Functional role of brain-body interactions in critically ill patients – Brain-Body-ICU

Critical illness is an acute, life-threatening condition with high mortality and lasting cognitive and psychological sequelae, known as post-intensive care syndrome. It involves organ failure driven by a maladaptive acute stress response. The autonomic nervous system (ANS), which connects peripheral organs to cortical and brainstem autonomic centers, coordinates this response through its effects on the heart, lungs, and the digestive and immune systems to restore homeostasis. Certain regions of the ANS, particularly the brainstem cardiorespiratory centers and limbic structures, are especially vulnerable during critical illness, potentially contributing to poor outcomes. Accordingly, brainstem dysfunction, assessed clinically or through evoked potentials, is a major prognostic factor in intensive care. Furthermore, studies in cognitive neuroscience suggest that ANS activity is better understood through brain-body interactions rather than isolated organ measurements. We hypothesize that these brain-body interactions are central to the pathophysiology of critical illness and can be quantified using multi-system electrophysiological signals, in order to better understand critical illness and develop robust prognostic tools.

To test this hypothesis, we will include 150 adult patients admitted to intensive care for a critical illness (with or without primary brain injury), characterized by neurological, cardiac, and respiratory failure. We will simultaneously collect non-invasive physiological data at the patient’s bedside: high-density EEG, ECG, and airway pressure/flow measurements (through the mechanical ventilator for intubated patients or through a nasal canula for spontaneously breathing patients). These recordings will be acquired at rest and following thermal stress induced by a cold pressor test to stimulate the ANS. Data collection will occur during three phases of critical illness: the acute phase (<72 hours after admission), the recovery phase (~4–12 days), and at 3 months post-ICU.

Using quantitative signal analysis and advanced mathematical modelling, we will develop static and dynamic measures of brain-body interactions (WP1). For heart-brain interactions, we will use synthetic data generation models linking topographic indices of cerebral networks activity with sympathetic and vagal cardiac indices to identify the neural networks involved in the acute stress response during critical illness (WP1a). For brain-lung interactions, we will analyze EEG activity associated with patient-ventilator asynchronies and varying levels of respiratory drive, automatically quantified through analysis of airway pressure and flow waveforms in mechanically-ventilated patients and through synthetic data generation models in spontaneously breathing patients (WP1b). We will characterize brain-body interaction profiles according to the degree of clinical encephalopathy and systemic inflammatory response, and track their evolution across the three ICU phases (WP2). We will assess the performance of brain-body interaction markers in predicting short- and long-term outcomes, particularly ICU mortality and cognitive impairment at 3 months (WP3). Finally, we will explore the role of persistent alterations in brain-body interactions in relation to the presence or absence of post-intensive care syndrome at 3 months (WP4).

This integrative, multi-systemic, and translational approach aims to identify precise pathophysiological biomarkers of the stress response, to improve our understanding and prediction of outcomes in critically ill patients, as well as of the pathophysiology of post-intensive care syndrome, which leads to psycho-cognitive impairments affecting quality of life. Furthermore, this work will pave the way for monitoring ANS activity and developing biomarkers for individualized, targeted therapies.