Vous trouverez dans cette rubrique les offres de stage, de thèses, de post doctorats, de CDD et CDI à pourvoir dans les laboratoires partenaires du réseau.

Liste des annonces

M2 Internship + PhD thesis IRAMIS Lab

case aims at assessing a learning procedure. It is useful
for instance to compare models and to be able to tell if one approach is significantly better than another one or if the difference is due to chance. How to properly perform such statistical assessment is still a matter of active research. Indeed, one needs to account for multiple sources of variance, not only coming from the testing set but also from the training set, from hyperparameter choices, from random seeds… How to do this remains an open problem even though some interesting directions have been proposed (Bouthillier et al, 2021).
Therefore, this project aims at studying statistical methods for assessing learning procedures for neuroimaging data analysis. It will include a back and forth between experimental aspects and theoretical aspects (new experimental results potentially leading to modification of statistical procedures which would in turn lead to new experiments).
To that purpose, the experiments will rely on ClinicaDL (https://clinicadl.readthedocs.io/), a software platform for reproducible deep learning in neuroimaging.
More specifically, the project will include the following aspects:
- Theoretical analysis of statistical approaches for assessing learning procedures
- Overview of existing practices and litterature
- Performing experiments on standard benchmark data
- Performing various experiments with ClinicaDL across different neuroimaging tasks (classification, regression, segmentation…) and various datasets (starting with datasets on Alzheimer’s disease and then moving to include datasets with other pathologies in order to: i) measure the different sources of variances of learning procedures and see if the conclusions made for general ML benchmarks hold in our context; ii) assessing if the existing proposed statistical approaches are adapted across our tasks of interest (to that purpose, we will compare them to more comprehensive but computationally heavy
- If needed, enrich ClinicaDL with new models and datasets if the ones currently available are not sufficient
- Based on experimental results, propose revised statistical procedures that are adapted depending on the context (task, number of samples, type of network, dataset…)
- Implement the statistical procedures in ClinicaDL
This is an ambitious project. Only a fraction of the aforementioned work can be performed within the timeframe of an internship. This is why we propose to potentially continue the internship by a PhD.
- O. Colliot, E. Thibeau-Sutre, Brianceau C, and N. Burgos, “Reproducibility in medical image computing: what is it and how is it assessed?,” 2024. https://openreview.net/forum?id=3fIXW9mFfn
- X. Bouthillier, P. Delaunay, M. Bronzi, A. Trofimov, B. Nichyporuk, J. Szeto, N. Mohammadi Sepahvand, E.
Raff, K. Madan, V. Voleti et al., “Accounting for variance in machine learning benchmarks,” Proceedings of
Machine Learning and Systems, vol. 3, pp. 747–769, 2021. https://arxiv.org/abs/2103.03098
- G. Varoquaux and O. Colliot, “Evaluating machine learning models and their diagnostic value,” To appear in Machine Learning for Brain Disorders, Springer, HAL preprint, vol. hal-03682454, 2023 https://hal.archivesouvertes. fr/hal-03682454

ingénieur.e de recherche EPSM Marne

L’ingénieur(e) de recherche assure le développement et la mise en oeuvre des méthodes d'acquisition des données expérimentales utilisées dans le cadre des projets de recherche du service, assure le traitement de ces données et participe à leur valorisation scientifique.
 Mise au point des outils techniques et des dispositifs utilisés dans les recherches (systèmes de
recueil de données, méthodes d'analyse en imagerie et en électrophysiologie) et contrôle de la
qualité du recueil et du traitement des données
 Acquisition et traitement de données IRM anatomiques et fonctionnelles, physiologiques,
comportementales, cliniques, et toute donnée issue des protocoles de recherche du service
 Travail en collaboration avec les équipes de recherche hospitalières et universitaires pour définir
les méthodologies adaptées aux objectifs de recherche
 Présentation, diffusion, et valorisation des aspects méthodologiques des travaux de recherche
 Conception d’une documentation technique, de procédures, protocoles, et modes opératoires
constitutifs des méthodologies de recherche du service
 Veille technologique avec des actions de formation et de conseil sur les méthodologies de
recherche en neurosciences cognitives et neuroimagerie

