Title Affiliations Abstract "Developing the next generation of robust Bayesian networks: theory and efficient inference algorithms for mixed credal networks" "Department of Electronics and information systems" "Credal networks are Bayesian networks with imprecise (interval-valued) local probabilities, thereby allowing for robust inferences. This project develops a new type of credal networks, called mixed credal networks. The advantage of this new type is that they do not suffer from the typical computational problems that occur for other types, which allows us to develop efficient inference algorithms." "Algorithms for reasoning in credal trees" "Department of Electronics and information systems, Department of Electromechanical, Systems and Metal Engineering" "Develop theory and efficient algorithms for inferences in credal trees, with emphasis on imprecise hidden Markov models (iHMM). These represent a systemU+2019s uncertain evolution through states, where we can only observe the states imperfectly, through uncertain outputs. I plan to adress: learning an iHMM model from sequences of observations, dealing with missing data and extending results to general credal trees." "Trade at the Crossroads of the Pacific and Indian Ocean Worlds: Credit Networks and the Financial Structure of Merchant Cooperation in Eighteenth-Century Manila" "Angela Schottenhammer" "Early Modern History (15th-18th Centuries), Leuven" "During the course of the Manila galleon era (1571–1815), merchant networks were essential to the operation of trans-Pacific and Indian Ocean routes, though until now we have had little systematic evidence of how these operated. The allocation of networkable resources, specifically credit, can be gleaned from Spanish archival record and acts as a proxy for a less tangible resource: mercantile trust. The National Archives of the Philippines (NAP) house a cache of maritime loans formalized before notaries in Manila during the 18th and early 19th centuries, which can be mined for data. These loans were a form of “respondentia,” where the principal’s cargo served as collateral. These credit contracts provide data on lenders and borrowers (ethnic Spaniards and Chinese) that can be used to recreate merchant networks through relational databases and Social Network Analysis. This kind of systematic study of Manila’s merchant population has been priorly impossible without the kind of extensive micro-level data captured in these contracts. Thus Spanish traders have been largely excluded from the literature on intra-Asian trade. These new archival sources and methodology present an empirical challenge to existing historical narratives of the Pacific and Indian Ocean economies, as well as being a contribution to the burgeoning study of self-organizing merchant networks in the early modern world." "Innovative credit scoring modeling using textual and social network data." "David Martens" "Engineering Management" "It is the purpose of this research project to come up with new, original and groundbreaking approaches for credit risk modeling through innovations in input data to model different aspects of credit risk. This research project will focus on the potential of social network and textual input data, and consists of four research objectives." "Innovative credit scoring modeling using textual and social network data." "David Martens" "Engineering Management" "It is the purpose of this research project to come up with new, original and groundbreaking approaches for credit risk modeling through innovations in input data to model different aspects of credit risk. For financial institutions, it is important to know which input data has the best prediction performance, how these data should be handled and which intrinsic characteristics should be taken into account to obtain the most accurate credit risk prediction. This research project will focus on the potential of social network and textual input data, and consists of four research objectives" "Gas cell stabilization in bread containing non-wheat cereals" "Jan Delcour" "Food and Microbial Technology (CLMT), Centre for Food and Microbial Technology" "Bread is an important staple food around the globe. In Europe and other parts of the world, it is most often made from wheat (Triticum aestivum L.) flour and mainly with a straight-dough process. This process starts by mixing flour, water, yeast, salt, and potentially a number of non-essential ingredients into viscoelastic dough. The dough is then fermented, which results in gas cell expansion and thus in an increased dough volume. Finally, the leavened dough is baked and the resultant bread cooled to room temperature.The loaf volume and crumb characteristics of bread are important quality characteristics which largely depend on the amount of gas cells incorporated during mixing and the degree to which they are stabilized throughout the bread making process. In wheat bread making, hydrated