior. Thus, we designed and evaluated one early prototype that allows people to judge each other’s food’s healthiness by giving ratings and checking 5 predefined heuristics. The outcome of this prototype study confirmed that people like the idea of expressing own thoughts about food and like to have others’ thoughts to validate owns. We further explored exactly what kinds of food beliefs people hold via a survey study with real food pictures. And we found 8 categories of food beliefs and also validated that even experts can not reach consensus on food’s healthiness and thus there is clearly a gap between what we know from nutrition literature and how interventions are actually designed. To further leverage people’s food beliefs, we need to first capture it. We developed three simple UIs to capture such food beliefs, and evaluated the UIs from three different perspectives. We also found two patterns from the data we collect which can help us better visualize the data and identify group of people. In the end of the thesis, we discuss the thesis work and present the future works.

Dans les pays occidentaux, la consommation de protéines animales, majoritaire, diminue depuis une décennie. Elle est, dans la majorité des cas, associée négativement à différents paramètres de durabilité, et les études modélisant des régimes durables ont montré que les différents paramètres n’étaient pas toujours compatibles. Des régimes améliorant largement différents paramètres de durabilité ont été modélisés, mais la prise en compte de l’acceptabilité culturelle est insuffisante et nécessite des approfondissements. L’objectif de cette thèse était de modéliser des trajectoires acceptables de réarrangement de la consommation de sources protéiques pour augmenter l’adéquation nutritionnelle et d’en évaluer les impacts sur la durabilité. L’étude des consommations de sources de protéines en France a permis de conclure que les apports en protéines sont adéquats pour l’ensemble de la population, mais qu’il existe différents profils de consommation protéique, caractérisés par des niveaux de sécurité nutritionnelle différents. Une étude réalisée en 2018 a permis d’établir que les niveaux de consommation de viande étaient prédits par les attitudes, les normes sociales, et l’auto-efficacité perçue vis-à-vis de la réduction de la consommation de viande. Pour les travaux de modélisation pas-à-pas de l’alimentation, il a été considéré acceptable pour un individu de consommer un nouvel aliment, si celui-ci était largement consommé par des individus au profil de consommation protéique similaire. Cette hypothèse a été validée par une enquête en 2018. Les travaux de modélisation ont permis d’identifier que certaines recommandations alimentaires étaient efficaces pour l’ensemble de la population, mais que d’autres étaient spécifiques à certains profils de consommation protéique, caractérisés par des profils nutritionnels et des répertoires alimentaires spécifiques. Enfin, des modèles ont permis d’identifier que viser systématiquement plus de protéines végétales lors des premières modifications diététiques permet, malgré une adéquation nutritionnelle légèrement plus faible, d’obtenir de meilleurs paramètres de durabilité. / In Western countries, the consumption of animal protein, which is the predominant protein source, has been decreasing over the last decade. This consumption has been negatively associated with different sustainability parameters in the majority of cases, and studies modelling sustainable diets have shown that the different parameters are not always compatible. Diets that significantly improve different sustainability parameters have been modelled, but the consideration of cultural acceptability is insufficient and requires further investigation. The objective of this thesis was to model acceptable trajectories of rearrangement of protein source consumption to increase nutritional adequacy and to evaluate its impacts on sustainability. The study of protein source consumption in France led to the conclusion that protein intake is adequate for the entire population, but that there are different profiles of protein intake characterized by different levels of nutritional security. A study conducted in 2018 found that meat consumption levels were predicted by attitudes, social norms, and perceived behavioral control related to the reduction of meat consumption. For the stepwise dietary modelling study, it was considered acceptable for an individual to consume a new food if it was widely consumed by individuals with a similar profile of protein intake. This hypothesis was validated by a survey conducted in 2018. The modelling work identified that some dietary recommendations were effective for the general population, but others were specific to certain profiles of protein intake characterized by specific nutritional profiles and food repertoires. Finally, models have identified that systematically targeting more plant proteins during the first dietary modifications allows, despite a slightly lower nutritional adequacy, to obtain better sustainability parameters.

