ABSTRACT
Electroencephalography (EEG) is perhaps the most widely used brain-imaging technique for paediatric populations. However, EEG signals are prone to distortion by motion-related artifacts, which can severely confound interpretation. Compared to adult EEG, motion during infant EEG acquisition is both more frequent and less stereotypical. Yet the diverse effects of motion on the infant EEG signal have not been documented. This work represents the first systematic assessment of the effects of naturalistic motion on infant and adult EEG signals. In Study 1, five mother-infant pairs were video-recorded during naturalistic joint- and solo-play with toys. The frequency of occurrence of 27 different facial and body motions was time-coded for both adults and infants. Our results suggested that movement was continuously present within dyads, and that different types of movement were observed when comparing social and non-social play, as well as adults and infants. In Study 2, one adult and one infant actor each re-created the most commonly occurring facial, limb and postural motions from Study 1, allowing us to assess the topological and spectral features of motion-related EEG artifacts, as compared to resting state EEG measurements. For the adult, all movement types (facial, limb and postural) generated significant increases in spectral power relative to resting state. Topographically and spectrally, the strongest contamination occurred at peripheral recording sites, and affected delta and high-beta frequency bands most severely. Exceptionally, at certain central and centro-parietal channels, virtually no motion-induced power changes were observed in theta, alpha and low-beta frequencies. By contrast, infant motions mainly produced a decrease in alpha power over fronto-central and centro-parietal regions, and was most pronounced for talking and upper limb movements. However, with the exception of peripheral channels, the infant theta band (3-6 Hz) showed little contamination by face and limb motions. It is intended that this work will inform future development of methods for EEG motion-artifact detection and removal, and contribute toward the development of common artifact-related resources and best-practice guidelines for EEG researchers in social and developmental neuroscience.