Autoregressive modelling of electroencephalographic signals for medical simulators

Abstract

The recording of brain electrical activity has led to a greater understandingof different neurophysiological states, has even made possible thediagnosis of some neuronal disorders, hence the importance of characterizationand understanding of the different morphologies that mayhave electroencephalography signals (EEG). The mathematical modelingof biomedical signals facilitates the development of simulators that canbe useful as medical training tools on computers or mobile devices. Thispaper presents the autoregressive (AR) modeling and simulation ofEEG signals in different physiological states: seizures, resting with eyesopen and eyes closed, and also under the presence of some of the mostcommon artifacts: muscle, eye blinking, electrode “pop”, and 60-Hz.The performance of the models has been validated in the time domainusing the percentage of fitting (FIT), which was always above 70%, andin the frequency domain through energy of the characteristic frequencybands of the EEG. The modeling methodology, figures of simulatedsignals and the values of the parameters evaluated are presented. Thewide variety of EEG signals modeled allow the development of brainsignals simulators for training of medical personnel, and also for theanalysis and characterization of EEG signals.