Inter-subject synchrony as an index of functional specialization in early childhood

Abstract

Early childhood is a time of significant change within multiple cognitive domains including social cognition, memory, executive function, and language; however, the corresponding neural changes remain poorly understood. This is likely due to the difficulty in acquiring artifact-free functional MRI data during complex task-based or unconstrained resting-state experiments in young children. In addition, task-based and resting state experiments may not capture dynamic real-world processing. Here we overcome both of these challenges through use of naturalistic viewing (i.e., passively watching a movie in the scanner) combined with inter-subject neural synchrony to examine functional specialization within 4- and 6-year old children. Using a novel and stringent crossed random effect statistical analysis, we find that children show more variable patterns of activation compared to adults, particularly within regions of the default mode network (DMN). In addition, we found partial evidence that child-to-adult synchrony increased as a function of age within a DMN region: the temporoparietal junction. Our results suggest age-related differences in functional brain organization within a cross-sectional sample during an ecologically valid context and demonstrate that neural synchrony during naturalistic viewing fMRI can be used to examine functional specialization during early childhood – a time when neural and cognitive systems are in flux.

Cortical temporal hierarchy is immature in middle childhood

“…However, given the fact that we also see group differences in low-frequency power, idiosyncratic long timescale processing may be unlikely since the power analysis is not affected by between-subject variability. These within-and between-group ISC findings in the current study are consistent with our previous work that suggests greater neural synchrony within the DMN in adults compared to children and that neural synchrony in the left TPJ becomes more 'adult-like' with age across childhood (Moraczewski et al 2018).…”

“…To index social-cognitive comprehension, we calculated a composite score of mental / (mental + non-mental) episode comprehension to examine regions that show greater ISCs as a function of mental state comprehension beyond general episode comprehension. Rather than using within-group ISC (i.e., adult to adult or child to child correlations), we used the child-to-adult ISCs (Moraczewski et al 2018), a metric also known as neural maturity (Cantlon and Li 2013). For this analysis, we predicted child-toadult ISC as a function of the fixed effects of a covariate (i.e., age and mental comprehension composite score), condition, and covariate by condition interaction, while also controlling for differences between episodes and crossed random effects.…”

“…However, the use of naturalistic stimuli with fMRI to answer developmental questions is growing (Cantlon and Li 2013, Emerson et al 2015, Moraczewski et al 2018, Richardson 2018) due to its similarity to real-world processing (Zaki andOchsner 2009, Hasson and while minimizing task demands and head motion (Vanderwal et al 2015, Vanderwal et al 2018. In a previous naturalistic viewing study, we demonstrated that adults exhibited greater neural synchrony in regions topographically similar to the DMN compared to 4-to 6-year old children and that responses in the left TPJ within the child group became more 'adult-like' as a function of age (Moraczewski et al 2018). Other recent work suggests that functional maturity within the theory of mind network, as well as the pain network, is positively related to anticorrelation between the two networks (Richardson 2018.…”

Untangling the Relatedness among Correlations, Part III: Inter-Subject Correlation Analysis through Bayesian Multilevel Modeling for Naturalistic Scanning

“…Their corresponding effectsb 1k andb 2k are presumably equal, but in the same vein as the practical implementation of subject-specific effects through two separate random-effects components as previously elaborated, the two fixed-effects components ofb 1k andb 2k that are associated with the explanatory variable x would also have to be estimated separately through data duplication. The situation with more than one explanatory variable would be similar except for an expanded form, and this modeling strategy has been applied at the whole-brain voxel level to a few studies in the literature (e.g., Moraczewski et al, 2018;Finn et al, 2018).…”

“…It has been argued that, under a context closer to natural environment, neural responses are more reproducible and reliable than traditional simple repetitive stimuli (Hasson et al, 2010) due to the involvement of extensive cognitive processing (such as working memory, judgment, reasoning, social cognition, etc.). Its adoption has been steadily growing in investigating various aspects of brain functions such as music imagery (Zhang et al, 2017), early childhood development (Moraczewski et al, 2018), personality traits (Finn et al, 2018) and cognitive differences between schizophrenics and controls. For typical task-related designs, the focus is usually on identifying regions activated by an explicit task or condition.…”