Personalizing the Pressure Reactivity Index for Quantifying Cerebral Autoregulation in Neurocritical Care

Area of Science:

• Neurocritical care
• Physiological monitoring
• Biomedical engineering

Background:

• Cerebral autoregulation is critical for managing patients with severe brain injuries.
• The Pressure Reactivity Index (PRx) is a widely used but sensitive metric for assessing cerebral autoregulation.
• Patient variability and hyperparameter selection can introduce significant errors in PRx calculations.

Purpose of the Study:

• To enhance the clinical utility of the Pressure Reactivity Index (PRx) for cerebral autoregulation assessment.
• To develop a personalized PRx algorithm (pPRx) to improve accuracy and robustness.
• To identify optimal hyperparameters for the standard PRx algorithm.

Main Methods:

• Quantified algorithmic errors using simulated and real-world multimodal monitoring data from traumatic brain injury patients.
• Identified patient heart rate as a source of PRx error using linear regression.
• Developed the pPRx algorithm by re-parameterizing PRx averaging to individual heartbeats.
• Determined ideal hyperparameters for the standard PRx algorithm to minimize errors.

Main Results:

• The standard PRx algorithm demonstrated sensitivity to hyperparameters and patient variability, with errors linked to heart rate.
• The personalized PRx (pPRx) methodology significantly reduced noise and sensitivity to patient variability and hyperparameter choices.
• Optimal hyperparameters for the standard PRx were identified as 10-second averaging windows and 40-sample correlation windows.
• pPRx demonstrated superior robustness and accuracy in cerebral autoregulation estimation.

Conclusions:

• The personalized PRx (pPRx) algorithm enhances the robustness and accuracy of cerebral autoregulation estimation.
• Addressing patient- and hyperparameter-sensitivity with pPRx is crucial for reliable clinical decision-making in neurocritical care.
• Improved cerebral autoregulation estimation can aid in identifying precision medicine targets and improving patient outcomes.