k-Shape clustering for extracting macro-patterns in intracranial pressure signals

Area of Science:

• Neurosurgery
• Biomedical Engineering
• Signal Processing

Background:

• Intracranial pressure (ICP) monitoring is crucial in neurosurgery.
• Telemetric devices enable broader ICP monitoring, including at home, revealing diverse ICP waveforms.
• Identifying ICP macro-patterns (seconds to minutes) may enhance understanding of patient symptoms.

Purpose of the Study:

• To establish a foundation for automatic ICP macro-pattern identification.
• To understand the distribution and clinical significance of ICP macro-patterns.
• To develop a standardized library of ICP macro-patterns.

Main Methods:

• Utilized k-Shape clustering algorithm for ICP signal analysis.
• Developed a methodology for automatic macro-pattern identification.
• Built a standard library of identified ICP macro-patterns.

Main Results:

• Successfully extracted seven distinct macro-patterns from ICP signals.
• Created a macro-pattern library for classifying ICP variations.
• Demonstrated the potential for disease-specific classification based on ICP patterns.

Conclusions:

• This study provides a computational framework for characterizing ICP recordings.
• Offers a starting point for researchers studying a wide range of neurological disorders.
• Facilitates a deeper understanding of ICP signal variations in clinical practice.