gluten proteins develop into a continuous, viscoelastic network which in the early stages of fermentation provides structural support to expanding gas cells and thereby stabilizes them. It has been suggested that this network fails to surround some areas of gas cells during the late stages of fermentation and early stages of baking as it ruptures as a result of dough expansion. From this moment onwards, proteins, surface-active lipids, and non-starch polysaccharides (NSPs) dissolved in a liquid film surrounding the gas cells supposedly take over their stabilization. These liquid films are believed to be part of the aqueous phase of dough. At least a fraction of this phase can be isolated from dough by ultracentrifugation. The supernatant obtained in this way is generally referred to as ‘dough liquor’ (DL).People today are aware of the potential health benefits of consuming mixed cereal breads. Indeed, partial replacement of wheat by for example rye (Secale cereale L.) or oat (Avena sativa L.) flour can increase bread dietary fiber and lysine (i.e. an essential amino acid) contents. However, mixed cereal breads are often of lower quality in terms of loaf volume and crumb structure than wheat breads because non-wheat cereals lack the typical wheat gluten proteins. Hence, it can be argued that the mechanism of gas cell stabilization by liquid films may be even more important in mixed cereal or non-wheat bread than in wheat bread making.To the best of our knowledge, no research has been conducted in this regard. Indeed, all studies available in literature today have focused on studying the chemical composition or functional properties of DL isolated from wheat dough. Thus, the potential of soluble constituents of non-wheat flour to stabilize gas cells in bread making has not yet been investigated, let alone that it would have been exploited.Against this background, the work in this dissertation was executed with the aim to explore the potential of soluble rye and oat flour constituents to stabilize gas cells in bread making. The work plan relied heavily on the use of DL as a model for the dough aqueous phase.In a first part, relations between (i) the chemical composition and (ii) the foaming and air-water (A-W) interfacial characteristics of wheat, rye, and oat DLs were established and hypotheses on the composition of DL stabilized A-W interfaces were brought forward. Wheat DL constituents produced a low amount of unstable foam. This was attributed to a low bulk phase viscosity and to them slowly developing a strongly viscoelastic mixed protein-lipid film at the A-W interface. In contrast, stirring rye DL solutions generated high volumes of foam ofpoor stability. The high initial foam volume was ascribed to a combined effect of a high bulk phase viscosity and a rapid formation of a strong predominantly viscous protein-dominated film at the A-W interface. The low initial foam volume produced from oat DL constituents was the result of lipids being the dominant constituents at oat DL stabilized A-W interfaces. This was deduced from a high total lipid content, very low surface tension, and absence of a viscoelastic film at the A-W interface of oat DL. As protein- or lipid-dominated A-W interfacial films are more resistant to deformations than mixed protein-lipid A-W interfacial films, rye and oat DL constituents seem to have more potential for stabilizing A-W interfaces than wheat DL constituents.In a second part, the hypotheses on the composition of the A-W interfaces stabilized by wheat, rye and oat DLs were tested and further refined by using DL modification strategies. First, the role of surface-active lipids in interfacial stabilization was studied by comparing the A-W interfacial properties of control and defatted wheat, rye, and oat DLs. Second, the role of NSPs was assessed by enzymatic depolymerization prior to studying DL bulk viscosity and A-W interfacial properties. Third, both treatments were combined to assess the extent to which the ability of DL NSPs to affect interfacial stability depends on the presence of lipids at the A-W interface. It was observed that NSPs contribute substantially to the bulk viscosity of wheat, rye, and oat DLs and thus likely also to the bulk viscosity of the aqueous phase in their respective doughs. In addition, it was established that by adsorbing at wheat and rye DL stabilized A-W interfaces lipids impair mutual interaction between adsorbed proteins. Surface tension measurements of control and defatted oat DL samples confirmed that lipids are the predominant DL constituent at oat DL stabilized A-W interfaces. Finally, irrespective of whether or not lipids were present at the A-W interface, wheat and rye DL arabinoxylan exerted a film weakening and strengthening effect respectively. This demonstrates that interaction between arabinoxylan and proteins at A-W interfaces in some but not all cases may improve their resistance to