Pattern separation is the means by which the brain discriminates similar experiences. It enables retrieval of individuated memories without confusing them with other memories. It is the reason one remembers where one parked the car today and does not mix it up with where one parked it previously. Adult neurogenesis refers to the ongoing production of neurons in the mature brain. One of the likely roles of adult neurogenesis in the hippocampus is facilitating pattern separation. Induced reduction of adult neurogenesis in non-human animals is associated with depression- and anxiety-like behaviors. One possible explanation is that reduced neurogenesis leads to reduced pattern separation, further leading to overgeneralization of threat situations. Instead of perceiving threats where it should, the animal risks perceiving threats everywhere. Emotional resilience is the ability to recover from adversity with a minimum of lingering negative effects such as depression or anxiety. This thesis investigates whether pattern separation in the human hippocampus supports emotional resilience. I performed a systematic review of studies that used the Mnemonic Similarity Task – a memory task commonly used to measure human pattern separation – to investigate the relationship between pattern separation and anxiety. The results are inconclusive but suggest a possible interaction effect whereby pattern separation and high-arousal states like stress predict anxiety. Together with the evidence from the non-human animal studies, this suggests that reduced pattern separation as caused by reduced neurogenesis could make one vulnerable to developing anxiety disorders.

Episodic memory binds together the people, objects, and locations that make up the specific events of our lives, and allows the recall of our past in the service of current and future goals. Recent models of memory have posited that the hippocampus instantiates computations critical for episodic memory including mnemonic discrimination, relational binding, and holistic retrieval. Collectively, this set of studies aim to chart the ontogeny of each key components of episodic memory. We found robust improvements in children’s abilities to form complex relational structures and to make fine-grained discrimination for individual items from age 4 to age 6. However, relational memory dependent on context discrimination appears to follow a more protracted development. Furthermore, relational binding and mnemonic discrimination (item and context levels) undergo age-related decrements in senescence. Despite relatively poor relational binding capabilities, children as young as age 4 are able to retrieve multi-element events holistically, such as successfully retrieving of one aspect of an event predicts the retrieval success of other aspects from the same event. Critically, the degree of holistic episodic retrieval increases from age 4 to young adulthood. This multi-process approach provides important theoretical insights into lifespan profile of episodic memory. / Psychology

In this work, we examined the executive functions of the hippocampus at the behavioral level as a so-called behavioral separation in adult rats. We studied an impact of day-to-day alternation versus sequential learning (and the order of learning) of two spatial tasks (Morris Water Maze and Active Allothetic Place Avoidance) testing different hippocampal functions (experiment 1), or an impact of sequential versus alternating learning of one task (Active Allothetic Place Avoidance) in two different rooms (experiment 2), on performance. We found out that rats are able to learn both tasks as well as to discriminate between the two contexts regardless of the order or alternating of learning. Because such executive functions are impaired in human patients suffering from schizophrenia, we used this procedure also in the rat model of schizophrenia induced by acute intraperitoneal application of dizocilpine (MK-801), glutamate NMDA receptors antagonist, in the dose of 0.08 mg/kg. We failed to selectively disrupt the behavioral separation, however, we observed general learning deficit and hyperlocomotion regardless of the alternation in the Active Allothetic Place Avoidance task in these rats. The cognitive impairments in connection with learning after such low dose of MK-801 in this task have not yet been...

