Development of relative power contribution ratio of the EEG in normal children: A multivariate autoregressive modeling approach

The relative power contribution ratio using a multivariate autoregressive model was applied to determine the spread of delta, theta, alpha and beta activity over the scalp in childhood. EEGs of 264 normal healthy subjects from 3 to 26 years were recorded from Fp(1), Fp(2), C-3, C-4, O-1 and O-2 with linked ears during the resting relaxed state with eyes closed. After selecting the epoch to remove artifacts, the relative power contribution ratio at each frequency band was calculated from the digitized EEG. The most noticeable developmental change was seen in the alpha frequency, where the relative power contribution ratio from its own area decreased significantly with age, and ratios from the other areas increased significantly. The change was larger at the frontal and smaller at the occipital region. Consequently, the occipital alpha wave was stable after 3 years of age. Developmental change at beta frequency was similar to alpha activity, but the number of components with significant change was smaller at delta and theta frequencies. Thus the relative power contribution of the EEG enabled us to observe how relationships among each location of EEG matured.