When a speaker produces an utterance, a variety of choices and decisions have to be made. The speaker needs to decide what (s)he wants to talk about (e.g. my Sunday morning or my Monday morning occupations?), then select the words that are most appropriate to express the message (e.g. should I say table or desk?), and also to retrieve certain words that are constrained by the retrieval of other words (e.g. when speaking French, will the next article be la or le?) These examples show that lexical selection is constrained to variable degrees and by variable sources of information: the language production system appears to take a variety of decisions on the basis of different mechanisms. One shared feature across current cognitive models describing language production is their primary interest in how linguistic information is represented, and how activation circulates in the representational system. This focus leaves outside the scope of the research an exhaustive investigation of the cognitive principles underlying activation and selection. This has happened despite the fact that the processes of response selection in other tasks (e.g. two-alternative forced-choice) have received a lot of attention, and have lead to a variety of articulated proposals. Are response selection and lexical selection similar events, or does language production rely on specific mechanisms to select information? We propose an interdisciplinary investigation to address this question. Our project is motivated by the lack of an articulated integration of language production theories with cognitive descriptions of response selection. We will combine a psycholinguistic framework with models of optimal decision making, and with an information-theoretical approach for computational modeling. The experimental evidence will come from behavioral and electrophyisiological experiments. Our goal is to understand better the sequence of cognitive events by which a word is activated and selected during language production.

The skin is a complex tissue, capable of controlling infections through efficient immune responses while maintaining its own integrity. Dendritic cells (DCs) constitute sentinels of the immune system that contribute to elicit and control skin immune responses. Using innovative mouse models (LangEGFP and LangEGFPDTR), we studied the Langerhans cells (LCs) located in the epidermis, we discovered a new DC subset in the dermis (Poulin, 2007) and described five different DC subsets in the skin (LCs, CD207+ CD103-, CD207+ CD103+, CD207- CD11b-, and CD207- CD11b+ dermal DCs (DDCs)). We hypothesized that this phenotypic heterogeneity is associated with functional differences. Accordingly, the main objective of the present project is to understand the biological role of these different skin DC subsets by focussing mainly on tolerance mechanisms such as clonal deletion of self-specific T cells and induction of regulatory T cells (Treg). Four lines of research will be developed : 1. We will take advantage of our ability to discriminate homogeneous skin DC subsets to develop a comparative transcriptomic analysis of such five skin DC subsets. Such approach should provide us important hints on the functional specialization of those various skin DC subsets. 2. Using a transgenic mice where a model antigen (OVA) is expressed in keratinocytes (K5-mOVA), we have showed that the CD207+ CD103+ DDC subset was the only one capable of cross-presenting such self-antigen to OVA-specific OT-I CD8+ T cells (Henri, 2009). One aim of the present proposal is thus to elucidate the mechanism underlying the capacity of the CD207+ CD103+ DDCs to cross-present antigens. More specifically, we will determine whether this cross-presenting capacity is due to their enhanced capacity to encounter OT-I CD8+ T cells via XCR1 chemokine receptor expression. Moreover, the K5-mOVA model should allow us to further understand the mechanism of peripheral tolerance that is based on clonal deletion. 3. Treg cells could be induced by DCs producing retinoic acid (RA) through the oxidation of retinaldehyde by retinaldehyde dehydrogenase (ALDH). We have recently showed that in the skin, only the CD207- CD11b+ DDC was expressing the ALDH at the protein level and was able to induce Treg cells in vitro (Guilliams, in press). We hypothesized that this important property also extends in vivo. To determine if the high frequency of Treg cells found in the skin is the consequence of the inductive function of the RA-producing CD207- CD11b+ skin DDC subset, we will develop an innovative knock-in mouse model allowing to track and ablate RA production by CD207- CD11b+ DCs in vivo. 4. Finally, we will attempt to visualize the interactions of DCs with T cells using confocal and intravital microscopy. To track DDCs both in the dermis and in the CLNs, we intend to use mice that co-express the Cre recombinase under the control of either the CD11c or the Langerin gene and a LoxP-STOP-LoxP-RFP red reporter cassette within the Rosa locus. Adoptive transfer of green (GFP+) OT-I T cells into Langerin-Cre x LoxP-STOP-LoxP-RFP x K5-mOVA recipient will give us the unique possibility to determine by confocal and 2 photon microscopy whether the GFP+ OT-I T cells colocalize with the cross-presenting RFPbright CD207+ DDCs that have captured OVA in the skin. If the expression of XCR1 chemokine receptor is of importance in such cross-talk, developing a K5-mOVA x Lang-Cre x LoxP-STOP-LoxP-RFP x B6.129P2-Xcr1/J mouse line will allow us to visualize whether XCR1-deficient CD207+ DDCs have lost their ability to contact CD8+ T cells. In conclusion, this research proposal aims to unravel if the phenotypic diversity of the skin DCs is associated with specialized functions in vivo. A better understanding of the mechanisms of tolerance contributing to maintain skin homeostasis will allow the selection of the most optimal DC target for therapy intended to intensify or dampen skin immune responses.

Le but du projet MEMOIRE est de développer de nouvelles mémoires à grille flottante à nanocristaux en Si ou Ge visant à remplacer les mémoires flash actuelles. Cette nouvelle technologie, qui a acquis une grande maturité ces dernières années, se confronte aujourd’hui encore à des problèmes de reproductibilité, de compréhension des mécanismes de stockage des charges et de développement des composants, qui doivent être résolus avant son introduction dans l’industrie. MEMOIRE est un projet interdisciplinaire alliant l’engineering des composants à la physique pure, dans lequel les principes physiques fondamentaux sont utilisés pour lever des verrous technologiques à la frontière des applications. Ce projet permet de créer une chaîne complète de compétences de la manipulation / élaboration des NC, à la modélisation, fabrication et caractérisation électrique des transistors et des mémoires. Le principal challenge du projet est d’obtenir le contrôle parfait des structures élémentaires fabriquées afin de permettre la compréhension des mécanismes électriques locaux et la modélisation appropriée des composants. Le travail sera réparti en quatre tâches : 1) fabrication de la grille flottante à NC et des composants; 2) modélisation de la nanostructuration du substrat et auto-assemblage des NC ; 3) modélisation des dispositifs ; 4) fabrication et caractérisation électrique des dispositifs élémentaires. Grâce à la complémentarité des partenaires, le projet permettra des avancées majeures tant dans la compréhension des mécanismes de base que dans le développement de nouvelles architectures de composants. Le consortium s’attachera aussi particulièrement au transfert technologique des procédés mis en oeuvre en laboratoire vers le partenaire industriel.