A hybrid wavelet based K-means clustering approach to detect intracranial hypertension

Intracranial pressure (ICP) refers to the pressure within the skull and is known to have significant importance in revealing the compliance state of the brain. Due to the invasive nature of the ICP measurement, many researches have attempted noninvasive approaches of colletcing ICP information with limited clinical applications. In this paper, a wavelet based K-means clustering analysis has been conducted to detect the patterns of physiological signals and study their changes. The analysis has been performed on 20 patients with traumatic brain injuries and the results show that the hybrid approach is a viable method to detect the patterns in an unsupervised manner.