L’exposition au régime maternel durant la période périnatale, induit des processus d’empreintes orientant à long terme le phénotype et la santé des individus. De plus, les orientations alimentaires, telles que celles concernant les protéines, évoluent quantitativement et qualitativement. Ces deux constats, encouragent la communauté scientifique à s’interroger sur les conséquences de ces variations de consommation en protéines sur les générations futures. Ce projet de thèse vise à évaluer chez la descendance femelle rat, les effets d’une alimentation maternelle variant par la teneur (riche versus normal) et la qualité (sources animales versus végétales) en protéines sur la modification des préférences alimentaires et sur les risques métaboliques.Deux études ont été réalisées chez le rat. Une première étude a évalué l’impact de l’excès de protéines à travers un régime hyperprotéique (HP) à base de protéines de lait pendant la gestation. Une seconde étude a évalué les effets d’un régime HP de source protéique spécifique (lait, pois ou dinde) pendant la gestation et d’un régime de source protéique spécifique (lait, pois ou dinde) pendant l’allaitement. Une fois sevrés et jusqu’à l’âge adulte (étude 1 : 15 semaines ; étude 2 : 10 semaines), les ratons femelles ont été soumis à des modèles de « dietary self-selection » (DSS) leur laissant la possibilité de choisir la composition en macronutriments, le niveau de consommation alimentaire et la source protéique (étude 2 uniquement). Indépendamment du régime maternel, ces deux études ont montré que lorsque les sources en macronutriments étaient séparées dans le modèle DSS, les ratons présentaient une hyperphagie liée à une consommation accrue de lipides au détriment des glucides.De plus, les résultats de la seconde étude ont montré que les ratons n’orientaient pas spécifiquement leur consommation de protéines vers la source protéique à laquelle ils avaient été exposés via le régime maternel périnatal. En revanche, les deux études ont montré que la consommation d’un régime HP pendant la gestation, quelle que soit la qualité des protéines le composant, induisait une augmentation de l’adiposité chez la descendance femelle adulte. Cette augmentation était majorée lorsque la descendance avait été soumise au régime de choix (DSS), leur permettant d’augmenter leur consommation de lipides au détriment des glucides.En conclusion, l’exposition périnatale à un régime HP de qualité variable en protéines augmente la sensibilité au surpoids chez la descendance femelle adulte rat. Nous avons évalué les relations entre ces données et : la sensibilité des voies centrales du contrôle de la prise alimentaire et de la récompense, la sensibilité des voies de contrôle du métabolisme énergétique périphérique et la composition et l’activité du microbiote de l’intestin.Ces travaux apportent un grand nombre de nouvelles données indiquant clairement qu’une alimentation équilibrée en quantité et en qualité de protéines pendant la grossesse, à travers le ratio protéines/glucides et le profil en acides aminés, pourrait jouer un rôle clé sur des paramètres phénotypiques de la descendance notamment lorsqu’elle est soumise à des choix alimentaires augmentés. / Abstract : Perinatal exposure to maternal diet induces programming processes of later individual phenotype and health. Additionally, food orientations like for protein, change in terms of quantity and quality. These observations enhance scientific community to evaluate consequences of protein consumption changes on future generations.This thesis project aims to determine the consequences of modifying protein quantity and quality in maternal diets on food preferences and metabolic risks in female rat offspring.Two studies were conducted in rats. The first study evaluated the impact of protein excess in the maternal diet during gestation, through a high-protein (HP) diet composed with cow milk protein. The second study evaluated effects of (i) a HP diet composed with different protein sources (cow milk, pea, or turkey) during gestation and (ii) these different protein sources (cow milk, pea, or turkey-derived) during lactation. From weaning to adulthood (study 1: 15 weeks after birth; study 2; 10 weeks after birth), female pups were subjected to “dietary self-selection” (DSS), which allowed them to choose their own macronutrient compositions, level of food intake and protein sources (second study only).Regardless of the maternal diet, these two studies showed that when DSS was composed with separate macronutrients, rats exhibited overfeeding and increased lipid intake coupled with a decreased carbohydrate intake. Moreover, the results of the second study indicated that rats did not orient their protein intake towards the maternal protein source to which they were exposed during perinatal period. Nevertheless, the two studies showed that the maternal HP diet during gestation caused an increased adiposity in female adult offspring, regardless of the maternal protein source. This increase was stronger when offspring were subjected to DSS condition, which allowed them to increase lipid intake and decrease carbohydrate intake.In conclusion, perinatal exposure to a HP diet varying in protein quantity and quality increases the risk of becoming overweight in female rat adult offspring. We assess the relationship between these data and the the sensitivity of central pathways of food intake and reward control, the sensitivity of energetic and peripheral metabolic pathways, and the gut microbiota composition and activity.This work provides new data indicating that a balanced diet in protein quantity and quality during gestation, through a protein/carbohydrate ratio and amino acid profile, could play a key role on offspring phenotypic parameters, especially when submitted to increased dietary options.

