Detection and evaluation of events in EEG dynamics in post-surgery patients with physiological-based mathematical models

As part of the new directions for vision and mission of Europe, patient well-being and healthcare become core features of a modern and prosperous society. That is, healthcare costs are optimized towards patient benefit and sideways effects such as cost-related reduction in medication, in frequency of post-operatory interventions, in recovery times and in comorbidity risk. In this paper, we address the incidence of events related to stroke, epileptic seizures and tools to possibly predict their presence from Electroencephalography (EEG) signal acquired in post-surgery patients. Wavelet analysis and spectrogram indicate graphically changes in the energy content of the EEG signal. Physiologically based neuronal dynamic pathway is used to derive fractional order impedance models. Nonlinear least squares identification technique is used to identify model parameters, with results suggesting parameter redundancy. There is a significant difference in model parameter values between EEG signal with/-out events.