Post-doctorant.e en Imagerie hyperspectrale computationnelle rapide

We are actively seeking a postdoctoral fellow or an engineer to spearhead the development of a high-speed computational imaging system tailored for hyperspectral imaging. This cutting-edge camera need to be developed to detect Protoporphyrin fluorescence signal, specifically in the context of glioma resection. The role, funded by ANR, will encompass the full spectrum of responsibilities, from conceptualization and simulation to system design, practical implementation, and comprehensive characterization. This exciting opportunity will involve the acquisition of hypercubes from ex-vivo and in-vivo glioma samples directly within the hospital’s operating room setting.
Currently, the total acquisition time is suitable for biopsy samples but proves impractical for in-vivo imaging not to mention the lake of spatial resolution. In response, our research group has developed compressive imaging strategies aimed at reducing the number of patterns required for acquisition, thereby decreasing the total acquisition time. This innovative approach involves deep neural networks to reconstruct the hyperspectral cube corresponding to the observed scene. Over time, we have proposed various deep network architectures [4], each designed to optimize the hyperspectral reconstruction process. These networks have proven to be very effective in denoising experimental data obtained with our hyperspectral imaging device. While this represents a significant advancement, it remains insufficient for clinical in-vivo applications. Hardware development are needed to enhance the speed by developing a second setup.
- the taking charge of the existing optical setup and making improvements
- the development of the second optical setup
- the development of the instrumentation control software (based on the SPAS package [5])
- the use of advanced reconstruction algorithm based on deep learning (based on the spyrit package[6])
- the characterization of the imaging device (e.g., sensitivity, spatial and spectral resolution).
More precisely, the project comprises two main concurrent tasks
1 - Optical Setup Enhancement and Biopsy Imaging : The first task involves taking charge of the existing optical setup and making improvements, both in software and hardware where necessary. The upgraded setup will be used to acquire hyperspectral data cubes from biopsies within the operating room of the hospital. These acquired cubes will be analyzed and compared to the anapathological analysis. This comparison aims to distinguish tumor cells from healthy ones, providing valuable insights for clinical applications.
2- Development of a 2D Sensor Spectrometer : The second task focuses on the creation of an entirely new optical setup aimed at accelerating the hyperspectral data acquisition process. The current spectral arm, which employs a commercial spectrometer with a 1D sensor, will be replaced by a custom-made 2D sensor spectrometer. This new configuration will consist of a grating, prism, and camera, enabling the simultaneous capture of the full spectrum of multiple image lines. In contrast to the current device, which acquires lines sequentially using a linear sensor, this advancement is expected to dramatically increase imaging speed. As an example, a 64x64 image can be acquired 64 times faster than with the existing setup. The acquisition software will be updated to handle 2D patterns corresponding to 1D Hadamard patterns that repeat in one direction. The successful candidate will be responsible for constructing the 2D spectrometer, involving the use of a cylindrical lens to compress one spatial dimension and a prism or grating (or a combination of both) to disperse and diffract light onto the 2D sensor. These two tasks represent the core components of our project, with the potential to significantly enhance our hyperspectral imaging capabilities and their practical applications in the medical field.

Chef.fe de projet INSERM du PEPR Santé Numérique

Mission principale :
Vous assurerez la coordination et le suivi du PEPR Santé Numérique. Rattaché au directeur de l’Institut Thématique Technologies pour la Santé, le poste sera à l’interface avec l’équipe de pilotage de cet Institut thématique et la direction en charge des programmes stratégiques.
Ce programme de 60 millions d’euros sur 7 ans, copiloté par l’Inserm et Inria, vise à développer les connaissances et les méthodes de l’acquisition jusqu'à l'exploitation des données par des modèles mathématiques et leur sécurisation et fiabilisation, avec comme domaines prioritaires initiaux les approches multi-échelles dans les maladies cardiovasculaires et neurologiques.

Activités principales :
• Accompagner les différentes instances de pilotage du PEPR et les consortiums
• Suivre les programmes scientifiques ciblés et les appels à projets en lien avec l’ANR
• Organiser les reportings scientifiques et financiers des programmes spécifiques (en lien avec ANR)
• Contribuer aux réflexions stratégiques concernant l’orientation du PEPR
• Coordonner les réunions/retraites de restitution annuelle
• Assister les responsables scientifiques et operationnels pour la mise en place de symposiums/colloques par thème
• Construire et entretenir les outils de communication scientifique autour du programme en lien avec les responsables scientifiques et opérationnels et le chargé de communication PEPR
• Assister les responsables scientifiques et operationnels dans la recherche de financement : l’identification et l’initiation des montages Europe, ANR …
• Assister les responsables scientifiques et operationnels sur les aspects de valorisation en lien avec InsermTransfert et les cellules de valorisation et transfert des organismes partenaires du PEPR.