Dans ce travail de thèse, j'ai cherché à comprendre les mécanismes de contrôle des fonctions mnésiques par Rxrγ ainsi que l’implication potentielle de la signalisation dopaminergique dans ces mécanismes. Dans ce cadre, j'ai focalisé mon étude sur les fonctions hippocampiques à différent niveaux. J'ai cherché (1) à identifier les populations cellulaires de l'hippocampe exprimant Rxrγ en utilisant des techniques histologiques(immunohistochimie, hybridation in situ) afin de pouvoir (2) étudier les fonctions électrophysiologiques de ces cellules en utilisant la technique du patch-clamp. Pour comprendre le rôle de Rxrγ dans le contrôle des fonctions autonomes de ces cellules et les conséquences sur le réseau neuronal environnant, j’ai étudié les effets de la perte de fonction chez les souris Rxrγ/.Étant donné que différentes sous régions de l'hippocampe sont impliquées dans différents processus mnésiques, et que notamment le gyrus denté a été associé au processus de pattern separation (Leutgeb et al., 2007), j'ai également (3) cherché à préciser les fonctions mnésiques qui dépendent de l'activité de Rxrγ en réalisant des analyses comportementales chez les souris Rxrγ/. Etant donné que Rxrγ possède une activité transcriptionnelle, entre autre sur le gène du récepteur D2 à la dopamine (Drd2) (Samad et al., 1997), j'ai également (4) étudié la signalisation dopaminergique dans l’hippocampe chez les souris sauvages et les mutants nuls Rxrγ/. Enfin, afin de démontrer la spécificité neuroanatomique et homéostatique du contrôle exercé par Rxrγ sur la mémoire,j’ai procédé à (5) l’inactivation spécifique de Rxrγ dans les hippocampes de souris floxées pour ce gène, au moyen d’un virus AAV exprimant la recombinase Cre. / The present thesis work is an attempt to understand the mechanisms of Rxrγ control of memory functions, as well as the potential involvement of dopaminergic signaling in these mecanisms. In this context, I focused my research on hippocampal functions at several distinct levels. The first part of my work (1) aimed at defining the hippocampal cell populations expressing Rxrγ using various histological techniques (immunohistochemistry, in situhybridization) in order to (2) study the electrophysiological functions of these cells using invitro patch-clamp.To identify the role of Rxrγ in the control of cell autonomous functions, as well as the consequences on the surrounding network, I have studied the effects of its loss of function in Rxrγ/mice.As the different subregions of the hippocampus are implicated indistinct aspects of learning and memory, and in particular the dentate gyrus being associated with pattern separation (Leutgeb et al., 2007), I have also tried to dissect the mnemonic processes that rely on Rxrγ activity by performing behavioral analyses of Rxrγ/mice. Considering the transcriptional activities of Rxrγ on Drd2, I have also (4) studied dopaminergic signaling in the hippocampus of wild type and Rxrγ null mutant mice. Finally, to demonstrate the neuroanatomical and homeostatic specificity of Rxrγ control on memory, I performed (5) specific inactivations of Rxrγ in hippocampi of conditional mutant mice that possessed floxed Rxrγ, using AAV vectors expressing recombinase Cre.

Optimization is a scientific discipline that lies at the boundary between pure and applied mathematics. Indeed, while on the one hand some of its developments involve rather theoretical concepts, its most successful algorithms are on the other hand heavily used by numerous companies to solve scheduling and design problems on a daily basis. Our research started with the study of the conic formulation for convex optimization problems. This approach was already studied in the seventies but has recently gained a lot of interest due to development of a new class of algorithms called interior-point methods. This setting is able to exploit the two most important characteristics of convexity: - a very rich duality theory (existence of a dual problem that is strongly related to the primal problem, with a very symmetric formulation), - the ability to solve these problems efficiently, both from the theoretical (polynomial algorithmic complexity) and practical (implementations allowing the resolution of large-scale problems) point of views. Most of the research in this area involved so-called self-dual cones, where the dual problem has exactly the same structure as the primal: the most famous classes of convex optimization problems (linear optimization, convex quadratic optimization and semidefinite optimization) belong to this category. We brought some contributions in this field: - a survey of interior-point methods for linear optimization, with an emphasis on the fundamental principles that lie behind the design of these algorithms, - a computational study of a method of linear approximation of convex quadratic optimization (more precisely, the second-order cone that can be used in the formulation of quadratic problems is replaced by a polyhedral approximation whose accuracy that can be guaranteed a priori), - an application of semidefinite optimization to classification, whose principle consists in separating different classes of patterns using ellipsoids defined in the feature space (this approach was successfully applied to the prediction of student grades). However, our research focussed on a much less studied category of convex problems which does not rely on self-dual cones, i.e. structured problems whose dual is formulated very differently from the primal. We studied in particular - geometric optimization, developed in the late sixties, which possesses numerous application in the field of engineering (entropy optimization, used in information theory, also belongs to this class of problems) - l_p-norm optimization, a generalization of linear and convex quadratic optimization, which allows the formulation of constraints built around expressions of the form |ax+b|^p (where p is a fixed exponent strictly greater than 1). For each of these classes of problems, we introduced a new type of convex cone that made their formulation as standard conic problems possible. This allowed us to derive very simplified proofs of the classical duality results pertaining to these problems, notably weak duality (a mere consequence of convexity) and the absence of a duality gap (strong duality property without any constraint qualification, which does not hold in the general convex case). We also uncovered a very surprising result that stipulates that geometric optimization can be viewed as a limit case of l_p-norm optimization. Encouraged by the similarities we observed, we developed a general framework that encompasses these two classes of problems and unifies all the previously obtained conic formulations. We also brought our attention to the design of interior-point methods to solve these problems. The theory of polynomial algorithms for convex optimization developed by Nesterov and Nemirovsky asserts that the main ingredient for these methods is a computable self-concordant barrier function for the corresponding cones. We were able to define such a barrier function in the case of l_p-norm optimization (whose parameter, which is the main determining factor in the algorithmic complexity of the method, is proportional to the number of variables in the formulation and independent from p) as well as in the case of the general framework mentioned above. Finally, we contributed a survey of the self-concordancy property, improving some useful results about the value of the complexity parameter for certain categories of barrier functions and providing some insight on the reason why the most commonly adopted definition for self-concordant functions is the best possible.