Background Middle-income countries are currently undergoing nutrition transition more rapidly than did high income countries. These populations are therefore at high risk of over-nutrition (obesity) and nutrition-related, non-communicable diseases before the elimination of under-nutrition and infectious diseases. Such nutritional double burden is most common in low SES groups of middle-income countries, such as the Maya of Mexico. Long-term poor environmental conditions during early life results in a population with high levels of chronic under-nutrition (stunting), and a consequent predisposition toward overweight/ obesity, and associated health risks later in life. It is important to be able to identify individuals at an increased risk of diseases related to double burden and to determine whether stunting impacts the ability to identify at risk individuals. Aims The overall purpose of this doctoral research was to examine double burden in the urban Maya, a low SES section of the Mexican population which is a current example of a population undergoing significant transition. The specific aims of this thesis were fourfold: 1) To describe the living conditions and population characteristics of the urban Maya of southern Merida in the spring and summer of 2010; 2) to determine whether body mass index (BMI) predicts adiposity indicators in a sample of women and children with a high prevalence of stunting, 3) to determine whether measures of linear growth in women can be predicted by their recalled childhood environment and 4) To determine whether the relationship between objectively estimated free-living energy expenditure and body composition is altered by stunting. Methods Data from interviews, anthropometric measurements and bioelectrical impedance analysis were collected on 58 Maya schoolchildren aged 7-9 years-old and their mothers living in the south of Merida, Mexico. Objective, free-living physical activity monitoring using combined accelerometry and heart rate monitoring of the children was also performed for one week. The interview data was used to describe the living conditions of the south of Merida. Whether stunting status or body proportions influenced the power of BMI to predict adiposity indicators was assessed for the mothers and the children. The mothers recalled early life SES was compared to their current measures of linear growth. The children s objectively estimated energy expenditure was compared to their body composition and measurements of body size. Results The Maya of southern Merida overall had access to clean drinking water and basic health care and have apparently eliminated acute under-nutrition. Yet they remained double burdened with simultaneous stunting and overweight/ obesity. Individual double burden levels were high, with 70% of the mothers were simultaneously stunted and overweight. Family level double burden was also high, with 28% of the families having an overweight mother and a stunted child. The rates of childhood malnutrition varied widely when using different cut-offs. Child stunting rates were between 15.5% and 37.9% when using -2 z-scores of Frisancho s Comprehensive (created using NHANES data) reference versus the 5th percentile of the WHO reference, respectively. Child overweight/ obesity rates were less than 10% when using weight-for-age on both the Comprehensive and WHO reference charts. Child overweight/ obesity as classified using BMI z-scores was between 27.5 to 34.5% using the Comprehensive and WHO reference, respectively, while child overfat was over 80% when using body fat percentage for age reference curves. BMI predicted adiposity indicators in these Maya children, explaining between 33 and 84% of the variance in arm fat index and waist circumference z-score, respectively. BMI was less strongly related to the mothers adiposity indicators, explaining between 19 and 70% of the variance in arm fat area and waist circumference. The relationship between BMI and adiposity indicators was unchanged by stunting or body proportions in either mothers or children. Mothers recalled early life SES was significantly related to but explained little of the variance in her measures of linear growth. Birth decade explained 5% of the variance in stature and the Modernisation index (urban/ rural birth, sugar sweetened beverage consumption, packaged food consumption) explained 5% of the variance in mothers leg length. Birth order, sibling number and consumer durable ownership were also significantly related to linear growth of the mothers. These Maya children had high levels of physical activity, as all exceeded the recommended 60 minutes of moderate-to-vigorous physical activity per day. After controlling for fat free mass, short stature did not predict lower resting energy expenditures in the children. However shorter stature did predict lower levels of activity energy expenditure, particularly in girls. Stunted girls had the lowest activity energy expenditures. Conclusions These urban Maya tend to have access to basic sanitation and services and are at a very high risk for NR-NCDs with the co-existence of chronic under-nutrition (stunting) and overweight/ obesity. The high rates of stunting do not impact the usefulness of BMI to estimate adiposity nor does stunting appear to impact children s energy expenditure. While BMI is useful to predict adiposity in these urban Maya children, it is not recommended for use in the mothers. Interventions to reduce childhood adiposity need to begin very early in life to most effectively reduce adiposity. Research into the low SES groups of middle-income countries, offers insight to what may occur in low-income countries as they advance in the nutrition transition.

