M C Bastiaansen1, T R Knösche. 1. Cooperation Centre Tilburg and Eindhoven Universities, The, Tilburg, Netherlands. marcel.bastiaansen@mpi.nl
Abstract
OBJECTIVES: A problem with the topographic mapping of MEG data recorded with axial gradiometers is that field extrema are measured at sensors located at either side of a neuronal generator instead of at sensors directly above the source. This is problematic for the computation of event-related desynchronization (ERD) on MEG data, since ERD relies on a correspondence between the signal maximum and the location of the neuronal generator. METHODS: We present a new method based on computing spatial derivatives of the MEG data. The limitations of this method were investigated by means of forward simulations, and the method was applied to a 150-channel MEG dataset. RESULTS: The simulations showed that the method has some limitations. (1) Fewer channels reduce accuracy and amplitude. (2) It is less suitable for deep or very extended sources. (3) Multiple sources can only be distinguished if they are not too close to each other. Applying the method in the calculation of ERD on experimental data led to a considerable improvement of the ERD maps. CONCLUSIONS: The proposed method offers a significant advantage over raw MEG signals, both for the topographic mapping of MEG and for the analysis of rhythmic MEG activity by means of ERD.
OBJECTIVES: A problem with the topographic mapping of MEG data recorded with axial gradiometers is that field extrema are measured at sensors located at either side of a neuronal generator instead of at sensors directly above the source. This is problematic for the computation of event-related desynchronization (ERD) on MEG data, since ERD relies on a correspondence between the signal maximum and the location of the neuronal generator. METHODS: We present a new method based on computing spatial derivatives of the MEG data. The limitations of this method were investigated by means of forward simulations, and the method was applied to a 150-channel MEG dataset. RESULTS: The simulations showed that the method has some limitations. (1) Fewer channels reduce accuracy and amplitude. (2) It is less suitable for deep or very extended sources. (3) Multiple sources can only be distinguished if they are not too close to each other. Applying the method in the calculation of ERD on experimental data led to a considerable improvement of the ERD maps. CONCLUSIONS: The proposed method offers a significant advantage over raw MEG signals, both for the topographic mapping of MEG and for the analysis of rhythmic MEG activity by means of ERD.
Authors: Anil M Tuladhar; Niels ter Huurne; Jan-Mathijs Schoffelen; Eric Maris; Robert Oostenveld; Ole Jensen Journal: Hum Brain Mapp Date: 2007-08 Impact factor: 5.038
Authors: Heng-Ru May Tan; Arthur C Leuthold; David N Lee; Joshua K Lynch; Apostolos P Georgopoulos Journal: Exp Brain Res Date: 2009-05-08 Impact factor: 1.972
Authors: Hanneke van Dijk; Jurrian van der Werf; Ali Mazaheri; W Pieter Medendorp; Ole Jensen Journal: Proc Natl Acad Sci U S A Date: 2009-12-22 Impact factor: 11.205