Developmental changes in functional EEG asymmetry: a multivariate autoregressive modeling approach.

: To investigate the functional asymmetry of the electroencephalogram (EEG) in the developing brain, we analyzed quantitatively the different components of simulated EEG (SEEG) in a model before and after cutting off the intercerebral interaction. Original EEG (OEEG) recordings were obtained from 374 normal children aged 3-12 years. The SEEGs were computed from the OEEG after applying 'ARdock', a new multivariate autoregressive (AR) modeling method. Subsequently, the univariate AR and component analysis was applied on OEEG and SEEG. The F value was computed from Mahalanobis' distance between both AR coefficients of component activities for age, EEG derivation and frequency band. Our analysis showed developmental lateralities in the alpha band at the frontal pole in 5 and 12 year old children, and at the central region in 8 and 11 year old children. Our results suggest that the intercerebral interactions develop at approximately 5 years of age at the left frontal pole during the period of maturation of alpha band activity and after 8 years of age when it replaced theta band activity. The present method of analysis may have a wide clinical application in the analysis of normal and abnormal EEGs.