Current treatments for obesity have been unsuccessful. It is essential that a patient-centred approach for obesity management is developed and for this to be successful other diet and lifestyle approaches need to be considered. A systematic review comparing low carbohydrate vs. low fat diets for the treatment of obesity was carried out. It found that low carbohydrate/high protein diets are as effective as, if not better, for treating obesity and cardiovascular disease risk factors. A randomised controlled trial investigating dietary approaches for the treatment of obesity and its co morbidities was carried out. Variables including weight and body composition, cardiovascular risk factors, adipokines, liver and kidney function, and health and lifestyle factors were measured. Those with metabolic syndrome were also examined. It was hypothesised that there are alternative ways of treating obese subjects depending on their phenotype. Those with a higher BMI tend to have a higher carbohydrate intake rather that a higher fat intake so may be better suited to a low carbohydrate/high protein diet rather than the conventional low fat/energy reduced diet. Subjects were initially treated with the standard dietary approach for obesity (health eating, HE) and if not successful after 3 months were randomised to either a very low calorie diet (Lighterlife, LL) or a protein sparing modified fast (PSMF). All three groups showed a significant weight loss and reduced risk for CVD at 12 months. Significant improvements were seen for plasminogen-activated receptor-1, adiponectin, leptin and IL-6 on HE and LL, but only adiponectin significantly improved on the PSMF. Neither diet showed any detrimental effects for those with a healthy liver and kidney function. Quality of life and levels of depression improved at 12 months. Of the 54 subjects with metabolic syndrome at baseline, 12 remained on HE and 32 were randomised to LL and PSMF. This indicates that most subjects did not suit a low fat dietary approach. They were successful at losing weight on LL and PSMF and showed improvement in MS risk factors, and adipokine levels at 12 months. In conclusion, the study demonstrates that a low fat diet may not necessarily be the first line of approach to treat obese subjects with a BMI over 35 kg/m2, including those with MS. A very low calorie diet such as LL or a PSMF may be better suited to the subject.

L’allergie alimentaire est un problème majeur de santé publique. Sa prévalence est estimée à 5 % chez l’adulte et à 8 % chez l’enfant dans les pays industrialisés. L’objectif de cette thèse est d’apporter une contribution à la diététique diagnostique et thérapeutique de l’allergie alimentaire. Nos travaux ont porté sur l’allergie alimentaire au pois jaune et au sésame. Le pois jaune est un allergène émergent à risque de réaction grave. Son seuil réactogène est bas, soit 0,13 g de protéines alors que la teneur des aliments peut atteindre 20 g/100g. L’étude de 35 personnes allergiques au sésame a permis de définir que le seuil réactogène pouvant déclencher une réaction chez 5% de la population allergique se situe entre 1,2 et 4,0 mg de protéines de sésame. Nous avons validé la progression optimale des doses de sésame pour le test de provocation orale en lien avec le consensus européen Europrevall. L’absence d’harmonisation internationale concernant les allergènes à déclaration obligatoire, nous a conduit dans une seconde partie, à étudier l’impact de l’allergie alimentaire sur la décision, l’organisation et le déroulement d’un séjour à l’étranger. Cette étude réalisée en collaboration avec l’Observatoire Régional de Santé et des Affaires Sociales de Lorraine et les associations française et belge pour la prévention des allergies montre, sur un panel de 102 répondants, que l’allergie alimentaire contraint 14 % des répondants à ne pas voyager à l’étranger et que 30% des personnes ayant voyagé ont présenté une réaction allergique. Les accidents sont survenus principalement dans les restaurants. Ces résultats soulignent la nécessité de développer un système d’information et de formation des personnes allergiques et des professionnels de l’alimentation pour mettre en oeuvre des stratégies de prévention du risque. / Food allergy is a major public health problem. Its prevalence is estimated at 5% in adults and 8% in children in industrialized countries. The objective of this thesis is to bring a contribution to the diagnostic and therapeutic dietetics in food allergy. Our work focused on yellow pea and sesame food allergies. Yellow pea is an emerging allergen with risk of a severe reaction. Its reactogenic threshold is of 0.13 g of protein while the content of food can reach 20 g / 100g. The study of 35 people allergic to sesame has enabled the identification of the reactogenic threshold that can trigger a reaction in 5% of the allergic population. It is estimated between 1.2 and 4.0 mg of sesame protein. We validated optimal progression dosing scheme for sesame for oral challenge test taking in account the European consensus Europrevall. Lack of international harmonization on mandatory allergens, led us in a second part of this thesis, to study the impact of food allergy on the decision, trip organization and a stay abroad of an allergic person. This study, conducted in collaboration with the regional health observatory and Social Affairs of Lorraine and the French and Belgian associations for allergy prevention shows, on a panel of 102 respondents, that food allergy forced 14% of respondents do not travel abroad and that 30% of people who have traveled had an allergic reaction. The accidents occurred mainly in restaurants. These results highlight the need to develop an information system and training of allergies and food professionals to implement risk prevention strategies.

