The main purpose of the GinesisLab ("Laboratory for Biomedical Image Management Applications") is to develop an academic and marketable system for the harmonized management of multimodal imaging databases. The project is expected to have a significant social impact for the future of medicine by developing predictive tools for diagnostic by comparing data of an individual patient with those of healthy subjects and cohorts with identified pathologies. The GinesisLab will be composed by staff from the Neurofunctional Imaging Group (GIN), a team from the Institute of Neurodegenerative Diseases (UMR5293, Bordeaux) and from the Cadesis company (Lyon / Paris / Nantes). The GIN is investigating the functional neuroanatomy, genetic and cognitive foundations of hemispheric specialization of the human brain. Its experimental approach is based on the construction and analysis of multimodal databases of psychometric data, neuroanatomical (anatomical MRI and diffusion) and neurofunctional (functional MRI) of the human brain. The GIN is affiliated and supported since its creation in 1989 by the CEA, whose mission is to combine basic biological research and advanced technology to contribute to the advancement of knowledge and to provide solutions in an area with strong societal challenges: healthcare. Cadesis is a SME founded in 1999 that has a global expertise in deploying information systems, and in defining and optimizing business processes. Cadesis dedicates 11 % of his turnover (10M€) to activities of R&D on two main axes: the management of data for the industry and for biomedical imaging. The management of biomedical imaging data system envisioned is based on the use a PLM (Product Lifecycle Management), an approach that has been in use in manufacturing industries for over fifteen years and that Cadesis is distributing and optimizing. PLM solutions have evolved with the requirements from various contexts and are capable of responding to similar constraints to those encountered in the field of scientific studies using medical imaging: heterogeneity, large volume of documents, balance between collaboration and data protection, traceability. Such constraints are particularly obvious in the field of population imaging that has emerged in the 2000s and is characterized by both an exponential explosion of the volume of data and their multimodal nature. GinesisLab research and innovation will focus on three points: 1- specification and development of tailored user-machine interfaces adapted to end-users and specific domain profiles, 2- integration of data processing from the design side processing chains to local or deported launch on data centers, 3- design, tests and implementation of statistical methods for analysis of variability within and between imaging databases: as part of the personalized medicine, we will study methods for detecting a patient in one or more groups of healthy controls or patients with identified pathology. The creation of a joint laboratory between GIN and Cadesis is the opportunity to conduct a research joining scientific and industrial cultures, its feasibility being attested by preliminary collaborative works carried out since 2013 within the framework of the BIOMIST projects. Beyond the application proposed in the neuroimaging, Cadesis will promote the results of this research in the entire biomedical sector andand more specifically within the framework of clinical research. At the end of the project, the two parties are planning to extend their partnership through the creation of a Technological Research Team within the framework of the 2021-2025 contract of the institutional research organism.

Resting-state functional MRI (rs-fMRI) has the potential to be a valuable tool for early diagnosis of certain brain pathologies. However, its accuracy is limited by the existence of a few confounding factors as drowsiness. Recent studies suggest that drowsiness can significantly affect brain connectivity measurements. Furthermore, the use of different correction techniques can result in varying cognitive messages. Unfortunately, the conditions of rs-fMRI examination can induce drowsiness, making it challenging to use this test for personalised medicine. Following our work, the DeepRest project aims to develop a new brain imaging toolbox based on a machine learning approach for detecting drowsiness during rs-fMRI to improve its sensitivity and specificity. This tool will be developed in collaboration with four partners, evaluated on a large cohort of patients, particularly for medical follow-up purposes, and eventually integrated into the SWOMed solution created by the industrial partner.