Table 1.
All participants demographic and behavioral data.
Demographic data, means and standard deviations for all behavioral measures of interest for included and excluded participants.
Fig 1.
Illustration of major data analysis steps.
A) Group ICA analysis was run on preprocessed subject data, resulting in 29 independent components (C = 29), 6 of which were identified as ICs in the DMN, ECN and SN. GICA1 back-reconstruction was used to estimate the time courses (Ri) and spatial maps (Si) for each subject. B) The sliding window approach was used to estimate dynamic FNC as the series of correlation matrices from windowed portions (W) of Ri, resulting in a concatenated data matrix of all IC-IC paired correlation values over time. C) K-means was performed on the concatenated data matrix as outlined in Allen et al., 2014. The optimal cluster number was k = 4 and each windowed FNC was assigned to a cluster. Each cluster centroid (also known as state) is represented by a correlation matrix. Clustering analysis allows for quantifying dynamic FNC through measures such as Mean Dwell Time and Fraction Time. This figure was adapted from El-Baba and colleagues [60].
Fig 2.
Correlation matrices showing the fraction time and pattern of cross-network functional connectivity (represented using z-scores) of each of the four dynamic FNC states. 74 participants entered State-1, 66 participants entered State-2, 54 participants entered State-3, and 50 participants entered State-4. dDMN = dorsal default mode network; pvDMN = posterior ventral default mode network; pdDMN = posterior dorsal default mode network; rECN = right executive control network; lECN = left executive control network; SN = salience network.
Fig 3.
Association between FNC in dynamic state-1 and attention problems.
Pearson correlations revealed that participants’ T-Scores for Attention Problems (as measured by the YSR) correlated with diminished functional connectivity between the Salience Network and the left Executive Control Network during Dynamic State-1 (No other correlations survived correction for multiple comparisons (FDR; p<0.05).