It is well established that improved nutritional strategies can enhance both health and exercise performance. Scientific developments in recent years have furthered our understanding of cellular metabolism, which in turn, has provided an additional platform to investigate the impact of diet on health and adaptation. The overall aim of this research programme was to build on the current understanding of dietary intake in athletes and the impact dietary manipulation has on cellular and metabolic adaptation at rest and in combination with endurance training. It is postulated that nutrition is the most controllable risk factor impacting long-term health and chronic disease (World-Health-Organization, 2003), and enhanced knowledge of nutrition has been associated with improved dietary choices. A number of nutrition knowledge questionnaires have been developed to assess this; however the validity of each tool is reduced if implemented outside the target population. A valid and reliable general and sport nutrition knowledge questionnaire had not yet been developed. Using a parallel groups repeated measures study design (N = 101) the aim of the first experimental Chapter (Chapter 4) was to develop a new tool to measure general and sport nutrition knowledge in UK track and field athletes. Following the questionnaire design 53 nutrition educated and 48 non-nutrition educated participants completed the questionnaire on two occasions separated by three weeks. The results of the process demonstrated face and construct validity from the development of the question pool, content validity (the nutrition educated group scored > 30% higher that the non-nutrition educated group), reliability (test - retest correlation of 0.98, p < 0.05) and internal consistency (Chronbach's alpha value > 0.7) as such establishing a new tool (Nutrition knowledge Questionnaire for Athletes (NKQA)) for the assessment of general and sport nutrition knowledge in track and field athletes. Athletes' diets are commonly reported as inadequate and previous work has demonstrated a weak positive relationship between diet quality and nutrition knowledge. Additionally a commercially available tool, the metabolic typing questionnaire, claims to identify individual metabolic function and subsequently prescribe a personalised diet to optimise health. Thus the aim of the second experimental Chapter (Chapter 5) was to quantify nutrition knowledge (using the questionnaire developed in Chapter 4), measure diet intake and quality and investigate the efficacy of the metabolic typing questionnaire in UK track and field athletes. Using a parallel groups repeated measures design participants (UK track and field athletes n = 59, and non-athletic control group n = 29) completed a food diary, the NKQA and the metabolic typing questionnaire at two time points through the year (October and April) to investigate seasonal change. The results of the metabolic typing questionnaire concluded that 94.3% of the participants were the same dietary type and would subsequently have been prescribed the same diet. Athletes possess greater general and sport nutrition knowledge the non-athletes (60.4 ± 2.0 % vs. 48.6 ± 1.5 %) and also had better diet quality (76.8 ± 10.5 % vs. 67.6 ± 2.6 %). However no relationship was observed between individual nutrition knowledge score and diet quality (r2 = 0.003, p = 0.63). No difference in dietary intake was observed between power and endurance athletes; average diet intake consisted of 57.0% carbohydrate, 17.1% protein and 25.9% fat. The metabolic typing diet is based around three different diets: high carbohydrate, high protein and mixed diet. The results from Chapter 5 identified that the metabolic typing questionnaire was not able to differentiate between metabolic function in healthy individuals. Additionally all athletes, independent of event (power vs. endurance), consumed similar diets. With such similarities a clearer understanding of the impact such diets have at a cellular level is required. Therefore for the remainder of the thesis it was decided to investigate the impact of dietary manipulation utilising more robust measures. Mitochondria are responsible for energy production; their quantity and density have been associated with improved health and endurance performance. External stressors such as energy reduction, carbohydrate restriction and exercise are potent stimulators of transcription markers of mitochondrial biogenesis. Thus manipulating carbohydrate