Characterizing the temporal dynamics and maturation of brain activity during sleep: An EEG microstate study in preterm and full-term infants

Abstract

By interfering with the normal sequence of mechanisms serving the brain maturation, premature birth and related stress can alter perinatal experiences, with potential long-term consequences on a child's neurodevelopment. The early characterization of brain functioning and maturational changes is thus of critical interest in premature infants who are at high risk of atypical outcomes and could benefit from early diagnosis and dedicated interventions. Using high-density electroencephalography (HD-EEG), we recorded brain activity in extreme and very preterm infants at the equivalent age of pregnancy term (n = 43), and longitudinally 2 months later (n = 33), compared with full-term born infants (n = 14). We characterized the maturation of brain activity by using a dedicated microstate analysis to quantify the spatio-temporal dynamics of the spontaneous transient network activity while controlling for vigilance states. The comparison of premature and full-term infants first showed slower dynamics as well as altered spatio-temporal properties of brain activity in preterm infants. Maturation of functional networks between term-equivalent age and 2 months later in preterms was linked to the emergence of faster dynamics, manifested in part by shorter duration of microstates, as well as an evolution in the spatial organization of the dominant microstates. The inter-individual differences in the temporal dynamics of brain activity at term-equivalent age were further impacted by sex (with slower microstate dynamics in boys) and by gestational age at birth for some microstate dynamics but not by other considered risk factors. This study highlights the potential of the microstate approach to reveal maturational properties of the emerging brain network activity in premature infants.

Keywords: brain development; electroencephalography (EEG); microstates; prematurity; resting-state.

Fig. 1.

Global fit of the template microstates for different number of microstate classes and different vigilance states. Gain in the global explained variance as a function of the number of microstate classes (3 to 15) for preterm infants at 0 months of corrected age (0 mCA) and 2 mCA in the overall recordings, without separating vigilance states. Each individual infant is color-coded based on the gestational age (GA) at birth. For a number of microstates higher than 7, the gain in the explained variance dropped below 1%.

Fig. 2.

Pipeline comparing microstate characteristics at 0 and 2 mCA in preterms in comparable vigilance states. (a) Group-level template microstates are identified separately for each age group in preterm infants and their head-surface topographical representations are illustrated. (b) Similarity between group-level microstates of the two age groups is computed using spatial correlation irrespective of polarity of the maps. Similarity between group-level microstates is also computed for each group separately (i.e., within-group), and their maximal values are used to threshold the matrix representing the between-group similarity. (c) Similar microstates between the two age groups are averaged, constituting the shared microstates (“MS shared”). Non-similar microstates are kept in each age group (0 mCA/2 mCA → MS0/MS2). The combination of shared and non-similar microstates constituted the final set of 7 template microstates for each age group, among which shared microstates are comparable. (d) The final template microstates for each age group are back projected to each individual infant recording, illustrating activation of different microstates over time. Microstate metrics are extracted for all the microstates and are compared between age groups only for the shared microstates.

Fig. 3.

Differences in microstate duration between preterm (dark green) and full-term (light green) infants at 0 mCA (left panel) and 2 mCA (right panel) during different vigilance states: (a) REM sleep; (b) NREM sleep. Asterisks represent significant differences between preterms and full-terms (**p < 0.005, *p < 0.05). For the other microstate metrics (coverage and occurrence), seeSupplementary Figure 1.

Fig. 4.

Evolution in microstate duration between 0 mCA (dark purple) and 2 mCA (light purple) in preterms, during different sleep states: (a) for the 5 shared microstates identified in REM sleep; (b) for the 6 shared microstates in NREM sleep. Asterisks represent significant differences between 0 and 2 mCA infants (**p < 0.005, *p < 0.05). For the other microstate metrics (coverage, occurrence, and global explained variance), seeSupplementary Figure 1.

Fig. 5.

Impact of perinatal factors on microstate duration at 0 mCA for infants born preterm. Impact of (a) Sex, considering the average duration over all MS; (b) Groups of GA at birth for specific MS of REM sleep. Asterisks represent significant differences between Sex and GA groups (**p < 0.005, *p < 0.05).