Morphology analysis of intracranial pressure using pattern matching techniques

We present a clustering algorithm based on dynamic time warping (DTW) to automatically classify intracranial pressure (ICP) beats based on their morphology. The algorithm detects, classifies and labels each beat as a low-pressure or high-pressure beat based on morphology. The trend is removed during preprocessing to ensure the classifications are independent of the mean ICP. An ICP beat detection algorithm is used to automatically detect each beat. We measured the performance of the algorithm compared to expert classification of ICP beats acquired from intensive care unit patients using linear and nonlinear temporal alignment techniques. The algorithm achieved a superior performance using non-linear temporal alignment.