and energy availability in vivo may enhance cellular adaptation and limited literature exists on the impact increased protein intake has on this. The aim of Chapter 6 was to investigate the impact of acute (7-day) continuous dietary manipulation on metabolic markers, body composition and resting metabolic rate (RMR). Using a repeated measures parallel group (N = 45) design, participants were randomly assigned one of four diets: high protein hypocaloric, high carbohydrate hypocaloric, high protein eucaloric or high carbohydrate eucaloric. The macronutrient ratio of the high protein diets was 40% protein, 30 % carbohydrate and 30% fat, the high carbohydrate diets were 10% protein, 60% carbohydrate and 30% fat. Energy intake in the hypocaloric diets was matched to resting metabolic rate (RMR). Participants consumed habitual diet for 7-days then baseline measures were collected (skeletal muscle biopsy, dual energy X-ray absorptiometry scan (DXA) and RMR, habitual diet was consumed for a further 7-days and repeat testing was completed (these time points were used as a control), the intervention diet was then consumed for 7-days and post measures were collected. The results of the skeletal muscle biopsy demonstrated no group x time interaction in any marker, however a pre-post time difference subsequent to the high protein hypocaloric diet (the diet which induced the greatest metabolic stress) was observed in four transcriptional markers of mitochondrial biogenesis (pre-post intervention fold increase: PCG1-α 1.27, AMPK 2.09, SIRT1 1.5, SIRT3 1.19, p < 0.05). The results of the DXA scan demonstrated that the high protein hypocaloric group lost significantly more fat mass than the high carbohydrate eucaloric group (-0.99 kg vs. -0.50 kg, p < 0.015). Irrespective of macronutrient ratio, no energy-matched between group difference was observed in lean mass (LM) loss. However when matched for macronutrient ratio the high protein diet attenuated LM loss to a greater extent that the high carbohydrate diet, suggesting an important role of increased protein intake in the maintenance of lean mass. No time point or group difference in RMR was observed. This data suggests that a high protein low carbohydrate hypocaloric diet may provide a stimulus to promote skeletal muscle metabolic adaptation. The aim of the final experimental Chapter (Chapter 7) in this thesis was to explore the impact exercise in combination with a high protein diet on metabolic adaptation, substrate utilisation and exercise performance in well trained runners. Using a parallel groups repeated measures study design the participants (well-trained endurance runners, N = 16) consumed normal habitual diet for 7-days, then 7-days intervention diet (high protein eucaloric or high carbohydrate eucaloric, same dietary ratios as Chapter 6) and finally returned to habitual diet for 7-days, training was consistent throughout. A pre exercise muscle biopsy was taken subsequent to each diet and immediately followed by a 10 km sub-maximal run and a time to exhaustion run (TTE) at 95% of velocity at maximal aerobic capacity (vV̇O2max). Post intervention the high protein group presented significant changes in sub-maximal substrate utilisation with 101% increase in fat oxidation (0.59 g·min-1, p = 0.0001). No changes were observed in substrate utilisation in the high carbohydrate group. A trend towards a reduction in average weekly running speed was observed in the PRO group (-0.9 km·h-1), the high carbohydrate group maintained the same training speed. TTE was decreased (-23.3%, p = 0.0003) in the high protein group subsequent to the intervention, no change was observed in subsequent to the high carbohydrate diet. / The high carbohydrate group demonstrated preferential increases in markers of metabolic adaptations (fold increase: AMPK = 1.44 and PPAR = 1.32, p < 0.05) suggesting that training intensity, rather than carbohydrate restriction, may be a more profound driver of metabolic adaptation. All performance measures, in both groups, returned to pre intervention levels once habitual diet was returned; however the increased gene expression observed in the high carbohydrate group remained elevated 7-days post intervention. The increased metabolic stress imposed by reducing carbohydrate intake did not increase transcriptional markers of mitochondrial biogenesis. For continuous endurance training and high intensity endurance performance a high carbohydrate diet is preferential to a high protein diet.

