Project Summary: Professor Brown’s group seeks to unravel one of medicine’s big questions: how anaesthesia works. Several statistical methods and non-invasive techniques, such as functional magnetic resonance imaging magnetoencephalography (MEG) and electroencephalography (EEG), are used in the group to answer to this question. In the past years Professor Brown’s group has made great progress on developing dynamic methods for solving the source localization problem for MEG and EEG. Source localization technique can be applied to data recorded from subjects under general anaesthesia condition in order to identify brain areas activated and brain areas switched off during this condition.

The main goal of Elvira’s project is to develop real-time implementations of the algorithms previously developed in the lab for use in tracking brain states under general anaesthesia. This research lays key group for the development of a new approach to monitoring the brain in patients undergoing general anaesthesia for surgery. The starting point of her research is an algorithm already applied for the source localization in MEG data. The algorithm has been shown to be reliable modelling the spatiotemporal covariance of MEG measurements obtained from human subjects in resting-state and estimating the effect of different stimuli in neural activity by MEG data recorded from human subjects in evoked-potential studies. As first task of Elvira’s project, she is seeking to apply the algorithm at EEG data that the group has recorded in 10 subjects across different states of general anaesthesia. The following objective will be to improve the algorithm in order to speed up its computation. The main advantage of a faster source localization would be the possibility of an extension to estimation in real-time. The core material of this project include Kalman filtering principles, EM algorithm techniques, state-space estimation paradigms and implementation of the algorithm on the super computer cluster of the group.