VASC Seminar: Pétia Georgieva
Particle Filter Framework for Localization of Dynamic EEG Sources (Joint VASC-CBI Seminar)
Assistant Professor, University of Aveiro, Portugal
October 08, 2012, 3pm-4pm, NSH 1305
Electroencephalography (EEG)-based brain computer interface (BCI) is the most studied non- invasive interface to build a direct communication pathway between the brain and an external device. However, the technology’s susceptibility to noise in EEG measurements seems to impede such efforts. Alternatively, building BCIs based on less noisy brain sources activity instead of their surface projections, measured by the noisy EEG signals, seems a promising but still not well explored direction. In this context, finding the locations and waveforms of inner brain sources represents a crucial task for advancing source-based non-invasive BCI technologies. A particle filter (PF) framework to simultaneously estimate the number of EEG sources, their locations in the 3D head geometry and their corresponding waveforms is the focus of this talk. Status-quo in EEG source localization assumes fixed source locations, independently of the different stimuli that excite the brain. The proposed PF framework presents a shift in the current paradigm by estimating dynamic EEG sources, which may vary from one location to another in the brain depending on internal and external stimuli that animate the brain. Our simulations, based on generated and real EEG data, show that the proposed particle filtering approach estimates the dynamic EEG sources with high fidelity and outperforms popular techniques for EEG source localization like beamforming.
Host: Fernando De la Torre
Appointments: Bernardo Pires (email@example.com)
Petia Georgieva is an Assistant Professor at the Department of Electronics Telecommunications and Informatics, University of Aveiro, Portugal and the Head of Signal Processing Lab of the Institute of Electrical Engineering and Telematics of Aveiro (IEETA). She is a visiting faculty in the framework of the program CMU-Portugal faculty exchange. Her recent research interests are in applying machine learning techniques as Particle filters and Beamforming for Inverse modeling, with application in brain neural activity recovering based on EEG recordings and source-based BCI.