© Alexander Edthofer

Understanding transitions between states of consciousness, such as wakefulness, sleep, and anesthesia, requires tracking changes in functional connectivity (FC) between brain regions. Whole-brain models offer a framework for studying these transitions and have primarily been developed using functional magnetic resonance imaging (fMRI). However, because fMRI is costly and has limited clinical applicability, mathematical models based on electroencephalography (EEG) have been developed because EEG is routinely available during surgical procedures.

Our goal is to translate an established fMRI-based FC model into an EEG-based framework while preserving the underlying dynamic structure. To this end, we identify EEG-based functional clusters, leading to a low-dimensional stochastic differential equation model. Based on this reduction, we derive a Hopf whole-brain model and investigate its dynamic properties. Specifically, we analyze bifurcation structures and transitions between dynamic regimes associated with altered states of consciousness.

EEG - Analysis

The information processing in the human brain is a complex process and accordingly our brain cells are differently active when thinking. EEG can be used to measure, analyze and graphically display this electrical activity in the brain. In the clinical field, the analysis of the EEG is of great importance, for example during an operation or in the sleep laboratory. The analysis of the different states of consciousness is carried out with the help of algorithms, which mostly work in the frequency spectrum. The aim of the activities in the cooperation with the biosignal analysis working group of the Technical University of Munich is to develop parameters that can characterize the state of consciousness in a physically and physiologically meaningful way and the associated analyzes are carried out in the time domain. For this we use the entropy of difference and compare it with the permutation entropy already used in the literature.

In this context, we also deal with the Granger causality and its application to the EEG. This measure describes the amount of information flow between two electrodes by using autoregressive models to assess whether previous information in one electrode helps predict current information in another electrode.

© Iris Feldhammer