A Novel Methodology for Intracranial Pressure Subpeak Identification Enabling Morphological Feature Analysis

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Intracranial pressure (ICP) monitoring is crucial for managing traumatic brain injury (TBI).
• Accurate ICP waveform analysis is essential for understanding physiological states and predicting secondary pathologies.
• Existing methods for ICP subpeak identification face challenges with waveform variability.

Purpose of the Study:

• To develop a novel methodology for accurate intracranial pressure (ICP) waveform subpeak identification.
• To integrate arterial blood pressure (ABP) and electrocardiogram (ECG) signals for enhanced ICP analysis.
• To improve the clinical insights derived from ICP waveform morphology in TBI patients.

Main Methods:

• Multimodal signal pre-processing and manual ICP waveform labeling.
• Semi-supervised training of a support vector machine (SVM) for ICP waveform classification.
• Dynamic time warping barycenter averaging (DBA) for template generation and derivative dynamic time warping (DDTW) for subpeak mapping.

Main Results:

• The proposed framework achieved an overall subpeak identification accuracy of 98.2% on 30,000 ICP waveforms.
• Demonstrated improvement over existing methodologies in ICP waveform subpeak identification.
• Showcased resilience to patient-specific ICP waveform morphological variations.

Conclusions:

• The novel methodology provides robust and accurate ICP subpeak identification.
• Incorporation of subject matter expertise enhances adaptability to new ICP morphologies.
• Enables detailed analysis of ICP features over time for clinical decision-making in TBI management.