No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

La cuisson est un procédé universel de transformation de la matière première carnée en aliment. Le chauffage de la matrice musculaire conditionne diverses qualités organoleptiques, technologiques, sanitaires et nutritionnelles des viandes cuites. La cuisson étant appliquée de plus en plus fréquemment en conditions industrielles standardisées, il est pertinent de modéliser certains mécanismes clés dans le déterminisme de ces qualités afin d’optimiser le procédé. Pour cela, une démarche expérimentale originale a été mise en place fondée sur une analyse quantitative, locale, dynamique et in situ de la viande pendant la cuisson. Cette démarche ne fait pas d’hypothèse réductionniste en étudiant un échantillon intact à l’échelle de l’aliment consommé, ni d’hypothèse simplificatrice en prenant en compte les variations spatiales de la température dans l’échantillon. Elle s’appuie sur des développements originaux en imagerie par résonance magnétique nucléaire à haut champ et en traitement d’images pour cartographier la déformation et la quantité d’eau. Des modèles robustes liant température, déformation et quantité d’eau ont été obtenus pour des muscles de teneur variable en tissu conjonctif. Les résultats montrent principalement une augmentation de la déformation avec la température en plusieurs phases dont les caractéristiques dépendent de la composition du muscle, et une diminution de la quantité d’eau avec la température. Tous ces résultats sont discutés et interprétés au regard du comportement à la température des différents composants du muscle. Ce travail montre d’abord que l’imagerie dynamique, quantitative et multiparamétrique permet de décrypter des mécanismes intervenant lors de la cuisson des viandes sans établir des hypothèses réductrices lors de l’interprétation de ces phénomènes. Elle a conduit de plus, à des développements méthodologiques applicables à d’autres champs et ouvre la voie à d’autres investigations dans le domaine de l’optimisation qualitative des produits carnés transformés. / Cooking is a general process which transforms the meat raw material into food. The heating of muscle matrix influences different organoleptic, industrial, health and nutritional qualities of cooked meat. Cooking being applied more and more frequently in standardized and industrial conditions, it makes sense to model some key mechanisms which determine the latter qualities in order to optimize the process. For this purpose, an original experimental approach has been developed based on a quantitative, local, dynamic and in situ analysis of the meat during cooking. This approach is not based on any reductionist hypothesis by studying an intact sample at the scale of the consumed food, nor by the simplifying assumption of taking into account the spatial variations of temperature in the sample. It is based on original developments in nuclear magnetic resonance imaging at high-field and on image processing in order to map deformation and the water content. Robust models linking temperature, deformation and water content were obtained for muscles differing from their content in connective tissue. The results mainly show a deformation increase with the temperature in several phases whose characteristics depend on the muscle composition, and a decrease in the water content with temperature. All these results are discussed and interpreted thanks to the temperature behavior of the various muscle components. This work first shows that quantitative, multi-parametric and dynamic imaging can decipher the mechanisms involved during meat cooking, without formulating simplifying assumptions in the interpretation of these phenomena. Furthermore, it has led to methodological developments applicable to other fields and paves the way for further investigations in the field of quality optimization of processed meat products.

Abstract The subject of this thesis is flow imaging by methods based on the nuclear magnetic resonance (NMR) phenomenon. The thesis consists of three related topics: In the first one the feasibility of measuring velocity maps and distributions inside a microfluidic device by pulsed field gradient (PFG) NMR has been demonstrated. The second topic was to investigate microfluidic gas flow using a combination of a special detection technique and a powerful signal enhancement method. The third topic is related to the unambiguous determination of velocities under challenging experimental conditions and introduces a new, improved velocity imaging sequence. In the first part, well established imaging methods have been used to study water flow inside a micromixer. A surface coil matching the region of interest of the mixer was home built and used in the measurements in order to gain a better signal-to-noise ratio. Velocities inside the mixer have been measured by phase-encoding velocity, with unprecedented spatial resolution. Two dimensional NMR imaging and velocity maps revealed clogging and different manufacturing qualities of the mixers. In addition to the velocity maps, which display an average velocity for spins within one pixel, complete velocity distributions (so called average propagators) were measured. It was found that in the absence of spatial resolution in the third dimension, the propagator data can provide valuable insight to the flow system by revealing overlapping flow passages. The next topic was gas flow inside a microfluidic device. It was investigated by time-of-flight flow imaging. The measurement of the weak gas signal was enabled by the use of two signal enhancement techniques: remote detection NMR and parahydrogen induced polarization (PHIP). The results demonstrate that a very significant signal enhancement can be achieved by this technique. In the future it may enable the investigation of interesting chemical reactions inside microreactors. The third and last topic of the thesis deals with measuring flow by the so called multiecho sequences. When multiecho sequences are used in combination with phase encoding velocity, an error may be introduced: the multiecho sequence may produce a cumulative error to the phase of the magnetization, if it is sensitive to RF pulse imperfections. The problem has been elaborately explained and various solutions discussed, among the newly proposed one. Experimental results demonstrate the performance of the new velocity imaging sequence and show that the new sequence enables the unambiguous determination of velocities even in challenging experimental conditions resulting from inhomogeneous radio frequency fields of the measurement coils.