Rejection of stimulus-related MEG artifacts using independent component analysis

Area of Science:

• Neuroscience
• Biophysics
• Signal Processing

Background:

• Electro-gustatory (EG) stimuli offer precise temporal and spatial accuracy for studying gustatory information processing in the human brain.
• Traditional magnetoencephalography (MEG) measurements face challenges in eliminating significant stimulus-related artifacts, limiting EG technique applicability.

Purpose of the Study:

• To develop and validate a method for removing electro-gustatory (EG) stimulus-related artifacts from magnetoencephalography (MEG) data.
• To enable artifact-free MEG recordings for enhanced investigation of gustatory processing.

Main Methods:

• Independent Component Analysis (ICA) based on the information maximization (infomax) approach was applied to raw MEG data from 122 magnetometer channels.
• Independent components were analyzed, and those representing EG stimulus artifacts were excluded from signal reconstruction.
• Artifact-free MEG signals were reconstructed by remixing the remaining components.

Main Results:

• The infomax ICA method effectively removed large EG stimulus-related artifacts from MEG waveforms.
• Independent components corresponding to blinks and magnetocardiography were also successfully separated.
• Components representing spontaneous alpha-band brain activity (8-13 Hz) were identified.

Conclusions:

• Independent component analysis provides an effective solution for artifact removal in EG-stimulated MEG studies.
• This technique enhances the feasibility of using MEG to investigate brain mechanisms of taste perception.
• The method also allows for the isolation of other physiological signals and spontaneous brain activity.