ED Sciences de la Vie et de la Santé
Fear memory engrams in the Anterior Cingulate Cortex
by Sarra GUEHAIRIA (Institut Interdisciplinaire de Neurosciences)
The defense will take place at 14h00 - Amphi BROCA Université de Bordeaux IINS - UMR 5297 - Bât. Neurocampus 146 rue Léo Saignat
in front of the jury composed of
- Philipe ISOPE - Directeur de recherche - CNRS strasbourg - Rapporteur
- Daniela POPA - Directrice de recherche - Institut de biologie de l'ENS (IBENS) - Rapporteur
- Julien COURTIN - Chargé de recherche - Équipe Apprentissage associatif des circuits neuronaux (Herry), Neurocentre-Magendie,INSERM, Bordeaux, France - Examinateur
- Aude PANATIER - Directrice de recherche - Équipe Relations neurone-glie (Oliet), Neurocentre-Magendie,INSERM, Bordeaux, France - Examinateur
Memory allows animals and humans to use past experiences to direct future actions and adapt flexibly in changing environments. Modern neuroscience increasingly sees memory as a predictive system: one that detects patterns in experiences to anticipate upcoming events and prepare appropriate responses. This predictive ability depends heavily on the brain's capacity to encode temporal structure, which not only shows what happened but also indicates when it is expected to occur. Thus, temporal information is a crucial part of memory that influences behavior across various species. Fear learning is a type of memory that provides a powerful framework for exploring temporal prediction. Contextual conditioning teaches animals that a context signals an upcoming unconditioned stimulus (US), such as a foot shock. Crucially, they learn the timing between multiple US presentations. This temporal map greatly influences the timing and strength of defensive responses like freezing. During recall, some animals freeze immediately upon entering the context, indicating a contextual approach, while others freeze later, aligning with the predicted US timing, suggesting a temporal strategy. These behavioral differences highlight the importance of temporal processing in fear memory and reveal underlying diversity in neural computations. Fear memories are stored in dispersed neural groups, also called "memory engrams," which activate after the experience and during recall. Recall reopens the reconsolidation window when memories can be updated, strengthened, or erased. This process is especially important for cortical regions like the anterior cingulate cortex, which are repeatedly involved in expressing remote memory. Recent research shows that cortical engrams change over days and months, and each recall can either stabilize or alter their activation patterns. The anterior cingulate cortex is part of the medial prefrontal cortex and serves as a key hub for integration tasks. However, its exact role in predicting dangers during natural behaviors is still not fully understood. Another open question is the role of US‑responsive neurons in the ACC. This group of neurons (US1 responsive cells) shows strong responses specifically during the aversive event and may form part of the cortical component of the fear engram. Their reactivation during recent and remote recall could reveal how aversive memory traces change over time and whether recall episodes affect their stability. The dynamics of US‑responsive ensembles in the ACC and their relationship to behavioral variability remain largely unexplored. Finally, there are notable individual differences among animals in the timing of freezing during recall; some show early, context-driven freezing, while others display delayed, temporally aligned freezing. Whether these behavioral differences relate to distinct patterns of ACC activity or US-ensemble reactivation remains unknown. My thesis addresses these gaps by combining longitudinal calcium imaging with detailed behavioral quantification to explore how the ACC encodes and updates temporal predictions of threat. It also examines how recall history influences ACC population dynamics, ACC-behavior coupling, and the stability of US-responsive ensembles by comparing animals that had a recent recall session in the shock-associated context to those that did not. Together, these analyses aim to clarify the role of the ACC in temporal prediction, fear regulation, and the long-term organization of aversive memory engrams.
ED Sciences Physiques et de l'Ingénieur
Susceptibility modeling of electronic component chains to Intentional Electromagnetic Interferences
by Antoine DUGUET (Laboratoire de l'Intégration du Matériau au Système)
The defense will take place at 9h45 - Amphithéâtre Jean-Paul Dom Laboratoire IMS, 351 Cours de la libération Bâtiment A31 33400 Talence
in front of the jury composed of
- Tristan DUBOIS - Maître de conférences - Université de Bordeaux - Directeur de these
- Guillaume ANDRIEU - Maître de conférences - Université de Limoges - Rapporteur
- Fabrice CAIGNET - Professeur des universités - Université de Toulouse - Rapporteur
- José LOPES ESTEVES - Ingénieur de recherche - ANSSI - Examinateur
- Geneviève DUCHAMP - Professeure émérite - Université de Bordeaux - CoDirecteur de these
- Jean-Baptiste BéGUERET - Professeur des universités - Université de Bordeaux - Examinateur
The Internet of Things (IoT) can be defined as a network of connected objects integrating sensors, actuators, and wireless interfaces. This architecture enables them to interact with their physical environment and exchange data with the digital world through an Internet connection. Today, these electronic objects are predominantly composed of commercial off-the-shelf (COTS) components, making them vulnerable to Intentional Electromagnetic Interference (IEMI) sources that are increasingly accessible to the general public. Previous studies have demonstrated that such interference sources can induce various effects on an electronic target, ranging from performance degradation to behavioral manipulation akin to hacking. To date, the study of this electromagnetic attack scenario has been primarily limited to experimental approaches in laboratory conditions. However, to deploy such studies more broadly, a simulation-based approach to electronic object susceptibility is preferred. This thesis work proposes several key elements for implementing a methodology to model the susceptibility of electronic objects to IEMI. Initially, the susceptibility of an electronic object specifically developed for this study is explored through radiated and conducted interference signal injections. Subsequently, the susceptibility of an analog electronic component chain is characterized and modeled using the ICIM-CI (Integrated Circuits Immunity Model – Conducted Immunity) and a field-to-trace coupling model based on transmission line theory. Throughout this study, the ICIM-CI model—traditionally employed for modeling component susceptibility to purely sinusoidal signals—has been extended to amplitude-modulated signals resembling the waveforms emitted by IEMI sources.