Hodnocení vlivu prostředí magnetické rezonance na akvizici a analýzu simultánního HD-EEG signálu

Abstract

Simultaneous recording of high-density electroencephalography (HD-EEG) and functional magnetic resonance imaging (fMRI), known as EEG-fMRI, allows the investigation of the relationship between brain activity and blood flow. However, the strong electromagnetic fields in the magnetic resonance (MR) environment pose significant challenges for EEG signal quality, including artifacts such as gradient switching, pulse-related, and motion-induced disturbances. This thesis aims to explore the impact of the MR environment on HD-EEG signal acquisition by comparing data recorded in a shielded cabin versus inside the MR scanner. A preprocessing pipeline is applied to both datasets to reduce noise and artifacts, using methods such as the independent component analysis (ICA). Moreover, a microstate analysis was then performed to identify stable topographic patterns in the EEG data, as well as to extract relevant spatiotemporal parameters for each recording setting and both combined. Statistical comparisons revealed significant differences in global explained variance (GEV), mean duration, and occurrence between the two environments. Despite some spatial similarities in dominant microstates, temporal dynamics were notably influenced by the measurement surrounding. Potential sources include remaining artifacts, postural differences between recording setups, and environmental factors such as MR scanner noise. These findings emphasize the importance of considering acquisition conditions when interpreting EEG microstate data.

Simultaneous recording of high-density electroencephalography (HD-EEG) and functional magnetic resonance imaging (fMRI), known as EEG-fMRI, allows the investigation of the relationship between brain activity and blood flow. However, the strong electromagnetic fields in the magnetic resonance (MR) environment pose significant challenges for EEG signal quality, including artifacts such as gradient switching, pulse-related, and motion-induced disturbances. This thesis aims to explore the impact of the MR environment on HD-EEG signal acquisition by comparing data recorded in a shielded cabin versus inside the MR scanner. A preprocessing pipeline is applied to both datasets to reduce noise and artifacts, using methods such as the independent component analysis (ICA). Moreover, a microstate analysis was then performed to identify stable topographic patterns in the EEG data, as well as to extract relevant spatiotemporal parameters for each recording setting and both combined. Statistical comparisons revealed significant differences in global explained variance (GEV), mean duration, and occurrence between the two environments. Despite some spatial similarities in dominant microstates, temporal dynamics were notably influenced by the measurement surrounding. Potential sources include remaining artifacts, postural differences between recording setups, and environmental factors such as MR scanner noise. These findings emphasize the importance of considering acquisition conditions when interpreting EEG microstate data.