Afin d’approfondir les connaissances sur les liens entre la structure et les propriétés nutritionnelles de pâtes alimentaires, des pâtes de structures contrastées ont été produites en modifiant la formulation et/ou le procédé de fabrication. Leur structure a été caractérisée aux différentes échelles, puis les caractéristiques définies ont été reliées à la digestibilité des protéines et de l’amidon. Le remplacement partiel ou total de la semoule de blé dur par des farines de légumineuses dans la fabrication des pâtes a engendré des changements structuraux à plusieurs niveaux. Un affaiblissement marqué de la structure du réseau protéique proportionnel à la quantité de légumineuses incorporées a notamment été observé, conduisant à une amélioration de la digestibilité in-vitro des protéines. La digestibilité in-vitro de l’amidon était en revanche maintenue à son faible niveau voir même réduite. L’application de différentes températures de séchage a amplifié les modifications de la structure des pâtes enrichies en légumineuses. Ces différences s’atténuaient toutefois après cuisson, n’engendrant de fait aucune modification marquée de la digestibilité in-vitro des protéines entre une pâte cuite préalablement séchée à basse ou à très haute température. Les teneurs en facteurs antinutritionnels étaient réduites par le procédé de fabrication et de cuisson des pâtes. A quantité égale de protéines consommées, l’étude in-vivo a révélé une augmentation de la croissance et de l’efficacité azotée chez les rats nourris avec les pâtes mixtes blé/fèverole (65/35) par comparaison aux rats nourris aux pâtes 100% blé. L’assemblage blé/fèverole constituait un mélange protéique efficace pour la croissance des rats jeunes, probablement du fait de son profil équilibré en acides aminés, et de la structure affaiblie du réseau protéique. Par ailleurs, les différentes températures de séchage appliquées aux pâtes n’ont pas engendré de différence au niveau de la rétention azotée in-vivo. Les spécialités pastières développées au cours de ce travail de thèse sont jusqu’à trois fois plus riches en protéines et mieux équilibrées en acides aminés, comparativement à une pâte alimentaire classique. Ces spécialités pastières sont aussi appréciées que leurs homologues qui se vendent sur le marché. En outre, la recette mixte blé/légumineuses peut être adaptée à tout consommateur souhaitant diversifier ses apports en protéines, mais également à certaines populations spécifiques, notamment aux personnes âgées. La recette 100% légumineuses peut concerner les personnes intolérantes au gluten. Ces spécialités pastières pourraient de ce fait, trouver une large application à l’échelle industrielle. / In order to reinforce our knowledge about the links between structural and nutritional properties of pasta, pasta with contrasting structures were produced by modifying the formulation and/or the manufacturing process. Pasta structure was then characterized using a multiscale approach and linked to the digestibility of protein and starch fractions. The modification of pasta formulation by partial or total substitution of durum wheat semolina by legume flours changed pasta structure at various scales. A marked weakening in the structure of pasta protein network proportional to the level of legume protein incorporated was observed and related to an improvement of the in-vitro digestibility of pasta proteins. Interestingly the in-vitro starch digestibility was maintained at its low level or even reduced in mixed pasta. Changing pasta drying temperature amplified changes in the pasta structure. These differences, however, were reduced after cooking pasta resulting in unchanged in-vitro protein digestibility between cooked pasta dried at low or very high temperature. Anti-nutritional factors were also reduced by the manufacturing and cooking steps. In-vivo study revealed a better growth and nitrogen efficiency in rats consuming the same quantity of proteins from wheat/faba (65/35) mixed pasta in comparison with animals fed with 100% wheat pasta. This was probably related to the better amino acid balance, and to the weakened protein network in mixed pasta. The change in drying temperatures did not allow modifying protein retention. Pasta developed in this thesis are up to three fold richer in proteins than traditional wheat pasta, and more balanced in essential amino acids. Produced pasta are appreciated by the consumer as well as their counterparts found on the marked. Mixed wheat/legume may be adapted to the elderly people, whereas 100% legume pasta may be consumed by gluten intolerant subjects. Produced pasta may therefore find wide application at the industrial scale.