Spécificité(s) et environnement du poste
Ce poste induit des relations fonctionnelles avec les instituts thématiques et les départements supports de l’Inserm (financier, juridique, RH, SI, communication), Inserm Transfert, ainsi que des relations avec le copilote INRIA, les chercheurs impliqués dans le PEPR, l’ANR, et des partenaires extérieurs (Universités, organismes de recherche, …)

Postdoctoral position in Biomedical Photoacoustic Imaging (M/F)

A postdoctoral fellowship is available at the Laboratory of Biomedical Imaging (Paris, France) in collaboration with Institut Galien Paris Saclay (Orsay, France). The candidate will work on the ANR funded project “Control of the optical Absorption Properties of nanovectors for PHOTOACoustic imaging – CAP-PhotoAc”. More specifically, she/he will image the biodistribution of labeled drug nanovectors at the site of inflammation in 3D and in high spatial resolution with photoacoustic imaging, and will correlate the biodistribution with the therapeutic outcome. The project involves the formulation and characterization of nanoparticles, and longitudinal in vivo experimental studies on murine models to monitor their accumulation and evaluate therapeutic efficiency.
- Description of the postdoctoral topic -
Nanoparticles as drug nanovectors are expected to profoundly change therapeutic treatments in the next 15 years and specifically benefit highly prevalent diseases such as rheumatoid arthritis [1,2]. Compared to conventional drug formulations, drug nanovectors need to overcome several biological barriers to achieve drug delivery at the site of inflammation. It has recently become increasingly evident that the heterogeneity of the NV accumulation profoundly impacts therapeutic response. Therefore, image-based selection of patients with an efficient accumulation for a given NV has been identified as a key point for improving therapeutic outcome. The objective of this postdoctoral project is to image the biodistribution of labeled NVs at the site of inflammation in 3D and in high spatial resolution and to correlate the biodistribution with the therapeutic outcome. The preclinical study will be performed in mice.
Photoacoustic imaging (PAI) is an emerging biomedical imaging modality which provides a molecular contrast and a sensitivity to dyes that can label drug nanovectors. The consortium of the ANR project “Control of the optical Absorption Properties of nanovectors for PHOTOACoustic imaging – CAP-PhotoAc” has recently developed BODIPY dyes that can label nanovectors in the NIR range [3-5]. The label concentration per particles was shown to enable a high molar absorption and an in vivo PAI detectability in healthy mice [3]. More recently the dye concentration per nanovector was even further increased and solid lipid nanoparticles (SLN) carrying a prodrug of dexamethasone (an anti-inflammatory drug) were labeled [5,6]. The labelled SLN were synthetized and characterized in vitro with a calibrated photoacoustic spectrometer [7]. They show very promising properties for a high detectability with a photoacoustic imaging scanner.
The postdoctoral fellow will carry a series of detection and accumulation measurements of labeled SLNs in murine models of arthritis using PAI. She/he will perform experimental studies and data analysis to attain the goal for 3D mapping of the biodistribution of NVs effectively delivered to treat an inflammation, and correlating the detected biodistribution with the treatment efficiency. This highly interdisciplinary project is at the interface between photoacoustic signal and image processing (Laboratoire Imagerie Biomédicale, LIB) and the formulation of drug nanovectors (Institut Galien Paris-Saclay, IGPS). The postdoctoral fellow will join the two research groups (the LIB and the IGPS). She/he will label and formulate NVs according to methods developed at the IGPS. She/he will characterize the labeled NVs with standard methods for the characterization of nanoparticles and will further measure their photoacoustic properties.

