Type of Publication: | Dissertation |
URI (citable link): | http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-9367 |
Author: | Hauk, Olaf |
Year of publication: | 2000 |
Title in another language: | Lineare Verteilte-Quellen-Analyse von EEG- und MEG-Daten: Theorie, Implementation und Anwendung in Studien zur Sprachproduktion |
Summary: |
The electric potentials and the magnetic fields measured at or above the scalp surface by the electroencephalogram (EEG) and by the magnetoencepahlogram (MEG), respectively, are correlates of brain activity with high temporal resolution. The bioelectromagnetic inverse problem consists of estimating the generators of these signals on the basis of the measured signals. In EEG and MEG experiments on cognition, the application of simple source models (like dipole models) is usually not justified. Therefore, mainly distributed source models will be considered in this work.
Using the concept of 'resolution kernels', it is argued that the classical minimum norm method (MN) provides a solution to the inverse problem with minimal modelling assumptions. Parameters are suggested to quantify spatial resolution, localisation accuracy, depth sensitivity and stability of minimum norm solutions. An implementation allowing convenient visualisation, appropriate parameterisation for the estimation of the above mentioned parameters, and efficient data reduction for statistical analysis, is suggested. It is demonstrated that the minimum norm method is suited for the analysis of spontaneous brain activity. In the second part of this thesis, the MN method was applied to data from psycholinguistic experiments on picture naming. A 'grapheme-monitoring task' was adopted for EEG and MEG experiments, based on Levelt's framework of language production. Subjects had to indicate by button press whether a given letter was present in the name of an object drawing or not. Assuming a 'left-to-right' monitoring of the picture name, this design allowed determining the time range of phonological encoding in picture naming. Source analysis of the EEG data associated left-frontal brain areas with this processing stage. No such lateralisation was found for the MEG. Discrepancies between the MEG and EEG results are discussed. |
Examination date (for dissertations): | Jan 29, 2001 |
Dissertation note: | Doctoral dissertation, University of Konstanz |
PACS Classification: | 87.19.Nn; 02.30.Zz |
Subject (DDC): | 150 Psychology |
Controlled Keywords (GND): | Physiologische Psychologie, Elektrophysiologie, Inverses Problem, Gehirn, Sprachproduktion |
Keywords: | Minimum Norm, Lineare Schaetzer, EEG, MEG, Bildbenennung, Minimum Norm, Linear Estimation, EEG, MEG, Picture Naming |
Link to License: | In Copyright |
HAUK, Olaf, 2000. Linear distributed source analysis of EEG and MEG : theory, implementation and application to studies on language production [Dissertation]. Konstanz: University of Konstanz
@phdthesis{Hauk2000Linea-10621, title={Linear distributed source analysis of EEG and MEG : theory, implementation and application to studies on language production}, year={2000}, author={Hauk, Olaf}, address={Konstanz}, school={Universität Konstanz} }
all_in_one2.pdf | 673 |