Attenuated reconfiguration of functional networks in extremely preterm-born children and adolescents

1515 

Abstract Submission 

Maksym Tokariev¹, Virve Vuontela¹, Anton Tokariev², Piia Lönnberg², Sture Andersson², Helena Mäenpää², Marjo Metsäranta², Aulikki Lano², Synnöve Carlson¹

¹Aalto University, Espoo, Finland, ²Helsinki University Hospital and University of Helsinki, Helsinki, Finland

Maksym Tokariev  
Aalto University
Espoo, Finland

Virve Vuontela  
Aalto University
Espoo, Finland

Anton Tokariev  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Piia Lönnberg  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Sture Andersson  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Helena Mäenpää  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Marjo Metsäranta  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Aulikki Lano  
Helsinki University Hospital and University of Helsinki
Helsinki, Finland

Synnöve Carlson  
Aalto University
Espoo, Finland

Large-scale brain networks undergo significant structural and functional developments since the third trimester of gestation and continue to mature into adulthood. These networks support a broad repertoire of higher cognitive functions, and their functional organization also reconfigures in response to external sensory stimuli and task performance. Extremely preterm birth (< 28 weeks of gestation) poses a high risk for the proper formation of the brain networks which can be reflected in altered functional connectivity (FC).
We collected functional magnetic resonance imaging (fMRI) data from extremely preterm-born and term-born children and adolescents during two brain-states, resting-state and performance of visuospatial n-back tasks, to investigate whether the FC of cognitive networks differs between the two groups. We aimed to identify networks that reconfigure FC between the brain-states differently between the groups. We also evaluated whether the FC strength of the networks associates with task performance.

Participants were extremely preterm-born (n=24, 15 males, mean age at fMRI 10.3 ± SD 3.2 y) and healthy term-born (n=22, 11 males, mean age 9.5 ± SD 2.5 y) children and adolescents with normal global cognitive performance. The fMRI data were preprocessed using a volume-based pipeline from CONN 21a toolbox, which included realignment, slice-time correction, and indirect normalization of functional and structural images to standard MNI152 space. The confounding effects, including motion parameters, were regressed out from the timeseries, and band-pass filtration within frequency band of 0.01-0.15 Hz was performed. The Schaefer atlas was used to parcellate the brain into 400 regions that were assigned to seven networks including the default mode (DMN), frontoparietal (FPN), visual (VN), dorsal attention (DAN), ventral attention, somatomotor, and limbic networks. For each subject, the matrices of resting-state and task FC were computed as pairwise Pearson's correlations between time courses of all regions. Differences in FC between the groups and brain-states, and the associations between FC strength and n-back task performance were analyzed using network-based statistics (significance level 0.05, followed by FWE-correction).

Statistical analysis produced a significant group-by-state interaction in DMN (p = 0.016, η2 = 0.22), DAN (p = 0.016, η2 = 0.23), and VN (p=0.022, η2 = 0.21). Controls had stronger FC during resting-state compared to tasks in both DMN (t21 = 14.54, p < 0.001) and VN (t23 = -0.72, p = 0.45), while in preterms the FC strength did not differ between the brain-states. In DAN, controls had stronger FC during tasks than during resting-state (t21 = -5,79, p < 0.001), while the opposite was observed in preterms (t23 = 3.18, p = 0.02).
Controls exhibited significant associations between task performance and network FC strength during tasks. Fewer incorrect responses were associated with stronger FC in DAN (R = -0.708, p = 0.026) and VN (R = -0.751, p = 0.028). Shorter response times associated with stronger FC in FPN (R = -0.848, p = 0.037) and weaker FC in DAN (R = 0.583, p = 0.017) and VN (R = 0.607, p = 0.034). No significant associations between performance and FC were observed in preterms.

We found that in preterms, the FC strength during task performance was either reduced or comparable to that during resting-state in several networks supporting cognitive performance. Moreover, preterms did not show significant associations between network FC and task performance. Controls, on the other hand, had a larger change in FC strength between the brain-states, and the FC strength of several networks during tasks was associated with better performance. Together these results suggest that flexible reconfiguration of network FC is important for successful performance of tasks that require attention and working memory, and this ability is reduced in extremely preterm-born children and adolescents.

Working Memory

Early life, Adolescence, Aging ²

Connectivity (eg. functional, effective, structural) ¹

Development

FUNCTIONAL MRI

Other - preterm birth