Post-doctoral Position – molecular imaging and theranostic agents for dysangiogenic diseases

The successful candidate will develop expertise in radiopharmaceutical development and molecular imaging (PET/CT and SPECT/CT) in the frame of the THERAnanoSTIC ANR project, and more widely in preclinical models of cardiovascular and oncologic dysangiogenic diseases. As part of a collaborative multi-disciplinary team, the postdoctoral associate will lead the radio-theranostic research initiative and will be responsible for:
• Development of novel imaging and therapeutic agents
• Testing lead agents in pre-established and new in vitro and in vivo model systems
• Perform corroborative ex vivo assays (e.g. flow cytometry, immunohistochemistry, gammacounting, autoradiography) assessment
The candidate will have access to state-of-the art lab equipment, departmental and interdepartmental core facilities to ensure success of this research initiative in C2VN and CERIMED, on the Health Campus “La Timone” in Marseille, France. The postdoctoral associate will be encouraged and supported in all aspects of career development and is expected to publish in leading scientific journals, present in national and international conferences, and participate in independent and collaborative grant applications.
Applicants with experience in radiochemistry, biochemistry, preclinical imaging and ex vivo assays
are encouraged to apply. Previous expertise with radiochemistry is not required, however, the applicant must be enthusiastic about learning and mastering these skills.

Developer on a medical image processing platform F/H

The objective of FLI-IAM node is to propose an infrastructure to store, manage and process in-vivo imaging data coming from human or pre-clinical procedures. We contribute to an archiving and management infrastructure of in-vivo images as well as provide solutions to process and manage the acquired data through dedicated software and hardware solutions. In addition, we have built a versatile image analysis and data management solution for in-vivo imaging that will allow the interoperability between distributed production sites and distributed users, heterogeneous and distributed storage solution implementing raw and meta-data indexing.

In this context and within the last years we have collected and maintained different kind of data, using the web-based image database, called Shanoir, and different kind of processing algorithms, using the Virtual Imaging Platform.

The Virtual Imaging Platform (VIP) is a web portal developed at CREATIS for the simulation and processing of massive data in medical imaging. One of the VIP main aims is to provide access to distributed computing resources in a transparent way for the end users. VIP has thus the capacity to manage large and complex workloads (generate, schedule and execute multiple jobs) automatically, while requiring no specific skills from its users. It is VIP developers and administrators that are in charge of making this possible. The VIP instance currently deployed at CREATIS uses mainly the storage and computing resources provided by the EGI e-infrastructure. A growing number of projects with various requirements (sometimes security driven) require access to computing and storage resources (e.g., local clusters, private/public clouds) that are not member of the EGI federation.

Within this context, the recruited developer will work on extending and adapting VIP for the use of such private computing and storage resources. He/She will be under the supervision of the manager of the VIP platform and will interact with the other VIP engineers and the FLI-IAM engineering team. He/She will be hosted at the CREATIS lab (Villeurbanne).
The main objectives of the position are:
• Requirement analysis and design
o Understanding of the current VIP implementation for job management on EGI
o Analysis of requirements and technical solutions for the integration of the new computing resources available
o Choice of the solution to be implemented and specifications
• Software development and testing
o Implementation of the chosen solution(s) within VIP and related dependencies
o Implementation of the associated tests
o Continuous integration (CI)
• Deployment and configuration
o Automation on the deployment and configuration procedure on the targeted infrastructure (ideally with Ansible scripts)

CDD Ingénieur.e developpement de codes de calcul haute performance (HPC)

L'équipe Ginkgo de l'Unité BAOBAB de NeuroSpin, dirigée par Cyril Poupon, recrute un.e ingénieur H/F en CDD en charge du développement de codes de calcul à haute performance (HPC), de visualisation à haute performance et de services web pour l’analyse et la mise à disposition de Big Data individuelles de très haute résolution en IRM de diffusion et en microscopie PLI.

Initiée dans le cadre du Human brain Project (HBP, https://human-brain-project.eu) depuis plus de dix ans, la collaboration entre les deux équipes s’est renforcée au travers de la création de l’Institut franco-allemand AIDAS créé conjointement par le CEA et le FZJ, qui vise notamment au déploiement des techniques de calcul à haute performance et d’intelligence artificielle pour cartographier le cerveau humain à des échelles spatiales jamais atteintes jusqu’à présent. Les équipes Ginkgo et Fiber Architecture s’intéressent plus précisément aux connexions cérébrales et ont développé depuis plus de 2 décennies des expertises en IRM de diffusion, en imagerie optique polarisée (ou PLI) et en reconstruction des fibres (ou tractographie) qui permettent aujourd’hui l’inférence des connexions cérébrales à partir de données d’IRM de diffusion
acquises in ou ex vivo, ou à partir de données microscopiques ex vivo de PLI. Récemment, dans le cadre du projet HBP, l’équipe de NeuroSpin a scanné une pièce anatomique (mise à disposition par l’équipe du Pr Christophe Destrieux de l’unité iBrain de Tours) pendant plus de deux années sur un instrument d’imagerie par résonance magnétique (IRM) doté d’un champ magnétique de 11.7 tesla qui a permis d’obtenir pour la première fois au monde sur un cerveau complet des données d’imagerie anatomique et de diffusion de résolution mésoscopique (~100um). En parallèle, les équipes française et allemande ont scanné en IRM et scannent actuellement en PLI une seconde pièce anatomique qui devrait permettre d’obtenir des
données de résolution microscopique de ces connexions sur l’ensemble du cerveau.
Les deux équipes développent actuellement de nouvelles méthodes d’inférence de la connectivité anatomique (dite également de tractographie) qui reposent sur un modèle de verres de spin permettant une approche globale du problème plus robuste au bruit des images et plus enclin à permettre la reconstruction du problème inverse d’inférence des connexions à partir de l’information locale de leur directionnalité. Elles développent également de nouveaux modèles locaux des processus biophysiques en jeu en IRM de diffusion et en PLI qui permettent d’estimer localement des paramètres caractéristiques de la microstructure du tissu cérébral.

