Level :
- Last year of Engineering Schools
Internship period or duration :
- 4 – 6 months
Context :
Comas are caused by an injury to the brain that can be due to increased pressure, bleeding or swelling of brain tissues or loss of oxygen. More than 50% of comas are related to head trauma, disturbances in the brain’s circulatory system (mainly due to cardiac arrest or stroke) or subarachnoid hemorrhage (mainly due to cerebral aneurysms). In each case, the different physiopathological processes induce cerebral white-matter alterations leading to brain disconnections and, thus, to temporary or permanent brain dysfunction. Recovery from coma ADSH_Ing_QC2017_Annonce_Stage_LIB_odtcan be complete, lead to neurological deficits and can eventually not occur with the patient remaining with a permanent disorder of consciousness (DOC), namely minimally conscious state or more rarely vegetative state.
The development of neuroimaging during the last decades allowed one to non invasively evaluate the integrity of brain tissues. One of the most promising techniques to provide useful clinical insights of the neuroanatomical substrates after brain injury is the Diffusion Weighted Imaging (DWI). This MRI technique quantify the diffusion of water molecules in tissue environments, which are influenced by the microstructural organization of tissues and their constituent cells, and can provide unique insights into pathophysiology, particularly in white matter. The use of such advanced imaging techniques in a day-to-day clinical context implies a high level of reliability. Indeed, a given imaging marker measured for a single patient must be reliable, reproducible and comparable with previous patients scanned in different machines from different centers. In that context, our group developed both a control quality pipeline of acquired data and a method to reduce inter-center bias. We argue that the combination of these two strata reduce significantly the variability of regional diffusion markers between healthy subjects and increase our ability to predict neurological long-term recovery in coma patients.
Objectives :
The main objective of the proposed internship is to develop a framework to evaluate the impact of different data preprocessing strategies in the reliability of diffusion markers. Thank to a large database of healthy controls as well as a multicentric cohort of coma patients acquired since 2005, it is possible to seriously address this issue in a clinical context.
Mission(s) :
The candidate will be in charge of designing the evaluation framework, developing the data processing pipelines, processing the data and reporting the main results.
Required skills :
- Good skills in Python development.
- Knowledge in image processing will be appreciated
Remuneration :
- Legal intership allowance (~520 €).