Indiana University-Purdue University Indianapolis (IUPUI) / Down syndrome (DS) is caused by three copies of human chromosome 21 (Hsa21) and results in an array of phenotypes including intellectual disability. Ts65Dn mice, the most extensively studied DS model, have three copies of ~50% of the genes on Hsa21 and display many phenotypes associated with DS, including cognitive deficits. DYRK1A is found in three copies in humans with Trisomy 21 and in Ts65Dn mice, and is involved in a number of critical pathways including CNS development and osteoclastogenesis. Epigallocatechin-3-gallate (EGCG), the main polyphenol in green tea, inhibits Dyrk1a activity. We have shown that a three-week EGCG treatment (~10mg/kg/day) during adolescence normalizes skeletal abnormalities in Ts65Dn mice, yet the same dose did not rescue deficits in the Morris water maze spatial learning task (MWM) or novel object recognition (NOR). Others have reported that An EGCG dose of 2-3 mg per day (90mg/ml) improved hippocampal-dependent task deficits in Ts65Dn mice. The current study investigated deficits in a radial arm maze pattern separation task in Ts65Dn mice. Pattern separation requires differentiation between similar memories acquired during learning episodes; distinguishing between these similar memories is thought to depend on distinctive encoding in the hippocampus. Pattern separation has been linked to functional activity of newly generated granule cells in the dentate gyrus. Recent studies in Ts65Dn mice have reported significant reductions in adult hippocampal neurogenesis, and after EGCG treatment, enhanced hippocampal neurogenesis. Thus, it was hypothesized that Ts65Dn mice would be impaired in the pattern separation task, and that EGCG would alleviate the pattern separation deficits seen in trisomic mice, in association with increased adult hippocampal neurogenesis. At weaning, Ts65Dn mice and euploid littermates were randomly assigned to the water control, or EGCG [0.4 mg/mL], with both treatments yielding average daily intakes of ~50 mg/kg/day. Beginning on postnatal day 75, all mice were trained on a radial arm maze-delayed non-matching-to-place pattern separation task. Euploid mice performed significantly better over training than Ts65Dn mice, including better performance at each of the three separations. EGCG did not significantly alleviate the pattern separation deficits in Ts65Dn mice. After the behavioral testing commenced, animals were given ad libitum food access for five days, received a 100mg/kg injection of BrdU, and were perfused two hours later. Coronal sections through the dorsal hippocampus were processed for BrdU labeling, and cells were manually counted throughout the subgranular zone of the dentate gyrus. The euploid controls had significantly more BrdU labeled cells than Ts65Dn mice, however, EGCG does not appear to increase proliferation of the hippocampal neuroprogenitor cells. This is the first report of deficits in Ts65Dn mice on a pattern separation task. To the extent that pattern separation depends on the functional involvement of newly generated neurons in an adult dentate gyrus, this approach in Ts65Dn mice may help identify more targeted pharmacotherapies for cognitive deficits in individuals with DS.