A Multimodal Approach to Personalized Tracking of Evolving State-Of-Consciousness in Brain-Injured Patients (PerBrain)

Improved treatment of severe brain injuries in Intensive Care Units has resulted in increased patients’ survival rates. While some of these patients regain consciousness after a transient state known as coma, other will develop a disorder of consciousness (DoC). Consciousness diagnosis in DoC currently relies on standardized behavioural assessment. However, this strategy may fail to detect consciousness due to major sensory and motor deficits. Moreover, DoC aetiology and pathophysiology are heterogenous and most likely result from the combination of factors whose interplay still needs to be clarified. Consciousness detection in DoC is of great importance in the context of pain management, prognosis and end-of-life decisions. Furthermore, caregiving of these patients is very stressful principally for the large uncertainty associated with them. For all these reasons is critical to develop personalised diagnosis and prognosis prediction tools that permit a stratified analysis at the single-patient level. The PerBrain Project will benefit from the multidisciplinary partners’ expertise, and the unique opportunity to perform longitudinal assessments in four clinical sites through behavioural, electrophysiological, neuroimaging and physiological techniques. Based on the collected data, we will develop a multimodal personalised diagnostic tool for DoC patients using machine learning trained to predict with high precision and recall the current state and outcome diagnosis. Critical to this objective we will develop an electronic data collection and processing system in which partners will upload their data in a simple manner and to launch pre-defined processing pipelines.The overall outcome of this project will allow drawing better single-patient predictions of state, prognosis, and rehabilitation strategies and furthermore, a better understanding the pathophysiological mechanisms behind DoC.