Ongoing doctoral projects in the framework of BoostUrCAreer programme

ACt4Autism: Action detection for improving the diagnosis of autism

Supervisors:

  • Research Director François Brémond, INRIA (French National Institute for computer science and applied mathematics) & CoBTEK (Laboratory of Cognition Behaviour Technology)
  • Doctor of Medecine Susanne Thümmler, Laboratory of Cognition Behaviour Technology & CRA of CHU-Lenval (Autism ressources Centre of the CHU-Lenval Children's Hospital of Nice)

International partners: Doctor Jean-Marc Odobez, IDIAP Research Institute, affiliated to the EPFL (Ecole polytechnique fédérale de Lausanne)

Presentation of the PhD topic: 

Deep Learning in computer vision, and in particular for Action Detection, is an effective solution for studying human behaviors of large population, and could be applied to children with autism. It allows capturing, in a non-intrusive and continuous way over time, behavioral patterns. Action detection from live video streams is an important task for monitoring patients, building robots for assisted living and other healthcare applications. Although several approaches, including Deep Convolutional Neural Networks (CNNs), have significantly improved performance on action classification, they still struggle to achieve precise spatio-temporal action localization in untrimmed video streams.

The PhD candidate involved in this project will design novel algorithms for detecting actions, taking advantage of the latest research in Deep Learning. These algorithms will be validated on various international video benchmarks and on a new video database on autism spectrum disorders (ASD) and be published in most prestigious conferences (e.g. CVPR). The early detection of ASD is a crucial issue because it makes it possible to set up intensive and early appropriate care management when certain developmental processes can still be modified.

Asthair: Active tranSporTation, HeAlth and envIRonment

Supervisors:

  • Professor Anne Vuillemin, LAMHESS (Laboratory of sport science),
  • Doctor Gilles Maignant, RETINES lab.

International partners: Senior Lecturer Audrey de Nazelle, Centre for Environmental Policy, Imperial College London

Presentation of the PhD topic: 

Air pollution, physical activity, road traffic injuries are important determinants of health that are affected by transportation patterns. Studies have demonstrated the potential for increased walking and cycling to benefit population health and the environment. The role of city planning and design in promoting population health is increasingly recognized as an essential and promising solution. To make such benefits apparent to decision makers and stakeholders, and further ensure success of such solutions, more work is needed in developing health impact modeling tools which address in a robust manner real world policies and conditions and integrate a variety of impacts.

The ASTHAIR PhD project aims at developing health impact models of proposed urban changes which consider multiple impacts, including co-benefits and trade-offs, integrates advanced knowledge on current activity patterns and other baseline conditions, and includes a framework for effectively communicating findings as feedback to stakeholders to ensure successful implementation and uptake. The results of this work should provide innovative solutions to promote and develop active transport. Industrial partners will be involved in the project and are interested in potential transfer.

BiomechanicsCapture: Deep learning based Real-time Biomechanics Capture

Supervisors:

  • Professor Tarek Hamel, I3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Doctor Andrew Comport, I3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Professor Emma Redding, Dance Science Department, Trinity Laban Conservatoire of Music and Dance (United Kingdom).

International partners:

  • Professor Emma Redding, Dance Science Department, Trinity Laban Conservatoire of Music and Dance (United Kingdom).
  • Google, USA
  • Youdome, Monaco

Presentation of the PhD topic: 
Capturing and tracking high-detail human motion in real-time is a hot research topic that is fundamental to a wide range of applications including e-health, sport performance analysis, human-robot interaction, augmented reality and many more. This multidisciplinary doctoral project aims to work across the domains of real-time computer vision, deep learning and bio-mechanics. The aim is to address the problem of acquiring the pose, shape, appearance, motion and dynamics (torques, forces and velocities) of humans in 3D using multi-camera environment in real-time. One of the major challenges in live motion capture is
the problem of dense modelling of non-rigid scenes.

The objective of this doctoral project will be to design an end-to-end approach such that the input to a training network will be the set of images from multiple
cameras observing the scene. The output of the network will be the high detail 3D geometry and dynamics acting on the human body. To this
end we aim to use RGB-D sensor-consistency to train the network in an unsupervised manner such that all images transform correctly to every
other image with minimal error. For the training phase we will use many sensors, however, the use of the network for reconstructing the bio-mechanics will use much fewer sensors (even potentially with a single sensor). Such a low-cost set-up with a single camera could be used by a medical (or sport) practitioner for diagnosis.
 
DTBC: Building a diagnosis tool to detect broncho constrictions

Supervisors:

  • Researcher Benjamin Mauroy, VADER center (Center for Virtual modeling of respiration) & LJAD (Jean Alexandre Dieudonné Laboratory of mathematics),
  • Professor Lisa Giovannini-Chami, Pneumology Department, Lenval Hospital,
  • Reseacher Angelos Mantzaflaris, Aromath (AlgebRa, geOmetry, Modeling and AlgorTHms), INRIA (French National Institute for computer science and applied mathematics).

International partners : Professor Olivier Debeir, Brussels Polytechnic School.

Presentation of the PhD topic : 

Amongst the most frequent lung’s diseases, many induce a shrinking of the bronchi, typically asthma, COPD (“smoker disease”), bronchiolitis in babies, cystic fibrosis, etc. One of the goals of the therapies is to correct those constrictions in order to restore normal air flows within the lung. It is however very difficult to know where the constrictions occur as no direct information can be obtained from routine lung’s explorations. Consequently, many therapeutic responses, such as chest physiotherapy, are empirical and are difficult to validate.

