Poster #5 - Early adolescent cortical thickness examined through two modeling approaches: Exploring sex and non-linear age effects

• In Event: Poster Session 08

Abstract

The childhood-to-adolescence transition is a notable period of change marked by pubertal development (e.g., Marshall & Tanner, 1969), which is thought to have widespread effects on neurodevelopment that may differ between male and females (Herting & Sowell, 2017; Raznahan et al., 2010). This study aimed to examine the effect of pubertal stage on cortical thickness (CT), both as a main effect and as it interacts with sex, in late childhood and early adolescence using two modeling frameworks. First, we examined these associations while controlling for chronological age. Second, we examined these associations while modeling pubertal stage and CT as potentially changing non-linearly across age.

In this project, we used data from two waves of the Adolescent Brain and Cognitive Development Study (Baseline ages 9-10 and Year 2 [Y2] ages 11-12), which resulted in a sample of 9,985 participants. To examine age-invariant effects, linear mixed-effect regression models examined the association between pubertal development, as a main effect and as it interacts with sex, to predict CT across 34 FreeSurfer regions of interest (see Casey et al., 2018). Penalized function-on-function regressions were then used to model the impact of puberty, and its interaction with sex, on CT across 34 regions of interest as a potentially non-linear function of age. Each of these models also controlled for age, sex, race, ethnicity, total family income, parental education, and marital status and controlled for family-wise error through Bonferroni correction of p-values.

There were no main effects of puberty on CT when assuming age- and sex-invariance. However, pubertal stage interacted with sex to significantly predict CT in 15 ROIs; These effects showed that pubertal stage is inversely associated with CT, and that this relationship is stronger for female youth than for male youth (Figure 1a). When modeled as a function of age, pubertal stage demonstrated sex-invariant effects on CT in seven regions, located primarily in frontal brain regions (Figure 1b). Significant sex by puberty interactions to predict regional CT modeled as a function of age were identified in five ROIs (Figure 1b). In general these results demonstrate that the effects of puberty on CT are negative and become stronger in the negative direction with age (Figure 2) and puberty by sex interactions generally show that these effects are stronger for female youth than male youth (Figure 2b).

Our results underscore the spatially heterogeneous nature of the relationship between puberty and brain development, with sex differences found in pubertal effects on CT in posterior brain regions and sex-adjusted, but age-variant, pubertal effects on CT found almost entirely in frontal brain regions. Developmental timing may be one factor driving this identified pattern, as posterior brain regions generally mature earlier in the life course (Gogtay et al., 2004; Tamnes et al., 2010) and frontal regions undergo protracted maturation well into early adulthood (Giedd et al., 1999; Tamnes et al., 2010). Taken together, these results indicate that, in this age range, puberty is most strongly associated with regional CT in females, which may have implications for adolescent social development.

Authors

• Andrea Wiglesworth, University of Minnesota - Twin Cities
  Presenting Author

• Mark B. Fiecas, University of Minnesota

• Meng Xu

• Aidan T Neher, University of Minnesota Twin Cities

• Laura Padilla, University of Minnesota Twin Cities

• Katherine Ann Carosella, University of Minnesota - Twin Cities

• Donovan J Roediger, University of Minnesota Twin Cities

• Bryon Mueller, University of Minnesota

• Bonnie Klimes-Dougan, University of Minnesota - Twin Cities

• Kathryn Cullen, University of Minnesota