CBFV Waveform Pattern Analysis in Noninvasive Ultrasound-based ICP Monitoring.

The noninvasive monitoring of intracranial pressure (ICP) from cerebral blood flow velocity (CBFV) obtained from transcranial doppler ultrasound has shown great promise in the recent years in terms of potential clinical use. Yet, our understanding of the variations of CBFV waveforms associated with the level of ICP remains limited. We present in this study a quantitative analysis of the CBFV waveform morphology in the context of noninvasive ICP monitoring. The predictive power of morphological waveform features and dimensionality reduction methods are analyzed side-by-side. In addition, a key component of the study is to provide a descriptive analysis of the CBFV waveform dynamic in presence of an increase in ICP. The analyses are performed on a dataset consisting of matched CBFV and ICP signals acquired on patients with traumatic brain injury (TBI) in a clinical setting. Our experimental results provide insights about which dimensionality reduction method best capture the changes in CBFV linked to ICP increase and also perform a quantitative evaluation of the morphological features of the CBFV signal that are linked to elevated ICP and could be the target of future non-invasive ICP models.