PhD Position in Machine Learning for Neuroimaging Data (Radiology Department)

CHUV - Centre Hospitalier Universitaire Vaudois Lausanne

Postuler
  • Date de publication :

    21 mars 2024
  • Taux d'activité :

    100%
  • Type de contrat :

    Durée indéterminée
  • Lieu de travail :

    Lausanne

PhD Position in Machine Learning for Neuroimaging Data (Radiology Department)

Recherche

PhD Position in Machine Learning for Neuroimaging Data (Radiology Department)

The team of the Translational Machine Learning Lab, is looking for candidates to fill a three-year PhD position in machine learning for behavioral outcome prediction from developmental neuroimaging data.

Contexte

The Lausanne University Hospital (CHUV) is one of five Swiss university hospitals. Through its collaboration with the Faculty of Biology and Medicine of the University of Lausanne and the EPFL, CHUV plays a leading role in the areas of medical care, medical research and training.

The Radiology Department has a strong research focus, advancing MR acquisition, image processing, and machine learning for radiology and clinical research. The department is part of the Center for Biomedical Imaging (CIBM), a joint undertaking of the EPFL, University of Geneva (UNIGE), UNIL, and the Geneva University Hospital (HUG).

HES-SO, the University of Applied Sciences and Arts Western Switzerland, excels in applied research, partnering with university hospitals and UNIGE's and UNIL's medical and psychology departments. Its industry ties, global engagement, and research centers address societal challenges, positioning it as a prominent player in Swiss higher education and interdisciplinary advancement.

The GEMMI lab (Geneva Musical Minds Lab) was founded in 2019 and is part of the Geneva University Neurocenter. It focuses on experience-driven brain and behavioral plasticity following musical, artistic, and physical activity over the lifespan aiming to boost child development and countervail age-related cognitive and cerebral decline.

Founded in 2020 at CHUV and UNIL's Radiology Department, the Translational Machine Learning Lab develops machine learning for diverse real-world imaging and medical data to enhance disease comprehension and patient outcomes.

RATIONALE OF THE PROJECT

Executive functions (EF) regulate goal-oriented behavior, encompassing initiation, shifting, inhibition, working memory, extending to complex tasks like abstract reasoning and planning.

Crucial for cognitive development, EF anchor memory, attention, problem-solving, decision-making, and language, consistently used daily, and pivotal for academic achievement.

The Art and Development project, funded by the Swiss National Science Foundation, aims at a breakthrough in understanding the development of EF, academic achievement, and its cerebral underpinnings in young school children.

A large-scale, long-term randomized controlled trial (RCT) in cognitive neuroscience is urgently needed, contrasting musical training with other arts and a control group. Through machine learning, this study will form a data-driven model of brain and EF growth (ages 6-8), appraising arts influence by comparing music and visual arts. Outcomes might reshape education, also considering children from disadvantaged backgrounds.

Mission

The goal of this doctoral position, in collaboration with a postdoctoral researcher within the same team and our other project partners, will be to develop new machine learning algorithms to model combined brain and behavioral development supporting executive function and academic achievements over the first two years of academic education (6 to 8-year-old children).

In this RCT, 150 children will receive 2-year intensive string instrumental training in an orchestra setting, or visual arts, or participate in an active control group (n=50 in each group).

At baseline, after one year and after two years, an extensive EF battery and four different types of MRI will be measured : T1-weighted MRI: MP2RAGE (gray matter assessment), Multishell Diffusion Weighted Imaging (DWI; white matter assessment); fMRI (auditory and visual working memory), and Resting State fMRI.

The candidate will develop and implement new algorithms, present work at conferences and in scientific publications, collaborate with local and international researchers, and interact with other lab members, clinicians, psychologists, bioinformaticians, physicists, engineers, and neuroscientists.

The candidate will apply machine learning to create a data-driven model of children's brain and EF progression with and without artistic enrichment. This involves comparing musical/artistic enrichment to the active control group, and translating findings into practical educational innovations. Initially, group-level analyses of imaging+behavior data, particularly brain connectivity data, will use multilinear algebra (e.g. sparse partial least squares) and signal processing (e.g. fusion ICA). Phase two focuses on individual behavioral score prediction through multimodal graph neural networks, incorporating transfer learning from large graph datasets and multi-output networks.

The "Art and Development" project, a collaboration among CHUV/UNIL, FPSE, CIBM UNIGE, and HES-SO Geneva School of Health Sciences (HEdS-GE), is overseen by Prof. Dr. Clara James (principal investigator and PhD co-director) and Dr. Jonas Richiardi (PhD Director). Abundant computing resources, GPUs, and active engagement with the Consortium and scientific partners from other institutions, along with two postdocs and MSc students, are integral. The role involves MSc student supervision and working primarily at CHUV/UNIL Lausanne, with weekly gatherings at the HEdS-GE.

Profil

  • MSc in computer science, applied mathematics, biomedical engineering, statistics, data science, statistical physics, statistics, electrical engineering, or related field
  • Good training in linear algebra, calculus, statistics
  • Good knowledge of Python 3, including relevant machine learning libraries such as scikit-learn, Pytorch 1.5+, and Tensorflow 2.
  • English proficiency is mandatory, and good French is an asset
  • Demonstrated previous experience in machine learning and deep learning
  • Genuine interest in the interdisciplinary field of the study
  • Experience in medical imaging, graph representations, image processing are an advantage.

You will be part of a welcoming team of scientists, Ph.D. students, and multiple partners, which requires a strong team spirit and professionalism. Excellent communication and interpersonal skills are as important as technical skills. This project also demands ingenuity, adaptability, and the ability to work autonomously as well as in a team.

Our team has a strong gender equity focus, and we encourage women to apply for this position.

Nous offrons

To become an employee of the renowned University Hospital Center from the Canton of Vaud is an assurance of:

  • First-rate social benefits such as a Paternity Leave of 20 days and a Maternity Leave of 4 months (there is also the possibility to obtain a complementary breastfeeding leave of 1 month)
  • 25 days of vacations per year
  • Regular salary progression adapted to your responsibilities
  • A right to at least three days of training per year, by accessing a wide offer of courses not only from the CHUV Training Center but also from external providers
  • Possibility to access one of the 500 furnished apartments offered in the surrounding neighborhoods in case of relocation in Switzerland
  • Discounts proposed on social and cultural events, goodies and other services, thanks to the “H-Oxygène” association
  • Signing up to our Mobility Plan and benefit from different advantages (discounts on public transportation, promotion of “Mobility” car fleet and discounts on electric bikes)
  • Being able to enjoy our high-quality corporate restaurants, located in every hospital building, with employees’ discount

Contact et envoi de candidature

Contact person in case of questions about this role : Dr Jonas Richiardi, jonas.richiardi@chuv.ch

All of our applications are processed electronically. For this reason, we kindly ask you to apply exclusively by clicking on the APPLY button at the bottom of the advertisement. Should you have any technical question, please contact our Recruitment Team (021/ 314-85-70 from 08h30 to 12h00 and from 14h00 to 16h30).

The CHUV applies the highest quality requirements as part of its recruitment process. In addition, mindful to promote workplace diversity and inclusion we strive to ensure equal treatment and avoid any discrimination. We are looking forward to receiving your application.

We would like to inform external recruitment agencies that any application inserted directly on our recruitment platform won’t be accepted and cannot be charged. Thank you for your understanding.

Postuler