28 A Graph Theoretical Approach to Understanding Associations between Structural Connectivity and Improvements in Behavior of Children Born Very Preterm Following a Positive Parenting Intervention

Abstract

Objective:

Children born very preterm (VPT; <32 weeks gestation) are at increased risk for long-term neurocognitive sequelae such as behavioral problems. These problems may be caused by disrupted brain development, particularly white matter abnormalities that affect network efficiency, as shown via diffusion magnetic resonance imaging (dMRI). There is evidence that short-term interventions for pediatric clinical populations can lead to behavioral improvements as well as associated neuroplasticity. Adapted from a previous parenting intervention for families of young children with traumatic brain injury, the novel Building Better Brains and Behavior (B4) program teaches responsive parenting skills for families of children born preterm. It is hypothesized that parent-reported externalizing symptoms will improve from pre- to postintervention and that these improvements will be mirrored by an increase in neural efficiency.

Participants and Methods:

VPT children between the ages of 3-8 with documented behavioral problems were recruited to participate in a single-arm pilot clinical trial. Families began with a baseline visit in which the Child Behavior Checklist (CBCL) was administered as a measure of behavior problems, and the child underwent dMRI. Parents then participated in the 7-session intervention integrating self-guided, online learning modules with live virtual coaching sessions with a therapist. Twenty three participants enrolled, 15 of which completed the intervention and baseline MRI scan; 4 children were excluded from analysis due to not meeting eligibility criteria, leaving 11 participants for analysis of intervention effects (8 males, Mage=5.42). At program completion, families returned for a follow-up that entailed another CBCL questionnaire and dMRI scan. Eight children completed the post-intervention scan and five were retained for analysis (4 males, Mage=5.83). Imaging data was analyzed using the Brain Connectivity Toolbox, which generated graph theoretical metrics to characterize the topological organization of anatomical networks.

Results:

A paired samples t-test showed significant reduction of externalizing behavior problems pre-intervention (M=61.12, SD=10.02) to post-intervention (M=55.00, SD=11.62; f(10)=3.09, p=0.01). At baseline, externalizing behavior problems were positively correlated with normalized clustering coefficient, r(10)=0.59, p=0.04, and small-worldness, r(10)=0.64, p=0.03. Change in externalizing symptoms pre- to post-intervention was positively correlated with baseline global efficiency, r(4)=0.94, p=0.02, and negatively correlated with mean local efficiency, r(4)=-0.89, p=0.03, and normalized characteristic path length, r(4)=-0.89, p=0.03.

Conclusions:

Preliminary results indicate that VPT children who exhibit higher levels of externalizing symptoms show higher normalized clustering coefficient (which is expected of networks with less integration), and higher small-worldness (which is unexpected). Greater behavioral improvements were associated with higher baseline characteristic path length as expected, but lower baseline global efficiency; this may indicate that children who had lower global efficiency to begin with benefitted from the intervention the most. Due to the small sample size and lack of corrections for multiple comparisons, these results are not definitive and further research is needed to elucidate associations between structural connectivity and behavioral intervention in children born very preterm.