Investigating brain mechanisms underlying natural reading by co-registering eye tracking with combined EEG and MEG

Abstract

Linking brain and behavior is one of the great challenges in cognitive neuroscience. Ultimately, we want to understand how the brain processes information to guide every-day behavior. However, most neuroscientific studies employ very simplistic experimental paradigms whose ecological validity is doubtful. Reading is a case in point, since most neuroscientific studies to date have used unnatural word-by-word stimulus presentation and have often focused on single word processing. Previous research has therefore actively avoided factors that are important for natural reading, such as rapid self-paced voluntary saccadic eye movements. Recent methodological developments have made it possible to deal with associated problems such as eye movement artefacts and the overlap of brain responses to successive stimuli, using a combination of eye-tracking and neuroimaging. A growing number of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) are successfully using this methodology. Here, we provide a proof-of-concept that this methodology can be applied to combined EEG and magnetoencephalography (MEG) data. Our participants naturally read 4-word sentences that could end in a plausible or implausible word while eye-tracking, EEG and MEG were being simultaneously recorded. Eye-movement artefacts were removed using independent-component analysis. Fixation-related potentials and fields for sentence-final words were subjected to minimum-norm source estimation. We detected an N400-type brain response in our EEG data starting around 200 ms after fixation of the sentence-final word. The brain sources of this effect, estimated from combined EEG and MEG data, were mostly located in left temporal lobe areas. We discuss the possible use of this method for future neuroscientific research on language and cognition.