deformations. That proteins did not seem to be present at oat DL stabilized A-W interfaces supports the observation that oat DL β-D-glucan neither weakened nor strengthened the A-W interfacial film. Thus, wheat and rye DL stabilized A-W interfaces are composed of a mixed protein-lipid film with arabinoxylan acting as secondary layer, whilst a lipid film is present at oat DL stabilized A-W interfaces.In a third part, the composition of wheat, rye, and oat DLs and the A-W interfacial properties of their constituents were related to the loaf volume and crumb structure of breads prepared from their respective flours. In terms of loaf volume, wheat bread had a high specific volume despite the poor foaming and A-W interfacial properties of wheat DL constituents. This was of course mostly due to the viscoelastic gluten network which by displaying strain hardening acted as the primary gas cell stabilizing entity. In contrast, even though rye and oat DL constituents seemed to have more potential for stabilizing A-W interfaces than wheat DL constituents, the volumes of rye and oat bread loaves were much lower than that of wheat bread. Thus, assuming that rye and oat dough aqueous phase constituents contribute to gas cell stability in rye and oat bread making, they cannot match the efficiency of the combined contributions of the gluten network and dough aqueous phase constituents in terms of stabilizing gas cells in wheat bread making. However, in terms of crumb structure more gas cells per surface unit were observed in rye than in wheat and oat bread crumbs. Bread making experiments in which a xylanase preferentially hydrolyzing the water-extractable arabinoxylan population of rye flour was used revealed that arabinoxylan contributes substantially to the fine grained crumb of rye bread. Indeed, arabinoxylan enzymatic hydrolysis resulted in rye bread crumbs with considerably larger mean gas cell areas and lower numbers of cells per surface unit than was the case for control rye bread. This implies that rye flour arabinoxylan delays gas cell coalescence during rye bread making presumably because of its contribution to the bulk viscosity of the dough aqueous phase. To further assess the contribution of DL constituents to bread loaf volume, breads were prepared from doughs containing blends of commercial wheat gluten and commercial wheat starch, with and without addition of wheat, rye, or oat DL constituents. Overall it was observed that wheat, rye, and oat DL constituents result in a pronounced increase in the volume of such model breads. This implies that not only wheat gluten proteins but also DL constituents contribute to gas cell incorporation and/or stabilization in bread making. However, it should be mentioned that the addition of DL constituents likely changed the bulk rheology of the model doughs which in turn may have contributed to the above mentioned bread volume increase. Notable was that the addition of wheat DL constituents resulted in the most pronounced bread volume increase. This did not match our expectations based on the foaming and A-W interfacial characteristics of wheat, rye, and oat DLs. Thus, the mechanism by which DL constituents contribute to gas cell stabilization in bread making remained unclear at this point.In this context, it is important that stability of gas cells is not only determined by the characteristics of the interfaces surrounding them, but also by those of the liquid films between them. Moreover, A-W interfacial properties can often only be studied at concentrations lower than that found in the supernatant after ultracentrifugation (i.e. the ‘native concentration’). Therefore, to better understand the role of dough aqueous phase constituents in bread making, in a fourth part the drainage dynamics of free-standing DL thin films (both at lower and at native concentrations) were assessed. Comparison of the drainage times and interferometry images of DL thin films at lower and native bulk concentrations demonstrated that the DL bulk concentration has a drastic impact on the structure and stability of the obtained thin films. Whereas protein aggregates dispersed in mixed protein-lipid A-W interfacial films were characteristic of wheat DL thin films at low bulk concentrations, lipids were the dominant constituent at A-W interfaces of wheat DL thin films at their native concentration. Moreover, they stabilized it by diffusing along the A-W interfaces and thus by exerting Marangoni-type effects. Lipids also stabilized oat DL thin film A-W interfaces both at low and at their native bulk concentrations by exerting Marangoni effects and presumably by forming an immobile monolayer, respectively. In addition, wheat and oat DL thin films at their native concentrations exhibited stratification. This essentially means that the thin films were made up of stacked layers of supramolecular structures, in this case likely lipid micelles. If at least two of such layers