Prospective longitudinal associations between harsh parenting and corticolimbic function during adolescence

Abstract

Childhood adversity is thought to undermine youth socioemotional development via altered neural function within regions that support emotion processing. These effects are hypothesized to be developmentally specific, with adversity in early childhood sculpting subcortical structures (e.g., amygdala) and adversity during adolescence impacting later-developing structures (e.g., prefrontal cortex; PFC). However, little work has tested these theories directly in humans. Using prospectively collected longitudinal data from the Fragile Families and Child Wellbeing Study (FFCWS) (N = 4,144) and neuroimaging data from a subsample of families recruited in adolescence (N = 162), the current study investigated the trajectory of harsh parenting across childhood (i.e., ages 3 to 9) and how initial levels versus changes in harsh parenting across childhood were associated with corticolimbic activation and connectivity during socioemotional processing. Harsh parenting in early childhood (indexed by the intercept term from a linear growth curve model) was associated with less amygdala, but not PFC, reactivity to angry facial expressions. In contrast, change in harsh parenting across childhood (indexed by the slope term) was associated with less PFC, but not amygdala, activation to angry faces. Increases in, but not initial levels of, harsh parenting were also associated with stronger positive amygdala-PFC connectivity during angry face processing.

Keywords: adversity; amygdala; corticolimbic; harsh parenting; socioemotional.

Figure 1.

Implicit emotional faces matching paradigm Note. This event-related task design included 100 trials, 20 each of the following facial expressions: angry, fearful, sad, neutral, and happy. Total task time was 8.75 minutes.

Figure 2.

Individual observed values of harsh parenting across childhood. Note. N = 4,144. Spaghetti plot of individual observed values of harsh parenting at ages 3, 5, and 9 years. Group average trajectory depicted in bolded orange. Model fit: X2 (1) = 5.62, p = .02; RMSEA = .03, 90% CI (.01, .06); CFI = 1.00, TLI = .99, SRMR = .01. Loadings for the latent slope factor were specified as 0 (age 3), 2 (age 5), and 6 (age 9), and all loadings for the latent intercept factor were set equal to 1.

Figure 3.

Harsh parenting in early childhood and increases in harsh parenting across childhood are associated with lesser corticolimbic activation during angry face processing. Note. N = 159. Results are from the most stringent models that control for youth gender, race, pubertal status, the intercept/slope term, and harsh parenting at age 15. (A) Greater harsh parenting in early childhood (i.e., the intercept from a linear growth curve model, set at age 3) associated with lesser left amygdala reactivity to angry versus neutral faces; (x, y, z) = (−26,−4,−22), t = 3.91, k = 16. (B). Increases in harsh parenting from ages 3 to 9 associated with lesser right dorsal anterior cingulate reactivity to angry versus neutral faces; (x, y, z) = (6,32,22), t = 3.66, k = 50.

Figure 4.

Increases in harsh parenting across childhood are associated with more positive left amygdala-left prefrontal cortex connectivity during angry face processing. Note. N = 159. BA = Brodmann’s Area. Results are from the most stringent models that control for youth gender, race, pubertal status, the intercept/slope term, and harsh parenting at age 15. (A) Increases in harsh parenting from ages three to nine (i.e., the slope) associated with more positive left amygdala-left orbitofrontal (BA10) connectivity during angry face processing in adolescence; (x, y, z) = (−8,46,16), t = 3.75, k = 40; (b) Increases in harsh parenting from ages three to nine (i.e., the slope) associated with more positive left amygdala-left mPFC (BA9) connectivity during angry face processing in adolescence; (x, y, z) = (−12,58,32), t = 4.26,%"k = 27; (C) Image of identified clusters in (A, in red) and (B, in blue).