Identifying neural dynamics using fluid dynamics methods

Brain activity can be recorded by an electroencephalogram (EEG). When investigating the data, it is notoriously difficult to extract neural dynamics related directly to e.g., a visual stimulus in an experiment, from the ‘background’ brain activity. A similar problem arises in fluid dynamics, when ordered dynamics, so-called coherent structures, are extracted from measurement data. There, multiple coexisting dynamics which are superimposed in the data need to be separated for further analysis. In recent years a technique called ‘Spectral Proper Orthogonal Decomposition’ (SPOD) was developed at TU Berlin and proved very helpful in identifying and separating dynamic structures in flows. This method is now applied to EEG data that will be recorded during an experiment in which subjects must perform cognitively demanding tasks. The specific patterns in brain activity correlating to the subject’s performance are identified. To predict when an individual’s ability to reliably solve those tasks decreases, a method of monitoring the identified patterns in close to real-time needs to be developed.

Faculty Supervisor:

Robert Martinuzzi

Student:

Partner:

Technische Universität Berlin

Discipline:

Engineering

Sector:

Technology; Information and Communications Technology; Health and Related Sciences & Technology

University:

University of Calgary

Program: