Title: The role of oscillations and Causality analysis: A systems modelling framework

Abstract: The study of different types of rhythmicities of the brain and their relation with different pathologies and functions has been an important subject of neuroscience, physiology and neurophysiology research. EEG is an important non-invasive technique for medical diagnosis in clinical neurophysiology, as well as for scientific study of brain function in cognitive neuroscience. Electroencephalographers describe the EEG brain activity from the spatial distribution on the scalp including the localized region frontal, posterior, lateral and bilateral as well as the dominant frequency components. Conventionally, EEG analysis mostly relies on visual inspection of relevant EEG signals. In many cases, however, visual inspection of EEG signals may be subjective and insufficient  because statistical information contained in EEG signals may not be adequately exploited and utilised.

This talk will describe a possible approach to obtain more relatively objective and reliable analysis results. I will focus on two ongoing research problems. The first part will describe our parametric approach to modelling clinically recorded EEG data sets from patients with epilepsy and explore how theoretical models may be useful to clinicians faced with the challenges of predicting, diagnosing, characterising and ultimately treating seizures. The second part will focus on research with collaborators in Clinical Neurophysiology and explore how model based hypotheses can alter biological understanding. The focus of this section is to establish a time-varying cause-effect modelling approach based upon the Non-linear AutoRegressive Moving Average with Exogenous input (NARMAX) model identification techniques, with application to EEG data to detect the directed causal influence between EEG channels. Using dynamic models in this way gives us both a clear set of modelling assumption as well as a set of powerful tools which can help clinicians interpret EEG signals.