are present, the layered structuring provides thin films with an additional degree of stability as it increases their disjoining pressure. Furthermore, protein aggregates in rye DL thin films at low bulk concentrations were surrounded by a relatively thick film. In addition, adsorbed proteins contributed to thin film stability by exerting steric and/or electrostatic repulsive protein-protein interactions. Finally, A-W interfaces of DL thin films at low bulk concentrations merged rapidly after drainage was forcibly induced, whilst DL thin films at their native concentrations were stable for up to at least three minutes of monitoring. This most important observation implies that DL constituents may contribute to the stability of gas cells in both wheat and non-wheat bread making. In conclusion, in this doctoral dissertation it was demonstrated that soluble wheat, rye, and oat flour constituents seem to have great potential for stabilizing gas cells in bread making. That wheat and oat DL thin films at their native concentrations had excellent stabilities combined with the observation that wheat, rye, and oat DLs increased the volume of model breads implies that gas cell stabilization by dough aqueous phase constituents is of importance both in wheat and non-wheat doughs. However, the volumes of rye and oat bread loaves were still much lower than that of wheat bread. This illustrates that the loaf volume of bread depends on the combined contributions of gluten proteins and of dough aqueous phase constituents. " "Additive Manufacturing for Self-Healing Robotics" "Guy Van Assche" "KU Leuven, Faculty of Engineering, Physical Chemistry and Polymer Science, Applied Mechanics, Robotics & Multibody Mechanics Research Group, Materials and Chemistry" "The increasing use of robots in close proximity to humans is giving a major impulse to the development of soft robotics, in which unexpected loads or impacts are dealt with by flexibility and adaptability rather than by making the structure stiffer and stronger than needed for its regular tasks. However, flexibility and softness of actuators and structures entails an inherent sensitivity to damage caused by sharp objects. Anyone having cut his fingers on the edge of a piece of paper can imagine that any human environment can be threatening to a robot with inflatable rubber fingers lacking our self-healing capacity. Over the past few years, we have been active in developing selfhealing rubbers and glasses, and their application in actuators for soft robotics, encompassing a selfhealing mechanical fuse, self-healing pneumatic fingers, and most recently self-healing pleated muscles. To create these actuators, classical production techniques like compression moulding were used, as well as a shaping-through-folding-and-self-healing method that exploits the self-healing properties of the materials. Nevertheless the shapes that can be produced are quite limited. Through the AMSeR project, we will bring additive manufacturing, with its inherent design freedom, to the world of self-healing soft robotics, jumping this field from manual labour into the 21st century make industry. This requires solving an intricate puzzle of material design, material behaviour, and processing methods." "A mechanistic understanding of the effects of yeast and yeast fermentation on the rheology of leavening cereal dough systems" "Paula Moldenaers" "Soft Matter, Rheology and Technology Section, Soft Matter, Rheology and Technology (SMaRT)" "Although bread-making has been practised for millennia, our fundamental scientific understanding of this process is still surprisingly limited. For the preparation of bread dough, only four ingredients are essential: wheat flour, water, salt and yeast. Wheat flour contains mostly starch and gluten proteins. During mixing, which involves substantial shear and extensional deformations, the gluten proteins tend to form a network, thereby encapsulating the individual starch granules. During the fermentation stage and oven rise, this gluten-starch matrix experiences an additional extensional deformation, as it is leavened by the expansion of carbon dioxide gas cells. A very delicate balance between material flowability on the one hand, and material stiffness on the other hand, is required to not only ensure a proper leavening of the dough but also to guarantee shape stability of the baked product afterwards. The rheological properties of dough are thus intrinsically linked to the final quality of the baked product. Hence, an in-depth understanding of how the different flour constituents determine the rheological behaviour of dough is highly desirable.To elucidate the individual contributions of gluten and starch to the overall dough behaviour, the rheological properties of dough and mixtures of different gluten-starch ratios were studied systematically in shear and extension, as both deformation types are frequently encountered in the