This interdisciplinary PhD thesis aims at giving a scientific basis to this empirical knowledge. The main goal is to build for the first time an artificial intelligence (AI) that will be able to relate data from routine exploration with the localisation of the constrictions.

e-MDR: Development of new antibiotics against clinical multidrug resistant bacteria from untapped marine microorganisms, a chemobiology approach

Supervisors:

  • Researcher Mohamed MEHIRI, ICN (Institute of Chemistry of Nice, UMR CNRS 7272),
  • Researcher Laurent BOYER,  C3M (Mediterranean Center for Molecular Medicine, INSERM U1065),

International partners: Professor Giovanna Cristina VARESE, MUT (MYCOTHECA UNIVERSITATIS TAURINENSIS), University of Turin.

Presentation of the PhD topic: 

Health problems and the quality of life are worldwide issues. The impact of antibiotic resistance on public health is considerable as it is estimated to be the leading cause of global mortality by 2050, resulting in more than 10 million deaths per year. Paradoxically, the pipeline for new antibiotics has experienced a long-term decline since 1987. The renewal of the therapeutic arsenal is therefore crucial in order to limit the impact of antibiotic resistance in the coming years.

Marine microorganisms represent an under-explored source of new natural products which exhibit in situ several biological activities (cytotoxic, antibiotic, antifungal, antifouling, etc.). Marine natural products have often original structures, different from those of the metabolites of the terrestrial environment, and exhibit potent pharmacological activities with novel mechanisms of action. They could therefore be used to address unmet medical needs such as antibiotic resistance.

In this context, the purpose of the e-MDR PhD project is the development of new antibiotics against clinical multidrug resistant bacteria from untapped marine microorganisms.

FED-Bionet:  Federated learning of biomedical data in large-scale networks

Supervisors:

  • Researcher, Marco Lorenzi, INRIA (French National Institute for computer science and applied mathematics),
  • Research Director Barbara Bardoni, IPMC (Molecular and Cellular Pharmacology Institute).

International partners : Doctor Andre Altmann, Centre for Medical Image Computing, University College London.

Presentation of the PhD topic :

This project envisions a novel paradigm for machine learning in healthcare based on the innovative concept of federated learning. Our goal is to exploit the power of modern learning methods at full capacity within the current clinical data scenario. To this end, we will focus on methodological, technical, and translational advances towards the development of a novel generation of federated learning methods for the analysis of private and large-scale multi-centric biomedical data.

This project will provide the fellow with highly competitive skills for securing a position in the tech industry, in particular in startup and companies in the domain of machine learning and artificial intelligence. Furthermore, the strong biomedical application tackled during the PhD project will allow the student to acquire solid competences in biomedical data management and analysis. This aspect may open up important career perspectives in the field of biotech, pharmaceutical, and clinical research.

The project will count on the expertise and collaboration of the partners of the ENIGMA consortium, a worldwide network of clinical centers providing data and expertise in dementia research. The project will also involve a 6 months visit period to the Centre of Medical Image Computing (CMIC) of University College London (UCL).

MEDSCAN:  Numerical solution for a hand-held high-resolution medical scanner

Supervisors:

  • Professor Claire Migliaccio, LEAT (Laboratory of Electronics, Antennas and Telecommunications),
  • Associate Professor Victorita Doelan, LJAD (Jean Alexandre Dieudonné Laboratory of mathematics).

International partners: Peter Barrowclough, Lincoln Agritech Ltd (New Zealand).

Presentation of the PhD topic:

Proton therapy is currently the most advanced radiation therapy available for cancer treatment. Due to its specificities, proton beam can destroy cancer cells without attacking the surrounding healthy tissue. However, the proton beam position and shape must be accurately measured before each radiation since it directly affects the treatment efficiency and the eventual collateral damages. We propose a new calibration approach by developing robust GaN semiconductor detectors alloying to increase the control of the irradiated dose while strongly reducing the system complexity and cost. This innovation may thus drastically improve the proton therapy.

In this context, the fellow will participate at all steps required to elaborate the GaN detectors at CRHEA-CNRS. He/she will benefit of the access to the regional technological platform CRHEATEC in order to develop the different processes of these devices fabrication. In a second period, he/she will characterize the devices directly on the medical site using the proton beam of the IMPT-CAL (Institut Méditerranéen de Proton Thérapie – Centre Antoine Lacassagne, https://www.protontherapie.fr). Subsequently, the fellow will spend 6 months in Professor Wieck's group at the University of Bochum to manufacture complete arrays of detectors but also to develop their interfacing with a commercially available readout circuit based on silicon

MitoIntegrOMICS:  an integrated multi-OMICS approach to increase the diagnostic power for mitochondrial diseases

Supervisors:

Professor Michel Riveill, I3S (Laboratory of Information and communication science of Sophia Antipolis), Research Director Silvia Bottini, MDLab (Medical Data Laboratory), University Côte d'Azur, Professor Véronique Paquis, IRCAN (Institute for Research on Cancer and Aging).


Presentation of the PhD topic:

Mitochondrial diseases (MD) are rare disorders caused by deficiency of the mitochondrial respiratory chain, which provides energy in each cell. MD are caused by alterations (variants) on genes involved in mitochondrial functions. The diagnosis of MD is based on the identification of the disease responsible gene(s), that will allow to be able to offer genetic counseling, prenatal diagnosis, to consider therapeutic approaches and to improve the care of patients. Nowadays, technologies currently used for detecting causal variants is far from complete, ranging from 25 to 50%.

To address these needs our research teams propose to gather three different domains: medical, bioinformatic and machine learning, in order to set up an integrated multi-omics approach to identify novel causal variants. We foresee that this project will contribute to set up new diagnostic tools to reduce the number of patients with a diagnostic stalemate. This study will settle the milestones to transfer the conjoint use of multi-omics technologies from research fields to diagnostic environment.

