Interpreting intracranial pressure waveforms by wave-nets

We propose a multiresolution based system to monitor intracranial pressure. The proposed technique exploits the concept of time-frequency analysis by using wavelet processing. The approach is compared with the standard techniques based on the Fourier analysis: with this novel method, the signals related to the breath and heart effects on the measured signal, can be easily separated. Furthermore, the time-frequency analysis allows one to implement a system that is not depending on the patient under consideration, or on the physiological conditions where the patient is. The multiresolution scheme is then exploited by using wave-nets which are used to implement a system that is able to predict the intracranial pressure signal from the observed samples.