bread-making process. The dough response in shear was studied by means of linear small-amplitude oscillatory tests and non-linear creep-recovery tests. The behaviour of dough under uniaxial extensional deformations was investigated with an extensional viscosity fixture mounted on a rotational rheometer. The starch component turned out to play a pivotal role in linear dough rheology. With increasing starch content, the linearity limit observed in oscillatory shear tests decreased as a power-law function. Starch also clearly affected the extensional viscosity at small strains. Consequently, in the linear region differences between different gluten systems may become obscured by the presence of starch. As bread-making qualities are known to be linked to the gluten network, it is imperative to probe the non-linear behaviour of dough in order to expose differences in flour quality. The quality differences between a strong and a weak flour type were revealed most clearly in the value of the strain-hardening index in uniaxial extension and the total recovery compliance in non-linear creep-recovery tests.The rheological properties of wheat flour dough are known to be very sensitive to small changes in water content and mixing time. At sufficiently high water levels, a free-water phase exists in dough, which attenuates the starch-starch and gluten-starch interactions. Increases in the water content were found to result in a parallel, downward shift in the dynamic moduli and the extensional viscosity at small to moderate strains, and a concomitant increase in the linear creep compliance. The impact of changes in the water content can thus be captured by a simple scaling law. Dough characterisation after different mixing times showed that overmixing may cause a disaggregation or even depolymerisation of the gluten network. The network breakdown, as well as the subsequent (partial) recovery, were clearly reflected in the value of the strain-hardening index, for which a maximum was reached at a mixing time close to the optimum as determined with the mixograph. Finally, the gluten proteins turned out to be much less susceptible to overmixing in an oxygen-lean environment, which demonstrates the significant role of oxygen in the degradation process.To improve the bread-making performance of wheat flours, enzymes such as glucose oxidase and transglutaminase are frequently included in the dough recipe, as these enzymes have the ability to considerably alter the viscoelastic nature of the gluten network. To evaluate the impact of these enzymes on a flour's bread-making performance, the rheological implications of adding glucose oxidase or transglutaminase to wheat flour dough were investigated with the adequate rheological toolbox developed previously. The enzymes enhanced the elastic character of dough until saturation was reached. In the bread-making process, the use of excessive amounts of enzyme turns out to be counterproductive. Whereas the dynamic moduli did not show a maximum as a function of enzyme content, the strain-hardening index clearly revealed this overcross-linking effect. Besides enzymes, the gluten network can also be reinforced by adding supplementary gluten proteins, which were indeed found to enhance the extent of strain-hardening as well.In this project, we also aimed at revealing the mechanisms responsible for the changes in the rheological properties of dough as a result of fermentation. Despite the obvious importance of the fermentation step in the bread-making process, the number of (fundamental) rheological studies dealing with fermented dough is surprisingly limited. By adding the main yeast metabolites (besides carbon dioxide) to unfermented dough at the concentrations observed in fermented dough, the associated rheological changes could be determined with our fundamental rheological techniques. Glycerol was found to have a softening effect on dough similar to water. Ethanol equally led to decreased values of the moduli, but its effect was not merely diluting: ethanol fundamentally altered the configuration of the gluten network, resulting in a decrease in the dough's extensional viscosity and extensibility. The stiffness and the extensional viscosity of the gluten network were also negatively affected by succinic acid and glutathione. Subsequently, the impact of these metabolites on the rheology of dough was also investigated in situ by examining the rheological behaviour of the dough matrix after fermentation had been completed. Compared to unfermented control dough, the fermented dough matrix exhibited reduced extensibility and a lower maximum extensional viscosity. The storage modulus was also negatively affected, but only at low frequencies. The observed changes could partially be accounted for by the yeast metabolites, yet it was clear that the rheological behaviour of the fermented dough matrix did not merely resemble a superposition of the rheological