International partners:

MyDataModels (France), Doctor Claudio Donati, Computational Biology Unit of the Research and Innovation Centre, Fondazione Edmund Mach (Italie).
Move-it:  Motion analysis by Video for gait Evaluation with Innovative Technology

Supervisors:

  • Professor Raphael Zory, LAMHESS (Laboratory of Human Motricity, Expertise, Sport and Health),
  • Researcher Laurent Busé, Aromath (AlgebRa, geOmetry, Modeling and AlgorTHms), INRIA (French National Institute for computer science and applied mathematics).

International partners: Associate professor Katia Turcot, Centre for Interdisciplinary Research in Rehabilitation and Social Integration, Laval University.

Presentation of the PhD topic:

In France, the expenses in physical rehabilitation increased from 7.3 to 8.4 B€ between 2010 and 2015, mainly due to the ageing population, the increase of chronic pathologies such as strokes or Parkinson, and the shortening of the hospitalization time. 70% of the activity of rehabilitation institutions is about gait (first step for the regain of autonomy). Accurate reliable knowledge of gait characteristics at a given time, and even more importantly, monitoring and evaluating them over time, may enable early diagnosis of diseases and their complications and help to find the best treatment. Three-dimensional motion analysis is the gold standard for clinical gait analysis (CGA), particularly in the presence of pathologies that hamper walking. Today, less than 1% of the patients benefit from CGA.

The main objective of this project is to develop a method based on an innovative low-cost motion analysis system and machine learning, enabling an accurate quantification of gait deviation parameters during functional tests, including spatiotemporal and full-body kinematic parameters.

MuExC3Po: Multimodal Functional Exploration of Cerebellar Cognitive Cues and Particularities in Neurology

Supervisors:

  • Professor Olivier Meste, I3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Medical Doctor Marie-Noële Magnié-Mauro, Neuroscience Department, CHU Nice (University Hospital of Nice).

International partners : Associate Professor Roberto Sassi, Biomedical Image and Signal Processing Laboratory, University of Milan.

Presentation of the PhD topic : 

The aim of this PhD thesis project is to improve, through interdisciplinary collaboration, our knowledge on the role of the cerebellum, especially in its functional asymmetries during cognitive, proprioceptive and motor processes. Multimodal functional explorations will be carried out through ECeG (ElectroCerebellarGrams) recordings coupled to f-MRI (functional Magnetic Resonance Imaging) or f-NIRS (functional Near Infrared Spectroscopy), concurrently with the development of ad-hoc signal processing methods.

Experiments in neuro-psychology will build upon interactive devices (as our current patent-pending tablet-based EEG coupled application) to fine tune the study and detection of cognitive and motor skills issues in different targeted populations (children with cognitive or motor particularities, patients with cerebellar syndromes …). With a motivation towards m-health (mobile e-health), this project will develop novel ways to help in the process of recovery/improvement by investigating synergies in complementary techniques like neuro-feedback/neuro-training with alternative interactive devices like Virtual Reality helmets or Wii-like motion sensitive remote sensors coupled with EEG/ECeG/f-NIRS recordings.

Patched-e-chembiol: Inhibition of Patched chemotherapy resistance activity: biocomputational, chemical and cellular approaches

Supervisors:

  • Researcher Isabelle Mus-Veteau, IPMC (Molecular and Cellular Pharmacology Institute)
  • Associate Professor Stéphane Azoulay, ICN (Institute of Chemistry of Nice).

International partners : Associate Professor Paolo Ruggerone,  Department of Physics, University of Cagliari.

Presentation of the PhD topic : 

Cancer drug resistance is a major problem of chemotherapy nowadays. Our team recently identified the Hedgehog receptor Patched as a drug efflux pump that participates to the resistance of cancer cells to chemotherapy. Thanks to a screening program, Panicein A hydroquinone (PAH), a natural compound purified from a marine sponge, was identified as an inhibitor of drug efflux activity of Patched. The synthesis of PAH allowed us to confirm that PAH increases the cytotoxic effect of several chemotherapeutic agents on melanoma cell lines in vitro and in vivo. The use of PAH in combination with chemotherapy may be a novel and innovative way to circumvent drug resistance, recurrence and metastasis of tumors.

The final objective is to obtain a clinical candidate that could be considered for clinical testing with a Pharma partner.

PRELUDE: Predicting the evolution of drug resistance

Supervisors:

  • Researcher Gianni Liti, IRCAN (Institute for Research on Cancer and Aging),
  • Researcher Agnese Seminara, INPHYNI (Nice Institute of Physics).

International partners : Associate Professor Marco Cosentino Lagomarsino, Physics Department, University of Milan.

Presentation of the PhD topic : 

The emergence of drug resistance is a major health problem that can thwart therapeutic control of a wide spectrum of diseases, from bacterial and viral infections to cancer. Drug resistance are regulated by multiple interacting quantitative trait loci (QTLs) as well as by novel mutations that evolve during the treatment process. Dissecting the genetic mechanisms underlying this phenotypic variation is a major challenge and this problematic apply to many human genetic diseases. Indeed, despite decades of genome wide association studies (GWAS), the genetic variants identified only explain a small fraction of the trait heritability, leaving the open question on whether accurate complex trait prediction can be achieved.

The PRELUDE project aims to understand how drug resistance arises and evolves using bacteria and yeast as genetic systems. To do so, the interdisciplinarity and the experimental and computational approaches using sequencing and large-scale genomic analysis make the project state-of-the-art, and will open endless possibilities in both the academic and the private sector.

Tornados: Targeting ofOncogenic microHNA with small molecules via Diversity Oriented Synthesis : toward newchemotherapies

Supervisors:

International partners: Doctor Roger Estrada Tejedor, IQS (Sarria Institute of chemistry) School of Engineering, University Ramon Llull (Spain).

