Early-Life Institutionalization and Structural Covariance: Data from the Bucharest Early Intervention Project

• In Event: Poster Session 09
  In Poster Session: PS 09 Topic - Developmental Psychopathology

Abstract

Neural development depends greatly on a variety of experiences and environmental inputs, particularly during sensitive periods (McLaughlin et al., 2017). Consequently, disruptions to the expected environment (e.g., institutionalization), can lead to profound deviations of behavioral, cognitive, and neural development. In particular, children who have experienced early institutionalization are at higher risk of developing psychopathology, experience higher rates of interpersonal problems, and are at increased risk for antisocial behaviors (Bos et al., 2011; Humphreys et al., 2018; Sheridan et al., 2018). It has been hypothesized that institutionalization hinders neural development during sensitive periods in childhood, and in turn, the neural effects of institutionalization underlie the increased risks for psychopathology.
Exploration of additional methodologies is necessary for furthering understanding of the neural mechanisms underlying the impact of psychosocial deprivation on cognitive and behavioral development. Structural covariance analysis involves the measurement of covarying interindividual differences in neural anatomy across groups, and has been posited to improve understanding of the fine-tuning of neural structure that occurs across development (Alexander-Bloch et al., 2013). We will apply this approach to data collected as part of the Bucharest Early Intervention Project (BEIP), which is the first and only known randomized control trial to examine foster care as an intervention for early institutionalization. The BEIP includes 136 institutionalized children (6- to 31-months-old) in Bucharest, Romania who were randomly assigned to a foster care intervention (foster care group, FCG) or to remain in the institution (care-as-usual group, CAUG), as well as Romanian children who were never institutionalized as controls (never-institutionalized group, NIG). Children from all groups completed MRI scans at approximately ages 9 and 16.
Typically developing children follow a nonlinear trajectory of structural covariance network properties across development: in middle childhood they display increased integration and decreased segregation of neural networks, whereas in adolescence there is decreased integration and increased segregation of networks. Thus, at age 9, we expect that the NIG will have increased integration and decreased segregation compared to previously institutionalized children (ever-institutionalized group, EIG). Within the EIG at age 9, we expect the FCG to exhibit increased integration and decreased segregation compared to the CAUG. At age, 16, we expect that the NIG will have decreased integration and increased segregation compared to the EIG, and, within the EIG, we expect the FCG to exhibit decreased integration and increased segregation compared to the CAUG. These findings would demonstrate that early psychosocial deprivation in the form of institutionalization impedes typical maturation of cortical networks into adolescence, and that randomization into foster care reduces the negative impact of institutionalization on structural brain development.
To construct structural covariance networks, the network will first be divided into nodes (brain regions) and edges (statistical similarity of cortical thickness between two regions) (He et al., 2007). We will calculate across-subject correlations between each pair of brain regions across each group of participants (Khundrakpam et al., 2017). In order to compare graph theoretical metrics between groups, we will generate 1,000 bootstrap samples with replacement for each age group (Khundrakpam et al., 2013).

Authors

• Sarah Furlong, University of North Carolina at Chapel Hill
  Presenting Author

• Katie McLaughlin, Harvard University
  Non-Presenting Author

• Mackenzie Woodburn, University of North Carolina at Chapel Hill
  Non-Presenting Author

• Kinjal Patel, University of North Carolina at Chapel Hill
  Non-Presenting Author

• Nathan A Fox, University of Maryland - College Park
  Non-Presenting Author

• Charles H Zeanah, Tulane University
  Non-Presenting Author

• Charles A. Nelson, Harvard Medical School
  Non-Presenting Author

• Margaret A. Sheridan, University of North Carolina at Chapel Hill
  Non-Presenting Author