changes associated with the main yeast metabolites. The differences could perhaps be attributed to other rheologically active components released by yeast during fermentation, or might reflect the time-dependent accumulation of metabolites in an already expanding gluten network during fermentation.Characterisation of the rheological properties of fermenting dough, including the carbon dioxide gas bubbles, is essential to understand the real dough behaviour during processing and to develop a firm understanding of the kinetics of dough fermentation. However, the rheological study of fermenting dough poses great challenges as the dough samples are extremely fragile and their properties change considerably over time. In order to track the time evolution of the dynamic moduli and the density of fermenting dough, a parallel-plate rheometer add-on with adjustable gap was developed. Overfilling effects were taken into account by establishing a calibration curve with unfermented dough. Over the course of two hours, both dynamic moduli exhibited a sharp decline, eventually reaching a steady-state value. As yeast produces several other metabolites besides carbon dioxide gas that are able to alter the viscoelasticity of the gluten-starch matrix, the decrease in the dynamic moduli with increasing fermentation time did not match exactly the time evolution of the dough density. Frequency sweep snap-shots at specific points in time were obtained in multiwave mode and indicated that already early on in the fermentation process, substantial changes occur in the rheological response of dough. The available level of salt (NaCl) and sugar (sucrose) had a clear impact on the rheological behaviour of (unfermented) dough and the fermentation kinetics. To study the latter, the results of the linear oscillatory tests were combined with gas production data obtained with a rheofermentometer. The presence of salt resulted in a stronger gluten network and a slower (and therefore better controllable) fermentation process. Following the addition of sucrose, the dough became softer as the free aqueous phase expanded in volume. The total amount of gas produced increased, even though initially a dip in the gas production rate could be observed as the yeast required time to adjust to the osmotic stresses induced by the high sucrose concentration.The combined operation of fundamental and empirical rheological techniques clearly constitutes a valuable means to study the rheological behaviour of wheat flour dough and to assess the impact of yeast fermentation thereon. The developed rheological methodology can be used further to obtain a deeper understanding of the role of the minor flour components (e.g. arabinoxylan, albumins and globulins, etc. ) in determining the rheological properties of dough and hence the final product quality. In addition, the procedures outlined in this dissertation allow to quickly screen a multitude of yeast strains (each strain having its own metabolic profile) in order to identify those yeast strains that have the potential to improve the stiffness and extensibility of the gluten-starch matrix via their excreted metabolites. This dissertation is thus part of the ongoing effort to further improve the bread-making process, as also in the 21st century bread is still considered to be the staff of life." "A colloidal approach to study the role of cereal flour constituents in mixed cereal breadmaking and product quality." "Jan Delcour" "Centre for Food and Microbial Technology" "Aeration during their production is important for the structure of many food products. In bread, proper gas cell stabilization results in a high loaf volume and a homogeneous fine grained crumb. In wheat breadmaking, the gluten proteins are initially responsible for stabilizing the gas cells incorporated during dough mixing. However, at a certain point during fermentation, a liquid film takes over gas cell stabilization. This film is made up by the dough aqueous phase. It contains surface active proteins, lipids, and carbohydrates, which impact its viscosity. In dough for mixed cereal bread systems, for which there is an increasing consumer demand, liquid films stabilizing gas cells during fermentation and baking are presumably more important than in wheat dough due to the weakened or absent viscoelastic gluten network. However, no information is available on gas cell stabilization by liquid films in mixed cereal bread systems. The importance of this mechanism in mixed cereal bread systems will be studied here by using an innovative in-depth colloidal approach to characterize their dough aqueous phases (surface tension, interfacial rheology,..). Their properties will be related to mixed cereal bread quality properties (loaf volume and/or crumb structure). Better understanding of factors contributing to gas cell stability in mixed cereal systems and hence bread quality will provide additional criteria for improved