Presentation of the PhD topic : 

One of the most amazing discoveries of the past decades in the domain of genetic oncology is that cancer is related to alterations of both protein coding genes and non-coding RNAs, such as microRNAs (miRNAs). The purpose of this project is the development of novel small-molecule drugs targeting specific oncogenic miRNAs production via original catalytic and green methodologies according to Diversity Oriented Synthesis.

To do so, the PhD candidate will conduct three concomitant tasks :

Task 1: Synthesis of small molecules via original methodologies according to Diversity Orientated ;
Task 2: Evaluation of the biological activity of the synthesized compounds on oncogenic miRNAs involved in gastric cancers, glioblastoma and colon cancer ; Task 3: Molecular modelling studies.

This is a highly challenging and very promising approach that would open the way for innovative targeted cancer therapy and will therefore guarantee employability in R&D companies or universities.

ThyroSonics: Learning based detection and classification of thyroid nodules from ultrasound images

Supervisors:

  • Researcher Hervé Delingette, INRIA (French National Institute for computer science and applied mathematics),
  • Medical Doctor Charles Raffaelli, CHU Nice (University Hospital of Nice).

International partners: Assistant Professor Guillaume Lajoinie, Physics of fluids group, TechMed center for technical medicine and Mesa+ institute for nanotechnology, University of Twente (Netherlands).

Presentation of the PhD topic: 

The prevalence of thyroid cancer is increasing worldwide, making it the fifth most common cancer among women. Owing to its low cost and high sensitivity, ultrasound imaging is unchallenged in the detection of thyroid nodules. The resulting diagnosis, however, heavily relies on the experience of the clinician and the interpretation is based on relatively subjective criteria.

Given the worldwide shortage of expert sonographers and the increasing prevalence of thyroid cancer, there is a strong need to assist clinicians in their analysis of ultrasound images. The proposed thesis aims at developing software solutions based on Artificial Intelligence and more specifically deep learning neural networks, in order to help sonographers i) to select the most relevant planes of acquisition, ii) to objectively detect thyroid nodules and iii) to classify the nodules based on their malignancy. The originality of the proposed research project compared to the state of the art is twofold. First, it will rely on a close collaboration with the University Hospital of Nice, providing clinical expertise, curation of an extensive imaging database and access to state-of-the-art ultrasound devices with support from various industrial partners. Second, it will exploit potentially three-dimensional ultrasound data but also the time series of the raw radio-frequency signals acquired by the ultrasound probes, which potentially contains more information about the tissues than the classical image modality (B-mode).

Other doctoral projects offered in the framework of BoostUrCAreer programme

AI4Psy: Sparse learning to discover biomarkers of reward circuit-dependent psychiatric diseases

Supervisors:

  • Professor Thomas Lamonerie, IBV (Institute of Biology Valrose),
  • Doctor Michel Barlaud, i3S (Laboratory of Information and communication Science of Sophia Antipolis).

International partners: Associate Professor & Doctor Elyanne M.Ratcliffe, Farncombe Family Digestive Health Research Institute, McMaster University (Canada).

Presentation of the PhD topic:

While susceptibilities to psychiatric diseases can be inherited, catalyzers of these susceptibilities, especially regarding anxiety and mood disorders, are stressful events, particularly those that happen early in life. Although it is clear that early life stress (ELS) is a catalyzer, causal mechanisms are not understood and predictive biomarkers to diagnose, stratify patients and prevent these diseases are lacking. The main reason to that is the difficulty to normalize data from patients with highly heterogeneous genetic background and trauma history. In addition, the complex composition of biological samples such as blood requires powerful analytical methods to highlight quantitative variations as well as advanced mathematical tools to identify reliable indicators. It is thus important to develop models of psychiatric diseases together with statistical methods applied to large sets of biological data to discover predictive or diagnostic parameters associated with these diseases.

This PhD project aims at using AI to identify signatures of risk of psychiatric disorders such as chronic anxiety and depression, that could be directly useful for clinicians. The ability to use AI to process large biological data sets is a highly sought-after skill in academics and in the food and drugs industry.

AI4VI: Pushing the limits of reading performance screening with Artificial Intelligence: Towards large-scale evaluation protocols for the Visually Impaired

Supervisors:

  • Researcher Pierre Kornprobst, INRIA (National Institute for Research in Digital Science and Technology) – Biovision Lab,
  • Professor Jean-Charles Régin, I3S – C&A (Constraints and Application) Lab
  • Reseacher Aurélie Calabrèse, INRIA (National Institute for Research in Digital Science and Technology) – Biovision Lab,

International partners: Professor Gordon E. Legge, Minnesota Laboratory for Low-Vision Research.

Presentation of the PhD topic:

Reading performance evaluation has become one of the essential clinical measures for judging the effectiveness of treatments, surgical procedures, or rehabilitation techniques for low-vision patients. To create accurate reading tests, one needs to design a series of equivalent sentences in terms of linguistics, length and layout.

However, because of their highly constrained nature, these sentences are hard to produce, leading to a very limited number of test versions. Your main mission will be to develop fundamental AI methods for automated natural language production, respecting strict constraints. Your results will have a direct impact on the design of powerful e-health diagnostic and rehabilitation tools that will benefit low-vision patients and their medical care staff. This project requires skills in computer science (algorithms, data structure, combinatorial optimization) and you should have a keen interest in medical applications.

BoostUrTeeth: Numerical modelling of the failURe of restored Teeth

Supervisors:

  • Associate Professor Yannick Tillier, CEMEF (Centre For Material Forming) Mines ParisTech,
  • Associate Professor Nathalie Brulat-Bouchard, University Côte d'Azur & CEMEF (Centre For Material Forming) Mines ParisTech.

International partners: Professor & Doctor Ivo Krejci, Department of Preventive Dental Medicine and Primary Dental Care, University of Geneva (Switzerland).

