Application of Gaussian Processes for Analyzing Interrelationship between PO2, CPP and ICP of Head Injury Patients

A. Tsaturyan, V. Bezhanyan, S. Sargsyan (Armenia), and L.A. Gabrielian (Australia)


Mechanical head injury, Gaussian processes, data analysis.


Objectives: To date mechanical head injury is one of the major health care problems. In management of patients with head injury intracranial pressure, cerebral perfusion pressure, and brain tissue oxygen tension are of critical importance. To improve the management of such patients the use of more efficient and revealing methods of data analysis can be helpful. We applied Gaussian processes to analyze clinically obtained data, previously analyzed with other methods, and compare the results. Study design/settings: Two deferent graphs were used where the correlation between brain tissue oxygenation and intracranial pressure or cerebral perfusion pressure were analyzed (Fig. 3 in [10]). Gaussian processes was used to analyze the acquired data and assess if it can reveal more details in the dynamics of studied parameters. Results: A total more than 300 data points were analyzed for each graph in this study. In [10] the Mann– Whitney test was used to study the acquired data. We applied Gaussian Processes (GPs) to the same dataset and the results demonstrate that GPs can reveal more dynamics of correlations between the analyzed parameters. Conclusions: Gaussian processes can be more revealing, reliable, accurate and helpful method to analyze data of head injury patients and improve their management and outcome.

Important Links:

Go Back