Inserm Chaire Neurotechnologies – Signal processing

The project is to develop new methodologies for processing electrophysiological brain signals (EEG, MEG, iEEG, ECoG, LFP, etc.) to i) develop physiological, pathophysiological and theoretical models of neural interactions and anatomo-functional connectivity, ii) to investigate the relationships between clinical evaluation and metrics of neural dysfunctions to identify biomarker(s) for accurate patient stratification and develop data-driven predictive models of functional outcomes and therapeutic responses and iii) to design assessment and/or therapeutic devices that can be easily used in clinical routine, transportable, usable by patients easily and with secure remote monitoring. On-line and transportable new approaches to optimize personalized neuropathological evaluation and care, that can be used in or out of the hospital, are central to LIB research. One of the best illustrations in this field in LIB is the development of portable device to collect in-ear EEG and mobile app interfaces that enables out hospital EEG monitoring. The recruited researcher will have strong expertise in signal processing of brain signals, have a high-quality scientific publication record and have demonstrated evidence in working in a multi-disciplinary team. In addition, experience in leadership of research groups and experience in the strategic planning, acquisition and implementation of R&D projects are desirable in this context. Lastly, good knowledge in deep learning would be appreciated.

The recruited researcher will reinforce teaching in courses related to computer sciences and signal processing within the bachelor’s and master’s level courses in computing and robotic automation at SU. The undergraduate courses provide an initial introduction to data science and to key approaches in artificial intelligence. Master’s courses are more specialized and organized around a dedicated cursus. The recruited researcher will be particularly called upon to strengthen courses related to advanced image and signal processing within the Master’s degrees in “Computer Science” or “Engineering of Intelligent Systems” or “Robotic Automation”. The recruited researcher will be encouraged to introduce practical laboratory instruction based on his/her research into all levels of training at SU and to contribute to the international program Computer Science Master-level.

More details following the link : https://eva3-accueil.inserm.fr/sites/eva/chaires/2023/session2/Documents/Job_profil_%20Chair_2023_U1146_Neurotechnologies_SU.pdf

Assistant ingénieur en expérimentation et instrumentation biologiques

Missions L’Assistant-e ingénieur-e rejoindra l’équipe de la Plateforme d’Imageries du Vivant pour assurer les missions principales suivantes :
1/ réaliser et adapter les protocoles d'imagerie échographique et de bioluminescence sur des modèles murins de pathologie.
2/ participer au fonctionnement général de la plateforme et de son animalerie.

Activités principales :
 Réalisation des protocoles : manipulation des animaux, anesthésie, échographie, Imagerie optique, analyse des images et pose de mesures.
 Adaptation des protocoles d’imagerie : A partir de protocoles de référence, adapter les conditions d’acquisition aux besoins spécifiques d’un projet. Analyser les résultats pour valider cette adaptation.
 Gestion de projets : réception des demandes, participation à la définition des protocoles, suivi des projets, remise des résultats.
 Compléter les informations de projet et d’utilisation machine pour le suivi de la qualité.
 Participation au fonctionnement de l’animalerie dédiée de la plateforme.
 Gestion des stocks et des commandes.

Activités associées
 Activités spécifiques ou transverses :
 Participation à la politique qualité de la plateforme.
 Tutorat d’apprentis ou de stagiaire (1/an).
 Coordination fonctionnelle de projet possible.