Presentation of the PhD topic:

A dentist spends as much time fixing defective restorations as dealing with initial tooth decay lesions! This is mainly due to the volumetric contraction of dental composites during the polymerization process. Replacing defective dental fillings costs a lot for the society (about $ 5 billion per year in the United States).

This project is part of a larger one that aims at designing and creating experimental and numerical tools that will be proposed to dental composite manufacturers for the development of longer lasting dental composites. The “BoostURTeeth” project is only focused on the numerical aspect. It aims at developing realistic multiscale 3D finite element models (FEM) in order to numerically evaluate the effects of filler contents and resin properties on their mechanical properties.

The work program has been designed to be as fluid as possible, starting (i) with generating the microscale model to describe all heterogeneities and resin/filler interactions, then (ii) developing the failure model to describe how cracks propagate at the interface (CZM models are usually preferred), (iii) to finish with the macroscale model to study the interfacial stresses increasing between the composite and the tooth during curing.

CellHetMod: Modeling the drug response heterogeneity to cancer drugs by cell-to-cell synchronization and feedback in the cell death signaling cascade

Supervisors:

  • Research Director Madalena Chaves, Biocore, INRIA (French National Institute for computer science and applied mathematics),
  • Associate Researcher Jeremie Roux, IRCAN (Institute for Research on Cancer and Aging).

International partners: Doctor Diego Oyarzun, School of Informatics, School of Biological Sciences, University of Edinburgh.

Presentation of the PhD topic:

Initiation of cell death is a critical cellular decision in tissue homeostasis and cancer emergence. However, substantial variability is observed in tumor cell populations, where a fraction of clonal cells commits to cell death while the other survives, contributing to the reduced efficacies of anticancer therapeutics. This PhD project is among the first to link high-content analyses from dynamic imaging and single-cell multi-omics, with state-of-the-art theoretical and computational methods to provide a global understanding of the origins of tumor cell heterogeneity in response to cancer drugs.

Working at INRIA and CNRS labs, the PhD candidate will develop an interactive numerical simulation platform based on mathematical models of cell signaling pathways, including stochastic components which she/he will develop with our partner during a visiting internship in the Biomolecular Control Group at University of Edinburgh. The PhD candidate will acquire a combined expertise in predictive modeling of heterogeneous single-cell data and dynamical systems, which are the fundamental assets of future research in interdisciplinary projects in academia and pharmaceuticals.

COMS: A multi-sensory environment: joining olfactory perception to musical listening in the context of well-being and performance

Supervisors:

  • Professor Sylvane Faure, LAPCOS (Laboratory of Anthropology and Clinical, Cognitive and Social and Psychology),
  • Professor Serge Antonczak, ICN (Institute of Chemistry of Nice).

International partners: Doctor Thanh Xuan Thi Nguyen, Danang International Institute of Technology, University of Danang (Vietnam).

Presentation of the PhD topic:

In this project, the PhD candidate will study the impact of olfactory and musical stimuli on participants' well-being and cognitive performances, considering individual experiences and culture as moderators. In line with the quest by consumers for naturalness and well-being, local (perfume industry in Grasse) or international companies have shown their interests in such aspects. The PhD candidate will thus perform double blind protocols with both subjective (questionnaires) and objective (blood pressure, heart rate, electrodermal response with the new technology of Cocolab Platform) measurements within a high collaborative framework associating psychologists and chemists of Côte d’Azur (France) and DaNang/HoChiMin (Vietnam) Universities. Therefore, under the supervision of Prof S. Faure, expert in cognitive and clinical neuropsychology and of Prof S. Antonczak, expert in the chemistry of aromas and perfumes, the PhD will have to:

  • propose enhanced protocols based on a revue of literature (model of multisensory integration, well-being and cognitive performance) ;
  • set up the experimental protocols for a Western population (physiological measurement with biopac®, emotional identification with FaceRader® of Noldus®, manage synchronization of odorant’s and music’s diffusion with TheObserverXT®, neuropsychological and psychometrics scales assessment) ;
  • replicate these studies to a non-Western population (Da Nang University, 6-months research stay) ;
  • value the results (publication, congress…) and search for new funding and partnerships.
Decision: Deep Learning for Clinical Score Prediction (towards Computer-Assisted Clinical Decision Support Systems in Psychiatry)

Supervisors:

  • Professor Lionel Fillatre, i3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Professor Nicolas Glaichenhaus, IPMC (Molecular and Cellular Pharmacology Institute).

International partners: Doctor Raquel Iniesta, Departement of Biostatistics and Health Informatics, King's College London.

Presentation of the PhD topic:

Datasets in medicine routinely involve thousands of heterogeneous variables, including clinical, neuroimaging, genomic and proteomic measures. The analysis of these datasets is challenging, especially when the number of measurements exceeds the number of individuals. Deep learning methods have brought breakthroughs in many fields including image recognition, video and sound analyses among others.

The DECISION PhD project aims to develop a novel clinical decision support system for diagnosis, prognosis and personalized treatment in the field of Psychiatry. It is worth noting that, in the European Union, more than 30% of people are affected each year by mental disorders.

The PhD student will process datasets consisting of both biological and clinical variables with a convolutional neural network. Her/his main objectives will be to show that such a deep neural network can make a piecewise linear approximation of the data manifold and that it can exploit this approximation to predict a (clinical) score defined over this manifold. Deep learning architectures are known to act as black boxes. By studying the theoretical properties of a deep architecture for linearizing the data manifold, we expect to make the results explainable.

DeepMedAnalysis: Large scale medical data analysis

Supervisors:

  • Professor Frederic Precioso, I3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Professor Pascal Staccini RETINES Lab - CHU Nice (University Hospital of Nice).

International partners: Professor Doctor Eduardo Alves Do Valle Junior, School of Electrical and Computer Engineering, University of Camerino.

