Heart Rate Variability via Poincaré Mapping as an Early Biomarker Post-Cardiac Arrest.

Area of Science:

• Biomedical Engineering
• Computational Neuroscience
• Cardiovascular Research

Background:

• Predicting neurological outcomes after cardiac arrest is a significant clinical challenge.
• Early and reliable prognostication is crucial for guiding neuroprotective interventions.
• Current methods often lack the precision needed for timely decision-making.

Purpose of the Study:

• To develop and validate a novel two-stage approach for predicting neurological outcomes post-cardiac arrest.
• To utilize advanced heart rate variability (HRV) features for enhanced prognostic accuracy.
• To establish a reliable method for early neurological assessment within the first hour of resuscitation.

Main Methods:

• Utilized a rodent model resuscitated after cardiac arrest.
• Extracted features from classic HRV and advanced Poincaré vector mapping.
• Employed Ant Colony Optimization with Dynamic Pheromone Decay and Knowledge Distillation (ACO-DPKD) for feature selection.
• Classified selected features using a support vector machine (SVM).

Main Results:

• ACO-DPKD effectively identified key HRV features for outcome prediction.
• Achieved 90% accuracy in predicting neurological outcomes within 1 hour post-resuscitation.
• Integration of advanced Poincaré metrics improved prediction accuracy by approximately 20% compared to traditional HRV features.

Conclusions:

• Optimized machine learning classification within the first hour post-cardiac arrest supports timely neuroprotective strategies.
• Advanced Poincaré vector features significantly contribute to early and accurate prognostic assessment.
• This approach provides a foundation for improved clinical management of cardiac arrest survivors.