Identification of neuromagnetic responses for real-time analysis in magnetoencephalography
- Identifikation von Neuromagnetischen Hirnaktivitäten zur Echtzeitanalyse in Magnetoenzephalographie
Breuer, Lukas; Shah, Nadim Joni (Thesis advisor); Steland, Ansgar (Thesis advisor)
Dissertation / PhD Thesis
Dissertation, RWTH Aachen, 2015
Magnetoencephalography (MEG) is an ideal tool for interactively studying the spatio-temporal organization of the human brain. Recently, MEG based real-time brain computing interfaces have been developed to study the acting and responding brain interactively. It is expected that real-time analysis combined with neurofeedback techniques will offer a great potential for the development of novel diagnostic and therapy routines in neurology and psychiatry. The analysis, however, is challenging and requires the design of novel strategies to cope with real-time data processing millisecond by millisecond. In particular, under real-time conditions whole-brain source analysis is challenging, as the signal-to-noise ratio is poor and inter-trial or across subject averaging cannot be applied. Previous attempts using MEG showed results where the analysis was restricted to predefined regions of interests, while artifact rejection was either neglected or simplified. In this thesis, a new real-time analysis workflow for MEG recordings is introduced. Thereby, novel concepts in real-time artifact rejection and whole-brain source localization are demonstrated using real data from six subjects. The new analysis approach introduced here reveals that real-time whole-brain MEG analysis now becomes possible, thus providing the necessary analysis steps needed for studying the information processes in the human brain online.
- Chair of Stochastics 
- Department of Mathematics