selection of cereal raw materials for mixed cereal breadmaking." "Study of the impact of thermo-active starch-modifying enzymes and surrounding wheat (Triticum aestivum L.) cell wall and protein matrices on in vitro starch digestibility in model and bread systems" "Jan Delcour" "Translational Research in GastroIntestinal Disorders, Food and Microbial Technology (CLMT)" "Starch is the main glycemic carbohydrate in human diet. High glycemic responses are an independent risk factor for developing diabetes. It is, therefore, vital to comprehend the factors that dictate starch digestion and to develop strategies to (re)formulate foods with enhanced nutritional functionality. In this context, the molecular structure of starch as well as the surrounding food matrix including the native plant cellular structure can play a decisive role on starch digestion. Targeted manipulations of these features could have a profound impact on starch digestibility and postprandial glycemic responses.Against this background, the objective of this doctoral dissertation was to explore the potential of native wheat endosperm cellular structures and enzymatic modifications of starch to modulate starch digestibility.In a FIRST PART, intact wheat endosperm cell walls were considered as a first tool to modulate starch digestibility and their role on swelling, gelatinization and digestibility of intracellular starch was studied by utilizing wheat endosperm milling fractions with different cell wall integrity levels. The swelling capacity of starch and the viscosifying potential of coarse farina (i.e. a wheat endosperm milling fraction with an average diameter of 705 µm), which contains a substantial portion of intact cells, were consistently lower than those of flour (i.e. the wheat endosperm milling fraction with an average diameter of 85 µm) and fine farina (i.e. a wheat endosperm milling fraction with average diameter of 330 µm) which contained no intact cells. This illustrates that starch present in intact cells did not swell to its full capacity. It was noted that the swelling capacity and the viscosity of only coarse farina increased upon enzymatic degradation of the cell wall polysaccharides and that, as a result of the treatment, they reached levels similar to those observed for flour and fine farina. Overall, it was demonstrated that the intact cell walls protected starch from extensive swelling upon heating in excess of water.The cell wall integrity negatively impacted in vitro starch digestibility. The extent of starch digestion was similar for all gelatinized samples, but the digestion rate constant was 42% lower for coarse farina (0.37 min-¹) than for flour (0.63 min-¹) and fine farina (0.64 min-¹). That indicated that the intact cells in coarse farina limit the diffusion of the pancreatic α-amylase and reduce the digestion rate constant. However, the similar extent of digestion suggests that the cell walls are permeable. Next, the cell walls were enzymatically degraded after hydrothermal processing to obtain starches differing in swelling behavior. In this case, the digestion rate constant became equal in all cases. This series of experiments showcased that the preservation of cell wall integrity can delay diffusion of amylolytic enzymes and lower the rate of digestion of intracellular starch irrespective of starch swelling. Coarse farina with intact cells could, hence, be used a functional ingredient to lower starch digestibility.In order to provide proof-of-concept, coarse farina containing intact cell walls was incorporated in a bread recipe and the impact on bread quality characteristics and the digestibility of its starch was studied. When substituting up to 80% of the white wheat flour with coarse farina, the bread loaf volume was decreased from 5.9 down to 4.5 mL/g and, in turn, the firmness of its crumb was increased from 0.8 up to 1.2 N. These observations were attributed to the bread having a disrupted gluten network in the presence of coarse farina particles and the unavailability of the gluten enclosed in cells to participate in the network formation. However, the quality of the bread produced from coarse farina containing recipes was acceptable. The cells in coarse farina remained seemingly intact in the bread crumb. Nevertheless, the extent and rate constant of starch digestion did not change, irrespective of the levels of coarse farina incorporation. The gluten network in regular breads may well be a sufficient barrier to starch digestion while in the breads produced from coarse farina containing recipes the contribution of starch encapsulation into the cells on starch digestibility does not compensate for the loss of the barrier effect from the disrupted gluten network.In a SECOND PART, enzymatic modifications of the molecular structure of starch were investigated as a second tool to modulate starch digestibility. The thermoactive amylomaltase (AMM) from Thermus thermophilus was employed to modify starch. This enzyme catalyzes the transfer of glucan segments from amylose (AM) to elongate the amylopectin (AP) chains via an intermolecular transglucosylation reaction. This enzyme was chosen since its potential to decreasing the digestibility of the modified starch has not been explored in detail, despite its unique action pattern and literature based evidence that longer AP chains in retrograded starch may be digested more slowly.  