Presentation of the PhD topic:

In the healthcare domain, datasets are geographically (hospital, practitionner’s office, patient him/herself) and timely distributed.

Besides health data collected during medical events, new flows of data are originated from the patient himself (quantified-self) as a record of his/her behavior, environment, life style, etc.

In recent years, we have seen an explosion of successful applications of deep learning in medical domain, including image analysis for diabetese retinopathy analysis, breast cancer detection, cardiological disease classification, Electronic health records analysis, Genomics analysis, etc.

Deep networks designed for these tasks have millions and billions of parameters that require enormous resources in terms of annotated data, huge memory and disk storage space, and computer power to manage them.

Although many open-source implementations leverage the performance of GPU programming, the resources required to learn the right settings for these architectures are considerable which are often not reachable for standard hospitals.

In this PhD, we examine the problem of convergence in a deep network with billions of parameters using several thousand GPU threads distributed within several GPUs not sharing memory, which is usually the case if several machines are assembled to solve a very large problem using a very large network.

DyATOT: Modeling Dynamics of White and Brown Adipose Tissues and Implications for Obesity Treatment

Supervisors:

  • Associate Professor Abderrahmane Habbal, LJAD (Jean Alexandre Dieudonné Laboratory of mathematics) & INRIA (French National Institute for computer science and applied mathematics),
  • Research Director Ez-Zoubir Amri, IBV (Institute of Biology Valrose).

International partners: Professor Pierre-Emmanuel Jabin, Department of Mathematics, Center for Scientific Computation and Mathematical Modeling, University of Maryland.


Presentation of the PhD topic:

According to the WHO, on a global scale, 1.9 billion people were overweight in 2016, comprising 650 million people with obesity. This high prevalence of obesity represents a serious threat to human health and well-being.  People with obesity suffer from stigmatization and a severely compromised quality of life.  Obesity strongly clusters with other comorbidities, in particular type 2 diabetes, arterial hypertension, dyslipidemia and certain cancers.
Obese state and its associated complications have emerged as the leading causes of death in Western countries, associated with estimated health care costs of 81 billion Euros per year in Europe only.

DyATOT is an interdisciplinary project intended to develop mathematical modeling, analysis and simulation of excessive accumulation of fat mass responsible of the progression of obesity and the associated metabolic disorders. The main aim is to establish comprehensive and predictive tools thereby leading to the development of efficient therapies to prevent and/or cure obesity.

The DyATOT program is expected to lead to the modeling of several nutritional, genetic and mechanical mediators responsible of the development of obesity in the functional crosstalk between white and brown adipose tissues. As obesity is considered world-wide pandemic and its incidence is increasing, there is a high potential for economic transfer of the gained expertise and findings, through industrial grants and start-up creation.

EmOdor: Design of odorants targeting emotion and motivation in elderly patients

Supervisors:

  • Professor Jerome Golebiowski, ICN (Institute of Chemistry of Nice)
  • Doctor Renaud David, Memory Resource and Research Center, Memory Resource and Research Centre, CHU Nice (University Hospital of Nice), Research Centre Edmond & Lily SAFRA , CoBTeK (Cognition Behaviour Technology laboratory)

International partners: Adjunct Professor Joel D.Mainland, Monell Chemical Senses Center, University of Pennsylvania

Presentation of the PhD topic:

Can we learn a computer how to smell or feel relaxed upon smelling? What is the impact of smelling on our mood and motivation? The PhD project aims to use machine learning and molecular modeling on properties measured through sensory analysis and psychophysiology experiments on human individuals. The goal is to design odorants to fight depression and anxiety using non-pharmacological approaches.

This research topic gathers two very exciting fields, i.e. numerical modeling and the measure of emotions around the sense of smell. It is associated with the proximity of the city of Grasse (the world capital of perfumes) and the technopole of Sophia-Antipolis, where numerical approaches are central. The research will be supervised by world-experts in chemosensory science and psychiatry. Pr. Golebiowski’s group is a world leader in the numerical modeling of smell and taste (http://chemosim.unice.fr/) and Dr. David’s group is a world leader in psychiatry related to autonomy (https://www.cmrr-nice.fr/?p=en-cobtek-presentation). They both published reference articles in their fields.

The candidate will mostly build numerical models to connect chemical structures to their effect on emotion and motivation on one part and on the odorant receptors on the other part. She/He will also partly oversee experiments because it is always better to master the data one tries to model.

ExtraCell: Characterization and modeling extracellular matrix from digital microscopy slides

Supervisors:

  • Research Director Xavier Descombes, INRIA (French National Institute for computer science and applied mathematics) & I3S (Laboratory of Information and communication Science of Sophia Antipolis),
  • Research Director Ellen Van Obberghen-Schilling, IBV (Institute of Biology Valrose).

International partners: Professor Alin Achim, Department of Electrical & Electronic Engineering, University of Bristol (United Kingdom).

Presentation of the PhD topic:

Pathological extracellular matrix (ECM) of tumor tissue contributes to the progression and spead of cancer. This is the case for head and neck cancer, the 6th most prevalent cancer worldwide. Immune-based therapies have shown promising results, yet only a fraction of patients responds. This project aims at better characterizing the ECM and its regulation of immune escape mechanisms using in vitro models developped at iBV and multi-parametric histological stainings of ECM components in human head and neck tumors.

The objectives of the project are twofold. First, the PhD candidate will develop a machine learning framework to characterize and classify the different types of ECM in healthy and pathological contexts from slide scanner data acquired at iBV (Nice, France). Secondly he/she will propose a model based on graphs of the ECM and derive the statistical tool to simulate and analyse the ECM.  The numerical and mathematical development will be performed in the Morpheme Team (Sophia Antipolis, France). The statistical framework will be developed in collaboration with Bristol University during a six-month stay.

