The evolution of complexity in human brain development: an EEG study

Analysis of the EEG as a signal from a deterministic non-linear system should, in principle, allow insights into the complexity of underlying brain activity. We examined the capability of this method to analyse the marked changes in brain activity during normal brain development. Resting EEGs of 54 healthy children (newborns to 14 years old) and of 12 normal adults were recorded digitally. The following parameters were calculated: correlation dimension, a measure of the complexity of the underlying system, and the first Lyapunov coefficient, indicating the system's 'unpredictability'. Analysis of variance (ANOVA) was performed with probands grouped by age. The subgroups of children older than 1 year was further examined by regression analysis. In all analysed epochs, Lyapunov coefficients were significantly positive (P < 0.0001. t-test). The presence of non-linear dynamics was asserted statistically in 64-76% of examined epochs. A highly significant increase in correlation dimension with age was found in all examined leads (P < 0.0001, ANOVA). In all age groups, marked differences in correlation dimension in different brain regions became evident (P < 0.01-0.0001, ANOVA). Evidence for the presence of non-linearity can be found even in newborns. Brain maturation was reflected in a marked and highly significant increase in correlation dimension (complexity). Our work indicates that non-linear dynamics analysis is suitable for measuring complexity of brain activity during maturation and provides age-dependent normal values as a basis for further study.