At low dosages of the enzyme in starch slurries, only partial shortening of the AM chains was observed which led to increased viscosity upon cooling due to the enhanced aggregation of the more mobile shorter AM chains. With increasing dosages, the peak viscosity in Rapid Visco Analysis upon heating increased from 2,825 to 3,050 mPa·s because of the quite extensive depolymerization of AM which, in its unmodified form, inhibits starch swelling. In parallel, the AP chains were gradually elongated and AM was gradually degraded to even result in its complete depletion. When the modified starches were stored at low temperature, the extent of starch retrogradation, here expressed as melting enthalpy, increased from 0.8 up to 7.7 J/g starch in line with the extent of AP chain elongation. Finally, the rate constant and extent of starch digestion were drastically reduced upon enzymatic modification of starch from 0.6 down to 0.2 min-¹ and from 54.5 down to 35.9 g digested starch per 100 g starch, respectively. The AP chain lengths were positively related with the melting enthalpies of retrograded starches, which, in turn, were negatively correlated with the extent and rate constant of in vitro digestion. Evidently, different enzyme dosages in starch slurries resulted in different molecular structures that induced different viscosity development while the considerable elongation of AP chains was found to be important for lowering the susceptibility of starch to digestion upon retrogradation.The next step was to in situ modify starch during bread making with starch-modifying enzymes [AMM and maltogenic α-amylase (MA)] and to investigate the impact on bread characteristics, starch retrogradation and digestibility. It was hoped that it would be possible to provide proof that enzymes can be used in food systems to lower starch digestibility. MA was also examined here as it is widely used in the bread making industry. The use of AMM led to partial shortening of the AM chains but not to a detectable level of AP chains elongation probably due to insufficient activity in the dough. The treatment had only a small negative impact on the rate of bread crumb firming upon storage and influenced neither starch retrogradation nor its digestibility. The addition of MA led to shortening of the AP chains. Indeed, the relative content of short chains with ≤ 8 glucose units increased from 8.8 to 13.8%. Starch retrogradation and crumb firmness were reduced upon storage of breads due to the incapacity of the shortened AP chains to re-crystallize. Interestingly, the extent of starch digestion declined from 46.7 to 38.0 g digested starch per 100 g crumb. Indeed, pancreatic α-amylase cannot bind close to the branching points of AP and when its chains are shortened, productive binding is hindered. Overall, AMM had only very limited functionality during bread making and no effect on starch digestibility, while MA presented a dual functionality in that it both decreased the crumb firmness upon storage and the extent of starch digestion.Lastly, we investigated the difference between postprandial glycemia after consumption of a cold-stored oatmeal porridge containing either unmodified or AMM modified starch. It was reasoned that oatmeal porridge is a food system the preparation of which can accommodate for sufficient AMM activity AMM. The extent of starch retrogradation, also here expressed as the corresponding enthalpy, increased from 0 (non detectable) to 5.3 J/g oats because of the enzymatic modification which probably resulted in AP chain elongation to an extent sufficient to promote re-crystallization. The postprandial glycemic responses were not significantly lowered and the enzymatic modification thus had no impact. This lack of effect may have been due to the texture-mediated effect of AM degradation on starch accessibility in the gastrointestinal tract, the potentially insufficiently increased extent of retrogradation as well as to the large inter-individual variation.In conclusion, this doctoral dissertation illustrates the potential of strategies such as starch encapsulation into the intact native wheat endosperm cells and enzymatic starch modifications for modulating starch digestibility. Further research should focus on the application of such insights in a wide range of common cereal-based food products such as other types of (whole-meal) breads, cookies or puddings as well as on validating any health benefits with randomized clinical trials."