GaNforPro: GaN for Proton Therapy

Supervisors:

  • Researcher Jean-Yves DUBOZ, CRHEA-CNRS (Centre de Recherche sur l’Hétéro-Epitaxie et ses Applications)
  • Researcher Joël Herault, Institut Méditerranéen de Proton Thérapie Centre Antoine Lacassagne

International partners: Professor Andreas D. Wieck, University of Bochum (Germany)

Presentation of the PhD topic:

Proton therapy is currently the most advanced radiation therapy available for cancer treatment. Due to its specificities, proton beam can destroy cancer cells without attacking the surrounding healthy tissue. However, the proton beam position and shape must be accurately measured before each radiation since it directly affects the treatment efficiency and the eventual collateral damages. We propose a new calibration approach by developing robust GaN semiconductor detectors alloying to increase the control of the irradiated dose while strongly reducing the system complexity and cost. This innovation may thus drastically improve the proton therapy.

In this context, the student will participate at all steps required to elaborate the GaN detectors at CRHEA-CNRS (http://www.crhea.cnrs.fr). He/she will benefit of the access to the regional technological platform CRHEATEC in order to develop the different processes of these devices fabrication. In a second period, he/she will characterize the devices directly on the medical site using the proton beam of the IMPT-CAL (Institut Méditerranéen de Proton Thérapie – Centre Antoine Lacassagne, https://www.protontherapie.fr). Subsequently, the student will spend 6 months in Professor Wieck's group at the University of Bochum to manufacture complete arrays of detectors but also to develop their interfacing with a commercially available readout circuit based on silicon.

ModeLInvasion: Computational modeling of the mechanical forces regulating tumor invasion

Supervisors:

  • Researcher Frédéric Luton, Institut de Pharmacologie Moléculaire et Cellulaire (IPMC),
  • Researcher Olivier Pantz, Laboratoire de Mathématiques J. A. Dieudonné,

International partners: Professor Keith E. Mostov, University of California, San Francisco.

Presentation of the PhD topic:

The proposed PhD project is transdisciplinary between computational modeling and cancer biology. Despite a considerable body of research, there is still no predictive marker to help clinicians discriminate between tumors that will remain benign from those that will develop into deadly invasive metastases.

We wish to address this major public health issue by combining computational modeling with the study of cell culture models of invasion in vitro. The ultimate goal is to draw out a predictive tool for invasive cancers. The primary objective of the PhD project is to complete the mathematical modeling directing in silico simulations of the mechanical forces required for tumor invasion. Key parameters used to feed and refine the computational model will be obtained using a 3D-cell culture system and confocal microscopy, in order to identify cellular forces controlling invasion.

The field of computational modeling, commonly used by all the big pharma but also other industries (aeronautic, automobile, civil engineering, etc), is perceived as an essential element in R&D and offers a wealth of career opportunities.

NeuroTech: Experimental study of a neurodevelopmental brain disease with high-resolution neurotechnologies and multi-scale computational solutions

Supervisors:

  • Research Director Michèle Studer, IBV (Institute of Biology Valrose),
  • Associate Professor Franck Grammont, LJAD (Jean Alexandre Dieudonné Laboratory of mathematics).

International partners: Senior Researcher Luca Berdondini, NetS3 Laboratory (Microtechnology for Neuroelectronics), Istituto Italiano di Tecnologia (Italy).

Presentation of the PhD topic:

A major challenge in the study of the nervous system, either normal or pathological, is to understand how complex brain functions are implemented and executed at the neural circuit level. We propose to use High Density MultiElectrode Array (HD-MEA) technology both on ex vivo and in vivo preparations to record the activity of thousands of neurons and apply innovative computational techniques to analyze how neurons modulate and synchronize their activity within neuronal circuits. The results of this work should provide innovative solutions to develop new implants for cerebral or medullar stimulation in humans and be of great interest both for biotechnological industry and medicine.

NUTOX-UME: Nuclear Chemical Toxicology - Understand, Model, Explain

Supervisors:

  • Professor Christophe Den Auwer, ICN (Institute of Chemistry of Nice),
  • Assistant Professor Sandra Perez, ESPACE laboratory,
  • Col. Pharmacist Denis Josse, SDIS (Fire and Rescue Department) des Alpes-Maritimes,

International partners: Researchers Johannes Raff, Satoru Tsushima, Helmoltz Zentrum Dresden Rossendorf (Germany); Professor Michael Kumke, University of Potsdam.

Presentation of the PhD topic:

Whether nuclear energy is being used as a source of energy or for other applications, it is subject to controversy: it tends to feed fears and diverse conspiracy theories at diverse scales. Behind those fears, risks in human contamination in case of an unprevented event are being questioned and the toxicology of plutonium in particular is a scientific challenge and a social stumbling block.

The objective of this PhD thesis is double: better understand and model a specific Pu-protein interaction involved in human nuclear toxicology; question how scientific knowledge impacts public opinion on nuclear safety and as a result, how this modulates risks and crises management. On the biochemistry side, a model protein (phosvitin) will be selected and basics of plutonium-protein interaction mechanisms will be explored. On the sociology side, a parallel will be drawn between fundamental research in this field and public perception through the elaboration of a public survey on nuclear toxicology.

Nutrimorph: Nutritional lipids and brain remodeling (measurement and characterization of glial plasticity using a cellular morphometric tool)

Supervisors:

  • Researcher Carole Rovere, IPMC (Molecular and Cellular Pharmacology Institute)
  • Researcher Eric Debreuve, I3S (Laboratory of Information and communication Science of Sophia Antipolis)

International partners: Research Director Denis Richard, IUCPQ-UL (Québec Heart and Lung Institute), Laval University

Presentation of the PhD topic:

Obesity and metabolic syndromes correspond to a state of chronic systemic inflammation that leads to deregulation of feeding behavior. Cell morphometric tools are becoming useful tools for studies associating cellular responses in the brain with feeding behavior.

