Developmental Trajectories of Higher-order Thinking Talk Differ for Typically-developing Children and Children with Brain Injuries

• In Event: Analogy and Cognitive Development

Abstract

Overview
Higher-Order Thinking Talk (HOTT; Frausel, Freeman, Richland et al, 2020) is a form of relational reasoning during communication, defined here as talk in which representations are linked through comparison, inference, abstraction and hierarchy. We examined children’s trajectories of using HOTT in typically-developing children and children with pre- or perinatal brain injuries, observed in naturalistic situations at home. We code children’s spontaneous talk to examine early higher-order thinking skills, identifying instances of structure HOTT (HOTT pertaining to abstract/conceptual relations) as well as surface HOTT (HOTT with more perceptual or surface-level inferences). These analyses provide insight into the relative roles of maturation and neurological foundation, socialization and socioeconomic status.

Research Design
The current project uses a rich set of naturalistic, longitudinal data of children in their home contexts to examine the nature and complexity of thinking and reasoning in children’s spontaneous talk. We test whether the onset and early developmental trajectory of surface and structure HOTT differ for typically developing children (TD) and children with pre- or perinatal lesions (PL).

Participants
For the current study, 64 TD and 46 children with unilateral pre or perinatal brain injury were videotaped in their homes every 4 months between 14 and 58 months while engaging in everyday activities, yielding over 1,500 hours of video.

Data
Data were analyzed using Hierarchical Linear Models (HLM; Raudenbush & Bryk, 2002) to characterize children’s HOTT use across development as a function of their intercept, growth, and acceleration. We report a two-level longitudinal HLM model (with age points at level 1 nested in individuals at level 2), with the number of child surface and structure HOTT utterances at a given age as a Poisson outcome (i.e., log link function), including session length in hours as an exposure variable. The level 1 model accounts for variation within-subjects over time, and the level 2 model accounts for variation between-subjects. Importantly, as shown in Table 1, both PL status and family income predicted amount of HOT use, both for surface and structure HOTT.
In Figure 1, we present model graphs, showing hypothetical trajectories of (a) surface and (b) structure HOTT for PL and TD children at the 25th income percentile ($42,500) and the 75th income percentile ($100,000). While we see main effects of brain-injury and income in both graphs, higher-income PL children have a similar surface HOTT trajectory to lower-income TD children (see two middle lines in panel a). However, TD children at both levels of income are predicted to use more structure HOTT than PL children regardless of family income.

In Summary
Our findings suggest that TD children and children with PL follow different trajectories of HOTT across early childhood—TD children use more HOTT and grow more quickly than children with PL. Income also predicts differences in HOTT such that children from higher-income families have an earlier onset and use more surface and structure HOTT than children from lower-income families. These findings highlight the possibility that these skills are malleable and responsive to parent and social input.

Authors

• Rebecca Frausel, Pennsylvania State University, Erie
  Non-Presenting Author

• Elayne Vollman, Lake Forest College
  Presenting Author

• Antonia Muzard
  Non-Presenting Author

• Lindsey Engle Richland, University of California - Irvine
  Non-Presenting Author

• Susan Goldin-Meadow, University of Chicago
  Non-Presenting Author

• Susan Cohen Levine, University of Chicago
  Non-Presenting Author