Bravo à nos lauréats !
Claire Altersitz
- Projet : Use of polygenic mouse models to identify diagnostic and predictive biomarkers of treatment response in depressive disorders
- Encadrement : Stéphane Jamain
- Unité : INSERM U955 – Université Paris-Est Créteil (UPEC)
- École doctorale : 402 – Life sciences and Health
- Recherche doctorale : Current rodent models of DD fail to recapitulate its complex etiology. To address this, we generated two mouse lines, H-TST and H-FST, with a complex genetic background selectively bred for their high immobility in response to stress, respectively the Tail Suspension Test (TST) or the Forced Swim Test (FST). Both were passively coping with TST and FST, but only H-TST mice showed anxiety-like, anhedonia-like behaviours and sleep disturbances. The two lines also responded differently to antidepressants, H-FST mice improving well with serotonin reuptake inhibitors where H-TST mice did not, thus modelling two subgroups of people with a DD. My doctoral work combines behavioural phenotyping, synaptic proteins quantification and gene expression analyses to decipher their underlying neurobiology.
To date, in-depth behavioural characterisation of H-TST mice showed they presented increased threat reactivity and decreased sociability, two transdiagnostic traits in some people with a psychiatric disorder. At the molecular level, PFC synaptic protein quantification demonstrated the modelled subgroups are underpinned by different alterations, as H-TST mice had increased excitatory synapses, and H-FST mice showed decreased excitatory but increased inhibitory synapses. Concordantly, PFC transcriptome analysis identified Grm7, encoding the metabotropic glutamate receptor 7, as overexpressed in H-TST but underexpressed in H-FST mice. This analysis also identified a 10-gene signature that stratified people with a DD based on their anxiety status, and predicted duloxetine response in an independent cohort.
Altogether, my doctoral work contributes to precision psychiatry by providing a novel gene expression-based signature to predict antidepressants efficacy in people with DD and increased threat reactivity and anhedonia. Therefore, testing innovative therapeutic approaches on these mice will accelerate the translation of targeted treatments tailored to patient endophenotypes
- Découvrez son profil ici
Valeria Finelli
- Projet : fMRI-based neurofeedback in drug-resistant auditory hallucinations
- Encadrement : Renaud Jardri, Paul Allen
- Unité : U1172 – Lille Neuroscience and Cognition – Université de Lille
- École doctorale : Ecole Graduée Biologie Sante de Lille
- Recherche doctorale :Auditory verbal hallucinations (AVHs) are among the most disabling symptoms of schizophrenia, affecting a substantial proportion of patients and frequently persisting despite optimized treatments. This therapeutic gap highlights the need for research into innovative interventions. Recently, there has been increasing interest in real-time (rt)-fMRI neurofeedback as a novel therapeutic method. In addition, recent developments in symptom capture methods now enable the identification of AVH episodes with high spatio-temporal precision.This study investigates whether patients with treatment-resistant AVHs can learn to regulate AVH-related brain states through an individualized rt-fMRI neurofeedback approach. The project combines multivariate pattern analysis (MVPA) with rt-fMRI neurofeedback, using a previously optimized between-subject classifier capable of decoding AVH-related brain states, which will be further personalized by integrating each participant’s own activation patterns. The efficacy of this image-guided intervention will be assessed in a randomized, double-blind, sham-controlled clinical trial including 90 patients with schizophrenia and refractory AVHs.To date, the neurofeedback protocol has been finalized and evaluated in a feasibility study involving 20 healthy participants trained to modulate an auditory verbal imagery-related brain state, providing initial support for the feasibility of MVPA-based rt-fMRI neurofeedback. In addition, the protocol has been successfully implemented in a first patient with schizophrenia and refractory AVHs, who completed the procedure without difficulty, supporting its clinical feasibility.By enabling personalized targeting of symptom-related brain states, this project represents a major step toward the development of fMRI-guided therapeutic interventions for treatment-resistant psychiatric symptoms and contributes to the advancement of precision psychiatry approaches in schizophrenia
- Découvrez son profil ici
Sina Kling
- Projet : Decoding oculomotor signals in human visual cortex: Towards deployable gaze tracking tools for clinical neuroimaging
- Encadrement : Guillaume S Masson / Martin Szinte
- Unité : Institut de Neurosciences de la Timone – Aix-Marseille Université
- Recherche doctorale : Every glance shifts the image on the retina, yet perception stays stable. The brain achieves this by combining visual input with information about its own eye movements, but how oculomotor signals are represented within visual cortex, and how they interact with visual maps, is still poorly understood. This gap is clinically meaningful: abnormal saccades and disorganised gaze during free viewing are among the most reproducible behavioral findings in schizophrenia, tied to altered oculomotor-related cortex including the frontal eye fields, pointing to disrupted cortical integration as a plausible origin. My thesis targets this integration directly. Because oculomotor activity is normally entangled with the visual response that inevitably dominates fMRI recordings during visually guided experiments, I pursued two routes to disentangle these signals. Computationally, population receptive field modelling decomposes mixed visuo-oculomotor activity during visually-guided tasks to isolate saccade-specific signals in higher visual areas. Experimentally, recording eye movements in darkness removes the visual confound. This route needed a separate source of ground-truth gaze that is not camera-based eye-tracking: a fine-tuned DeepMReye, a deep-learning model recovering eye position from the fMRI volume itself. My work yields three outputs: an ongoing computational framework mapping saccade-specific signals near the precentral and intraparietal sulci, illuminating the cortical basis of perceptual stability; a fine-tuned and validated eyes-open and eyes-closed DeepMReye pipeline recovering gaze from any fMRI scan lacking eye-tracking hardware; and « EyePrep », a preprocessing and quality-control tool for gaze data. These converge on one aim: deploying camera-free gaze measurement across FrenchMinds sites, with EyePrep standardising and quality-checking the decoded gaze, to derive oculomotor biomarkers from movie-watching in schizophrenia and other disorders
- Découvrez son profil ici