A dose response relationship between carbohydrate (CHO) ingestion and exercise performance has not been consistently reported. Additionally the underlying metabolic and neuromuscular explanations for an improvement in performance with increasing doses of CHO have not been fully explained. In Chapter 2 of this thesis 20 male cyclists completed 2 h of submaximal exercise followed by a time trial task (531 ± 48KJ). Three CHO electrolyte beverages, plus a control (water), were administered during a 2 h ride providing 0, 20, 39 or 64 g CHO·h-1 at a fluid intake rate of 1 L·h-1. Performance was assessed by time to complete the time trial task, mean power output sustained, and pacing strategy used. Mean task completion time (min:sec ± SD) for 39 g·h-1 (34:19.5 ± 03:07.1, p=0.006) and 64 g·h-1 (34:11.3 ± 03:08.5 p=0.004) of CHO were significantly faster than control (37:01.9 ± 05:35.0). The mean percentage improvement from control was -6.1% (95% CI: -11.3 to -1.0) and -6.5% (95% CI: -11.7 to -1.4) in the 39 and 64 g·h-1 trials respectively. The 20 g·h-1 (35:17.6 ± 04:16.3) treatment did not reach statistical significance compared to control (p = 0.126) despite a mean improvement of -3.7% (95% CI -8.8 to 1.5%). These data demonstrate that consuming CHO at a rate between 39 to 64 g·h-1 is likely to be optimal for most individuals looking to utilise a single source CHO as an ergogenic aid during endurance performances lasting less than 3 hrs. Attempts have been made to try and understand the acute metabolic regulation that occurs when ingesting increasing amounts of CHO. However, no one study has fully investigated the metabolic mechanisms underlying graded increments of CHO ingestion. In Chapter 3 we aimed to utilise stable isotopes and blood metabolite profiles to examine the integrated physiological responses to CHO ingestion when ingested at rates throughout the range where performance gains appear greatest. Twenty well-trained male cyclists completed 2 h constant load ride (95% lactate threshold, 185 ± 25W) where one of three CHO beverages, or a control (water), were administered every 15 min, providing participants with 0, 20, 39 or 64 g CHO·h-1 at a fixed fluid intake rate of 1L·h-1. Dual glucose tracer techniques (6,6,2H2 glucose and U13C labelled glucose) were used to determine glucose kinetics and exogenous carbohydrate oxidation (EXO) during exercise. Endogenous CHO contribution was suppressed in the second hour of exercise when consuming 39 and 64 g·h-1 in comparison to 0 g·h-1 (-7.3%, 95%CI: -13.1 to -1.6 and -11.2%, 95%CI: -16.9 to -5.5 respectively). Additionally, consuming 64 g·h-1 suppressed the endogenous CHO contribution by -7.2% (95%CI: -1.5 to -13.0) compared to the 20 g·h-1 treatment. Exogenous CHO oxidation rate increased by 0.13 g·min-1 (95%CI: 0.10 to 0.15) and 0.29 g·min-1 (95%CI: 0.27 to 0.31) when consuming 39 and 64 g·h-1 in comparison to 20 g·h-1 of CHO. Peak exogenous CHO oxidation rates were 0.34 (0.06), 0.54 (0.09) and 0.78 (0.19) g·min-1 for 20, 39 and 64 g·h-1 respectively. Plasma NEFA concentration was 0.10 (95%CI: 0.07 to 0.13), 0.12 (95%CI: 0.10 to 0.16) and 0.16 (95%CI: 0.13 to 0.19) mmol.L-1 higher when consuming 0 g·h-1 in comparison to 20, 39 and 64 g·h-1 respectively. Both 39 and 64 g·h-1 were effective at sparing endogenous CHO stores of which it is estimated that most of this is liver glycogen sparing, but the measured response was highly variable between individuals. Consuming 39 g·h-1 of CHO appears to be the minimum ingestion rate required to have a significant metabolic effect that results in an increase in performance. Recent research has indicated a key role of endogenous CHO sensing and oral glucose sensing in maintaining central drive and peripheral function during endurance exercise tasks. Consuming 39 and 64 g·h-1 of CHO elicits the greatest improvements in performance and also demonstrate a similar metabolic response. The improvement in subsequent time trial performance when consuming 39 and 64 g·h-1 coincided with significant alterations in whole body substrate usage that lead to endogenous CHO sparing at the same ingestion rates. In Chapter 4 we aimed to utilise gold standard neuromuscular function assessment techniques, alongside novel measures, to investigate the effect of consuming different rates of CHO on neuromuscular function during and following prolonged cycling exercise. In a double-blind, randomised cross-over design, well-trained male cyclists (n=20, mean±SD, age 34 ± 10 y, mass 75.8 ± 9 kg, peak power output 394 ± 36 W, V̇O2max 62 ± 9 ml·kg-1·min-1) completed 2 familiarisation trials then 4 experimental trials. Trials involved a 2 h submaximal ride followed by a high intensity time trial task lasting approx. 35 min with each of 0, 20, 39 and 64 g·h-1 CHO ingestion rates during submaximal exercise. Each trial involved pre and post exercise assessments (MVC, Mwave twitch potentiation and force, motor unit recruitment and firing rate assessment using high density EMG) and during exercise (gross EMG amplitude). MVC peak torque values were reduced post exercise by -20.4 nM (95%CI: -26.5 to -14.4) in comparison to pre value on all trials with no differences between trials. The firing rates of early recruited motor units significantly increased by 1.55 pps (95%CI: 0.51 to 2.59) following exercise in comparison to pre-exercise rates. Gross EMG during the 2 h cycling bout revealed a main effect of treatment (p<0.01) but post hoc comparisons provided no clarity and likely reflect methodological issues. Consuming CHO at ingestion rates between 20 and 64 g·h-1 had little to no impact on the neuromuscular function of well-trained cyclists when comparing pre and post fatiguing exercise values. Despite differences in time trial completion time between trials, following exercise to fatigue in an endurance task, no post exercise differences were detected.