Non-invasive intracranial pressure estimation using support vector machine.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Medical Monitoring

Background:

• Intracranial Pressure (ICP) monitoring is crucial for diagnosing and managing vascular brain disorders.
• Current invasive ICP measurement methods pose significant risks, especially for patients with traumatic brain injury (TBI).
• There is a need for non-invasive ICP assessment techniques for pathologies where invasive methods are contraindicated.

Purpose of the Study:

• To develop and validate a non-invasive method for estimating ICP.
• To utilize non-invasively obtained Arterial Blood Pressure (ABP) and Cerebral Blood Flow Velocity (CBFV) for ICP estimation.
• To enable real-time ICP monitoring with adequate temporal resolution.

Main Methods:

• Employed a non-linear Support Vector Machine (SVM) algorithm.
• Integrated ABP and CBFV signals as inputs for ICP estimation.
• Validated the method against real ICP measurements.

Main Results:

• Achieved a low error rate between the estimated and actual ICP signals.
• Demonstrated the feasibility of on-line ICP estimation.
• Confirmed adequate temporal resolution for clinical relevance.

Conclusions:

• The proposed non-invasive method using ABP and CBFV effectively estimates ICP.
• The Support Vector Machine approach allows for accurate and timely ICP monitoring.
• This technique offers a safer alternative to invasive ICP measurements for specific patient populations.