Thanks to an innovative technological approach, this project aims to understand the cell based mechanisms involved in the cerebral inflammatory response induced by different types of fat diets. The candidate will develop an image analysis procedure to automatically, i.e. a reliable and investigator independent procedure, measure the changes in the morphology of astrocyte and microglial cells to determine the degree of cell activation by fat diets.

This objective will be decomposed into three main steps:

  1. development of specific image processing tools and pipelines to automatically detect glial cells on images
  2. characterization of these detected objects, and
  3. analysis of these data using machine learning.

We will then attempt to define, using pharmacogenetic tools, whether inhibition of early postprandial activation of glial cells prevents food intake and obesity in order to be able to offer innovative therapeutic management for the treatment of obesity.

P2MT: Preclinical Mouse Model Tracking

Supervisors:

  • Researcher Philippe Le Thuc, LEAT (Laboratory of Electronics, Antennas and Telecommunications),
  • Researcher Georges F. Carle, TIRO-MATOs, CEA (Alternative Energies and Atomic Energy Commission),

International partners: Assistant Professor Hardik J. Pandya, Department of Electronic Systems Engineering, Indian Institute of Science, Bangalore.

Presentation of the PhD topic: 

While pre-clinical research remains today a pre-requisite in the validation of new drugs and therapies both for pharmaceutical companies as well as academic research, the need for reducing animal use (3R rules) requires the development of novel tracking systems able to collect biological data throughout their whole life. In order to match this challenge, we developed a communicating implantable tag combined with a set of antenna capable of recording the behaviour of hundreds of mice in their standard housing environment.

The PhD candidate will participate in the upgrading of this pre-clinical eHealth system with the final goal of recording in real time biological and biochemical parameters H24, 7 days a week. To achieve this objective, several challenges will be addressed :

  • the miniaturization of the implanted device ;
  • the development of algorithms for Big Data analysis and the exploitation of these data by deep learning or AI (Artificial Intelligence) for better monitoring and health status prediction ;
  • the addition of sensors capabilities (temperature, pressure, enzymatic activities,…) to the identifying function of the tag.
RV-Model : Electromechanical Modelling of the Right Ventricle (the Forgotten Chamber)

Supervisors:

  • Research Scientist Maxime Sermesant, INRIA (French National Institute for computer science and applied mathematics),
  • Assistant Professor & Doctor Pamela Moceri, CHU Nice (University Hospital of Nice).

International partners : Research Professor Bart Bijnens, IDIBAPS (Biomedical research Institute August Pi I Sunyer), ICREA (Catalan Institution for Research and Advanced Studies).

Presentation of the PhD topic : 

Despite AI important success in the recent years, its limited robustness to variations in input data makes it challenging to apply in healthcare. One reason is the lack of prior knowledge on human anatomy and physiology. Biophysical modelling is a principled mathematical framework to describe physiology which can encode prior medical knowledge. Electromechanical modelling of the heart has been an active research area in the last decades, however most of the focus has been on the left ventricle, while the right ventricle has been mostly ignored. Right ventricular (RV) function evaluation is of utmost importance in heart failure, congenital heart disease, pulmonary arterial hypertension, pulmonary embolism, and most of respiratory diseases.

Healthcare and biomedical engineering have one of the strongest recruitment increase in the last years, and skills acquired through this project will position well the fellow for his career. This project will utilise computational approaches in healthcare, which is a research area with an important growth. The interactions with academic and industrial partners will ensure employability in these two sectors. Medical imaging companies are currently developing new tools for shape and deformation analysis of the right ventricle. Such modelling approach is very complementary and could extend the possibilities of such products. Therefore there is an important potential for technology transfer. Finally, the Digital Twin concept which aims at creating a digital version of a patient to help diagnosis and therapy planning is currently promoted by large healthcare companies (Philips, Siemens,...). This electromechanical modelling project is perfectly in line with this concept, and should be of interest to these companies.

Super-Dynamic-I2Fun: Super-resolution Dynamic Imaging of Intracellular Compartments in a Human Fungal Pathogen

Supervisors:

  • Research Director Robert Arkowitz, IBV (Institute of Biology Valrose),
  • Research Director Laure Blanc Feraud, I3S (Laboratory of Information and communication Science of Sophia Antipolis).

International partners:

  • Professor Neil A. R. Gow, University of Exeter,
  • Professor Michael Unser, Biomedical Imaging Group, EPFL (Ecole polytechnique fédérale de Lausanne).

Presentation of the PhD topic: 

Worldwide, fungal infections cause significant morbidity and mortality and Candida species are major etiological agents of such life-threatening infections. Candida albicans, a normally harmless commensal, is found on mucosal surfaces in most healthy individuals, yet it can cause superficial as well as life-threatening systemic infections. Its ability to switch from an ovoid to a filamentous form, in response to environmental cues, is critical for its pathogenicity. The apical zone of the filament is densely packed with multiple highly dynamic membrane compartments, including secretory vesicles and Golgi cisternae. To understand the exquisite regulation of apical polarized growth, it is critical to follow the movement of these compartments in 3D, with high spatial and temporal resolution. This project will develop, optimize and apply super-resolution imaging approaches, in particular those taking advantage of fluorescent molecule blinking and their independent fluctuations in time, to study membrane traffic reorganization during filamentous growth in this Human fungal pathogen.

The recruited PhD candidate will follow different fluorescent protein fusions expressed in C. albicans live cells in super-resolved images obtained by reconstruction from wide-field acquisition. The entire acquisition pipeline will be optimized, from the experimental conditions to the reconstruction algorithm, for quantitative analysis of C. albicans hyphal, subcellular structure and dynamics.