Developmental trajectory of MEG resting-state oscillatory activity in children and adolescents: a longitudinal reliability study

Abstract

Neural oscillations may be sensitive to aspects of brain maturation such as myelination and synaptic density changes. Better characterization of developmental trajectories and reliability is necessary for understanding typical and atypical neurodevelopment. Here, we examined reliability in 110 typically developing children and adolescents (aged 9-17 years) across 2.25 years. From 10 min of magnetoencephalography resting-state data, normalized source spectral power and intraclass correlation coefficients were calculated. We found sex-specific differences in global normalized power, with males showing age-related decreases in delta and theta, along with age-related increases in beta and gamma. Females had fewer significant age-related changes. Structural magnetic resonance imaging revealed that males had more total gray, subcortical gray, and cortical white matter volume. There were significant age-related changes in total gray matter volume with sex-specific and frequency-specific correlations to normalized power. In males, increased total gray matter volume correlated with increased theta and alpha, along with decreased gamma. Split-half reliability was excellent in all frequency bands and source regions. Test-retest reliability ranged from good (alpha) to fair (theta) to poor (remaining bands). While resting-state neural oscillations can have fingerprint-like quality in adults, we show here that neural oscillations continue to evolve in children and adolescents due to brain maturation and neurodevelopmental change.

Keywords: magnetoencephalography; neurodevelopment; reliability; resting state; spectral power.

Fig. 1

Rest task design. Resting-state data were collected with 5-min eyes closed and 5-min eyes open fixated on a white cross. Split-half reliability determined absolute agreement in MEG spectral power within each year. Test–retest reliability determined consistency in MEG spectral power across 3 annual visits.

Fig. 2

Eyes-closed global normalized power. Individual and group mean trajectories with 95% prediction bands for eyes-closed global normalized power by frequency band. Solid line fit indicates significant relationship between power and age. Dashed line fit indicates that the relationship between power and age was not significant. Males had significant linear age-related decreases in delta and theta, while also exhibiting significant age-related linear increases in beta and gamma normalized power. Females had no significant age-related changes in any frequency band. Different developmental trajectories exist between males and females in delta, beta, and gamma normalized power.

Fig. 3

Eyes-open global normalized power. Individual and group mean trajectories with 95% prediction bands for eyes-open global normalized power by frequency band. Solid line fit indicates significant relationship between power and age. Dashed line fit indicates that the relationship between power and age was not significant. Males had significant linear age-related decreases in delta, while also exhibiting significant age-related linear increases in beta and gamma normalized power. Females had significant quadratic age-related changes in alpha, while also exhibiting significant age-related linear increases in gamma normalized power. Different developmental trajectories exist between males and females in delta and beta normalized power.

Fig. 4

Structural MRI by age and sex. Mean volume by age for females and males, along with group average values by sex in A) total gray matter, B) subcortical gray matter, and C) cortical white matter. Group mean shown with 95% confidence bands. Solid line fit indicates significant relationship between volume and age. Dashed line fit indicates that the relationship between volume and age was not significant. Males had significant age-related change in total gray matter, subcortical gray matter, and cortical white matter volume, while females had significant age-related change in total gray matter volume. Different developmental trajectories exist between males and females in total gray matter volume and cortical white matter volume. Males also had significantly more total gray matter volume, subcortical gray matter volume, and cortical white matter volume than females. Asterisks denote p < 0.001.

Fig. 5

Split-half reliability. ICCs of source space normalized power during the eyes-closed and eyes-open resting state were calculated within-session to compare first-half and second-half session averages. ICC values were calculated for each frequency band and region separately. Within-session ICC values ranged from 0.83 to 0.99 (excellent) across all frequency bands. Individual figures are plotted with frequency-specific data ranges to show slight regional differences. In typically developing children and adolescents, within-session reliability was excellent in all frequency bands and source regions during an eyes-closed state (ICC range = 0.83–0.99) and during an eyes-open state (ICC range = 0.85–0.99).

Fig. 6

Test–retest reliability. ICCs of source space normalized power during the eyes-closed and eyes-open resting-state were calculated across 3 yearly 10-min sessions. ICC values were calculated for each frequency band and region separately. Within-session ICC values ranged from 0.04 to 0.75 across all frequency bands. In typically developing children and adolescents, between-year reliability was highest in the alpha frequency band, ICC range = 0.53–0.75 during eyes-closed and ICC range = 0.48–0.65 during eyes-open resting state.