Grants and Contracts Details
Description
Using Big Data to Comprehensively Delineate the Neurobehavioral Phenotype of Children with
Neurofibromatosis Type 1
Background: Children with neurofibromatosis type 1 (NF1) tend to exhibit more neurobehavioral problems
including cognitive impairments, learning disabilities, behavioral problems, and socioemotional difficulties
compared to typically developing populations. Despite increasing research on the neurobehavioral phenotype of
NF1, there are several critical gaps in the extant literature. First, there is limited knowledge on how
neurobehavioral functioning changes across age. Second, few studies have analyzed the associations among these
neurobehavioral problems and whether there are subpopulations with different neurobehavioral profiles. Third,
predictors of within-group differences in NF1 neurobehavioral problems are understudied with most studies
focusing on comparing children with NF1 to a control group. Identifying developmental trajectories, phenotypic
subpopulations, and predictors of neurobehavioral functioning by using larger samples and advanced statistical
methods is critical for the effective management and treatment of neurobehavioral problems.
Objective/Specific Aims/Hypothesis: The overall objective is to provide a more comprehensive understanding
of the neurobehavioral phenotype of children with NF1 by pursuing the following aims:
1. To delineate the developmental trajectories of neurobehavioral functioning in children with NF1. Based
on our preliminary findings, we hypothesize that, compared to normative peers, some neurobehavioral domains
(e.g., math, writing, inhibitory control) may develop at a slower pace whereas other neurobehavioral domains
(e.g., reading, verbal IQ) may develop at a similar pace in children with NF1. That is, the functioning gap between
children with NF1 and normative peers would widen across age for certain affected neurobehavioral domains.
2. To analyze how cognitive functioning relates to academic, behavioral, and socioemotional functioning
and explore NF1 subpopulations with different profiles of various neurobehavioral problems. Based on
prior research and our preliminary study, we hypothesize that 1) impairments in some specific cognitive domains
would be particularly related to certain types of academic difficulties, behavioral problems, and socioemotional
problems, and 2) there may be multiple phenotypic subpopulations: Group 1 with high levels of impairments
across all domains; Group 2 with low levels of impairments across all domains; Other groups characterized by
predominant impairments in some domains but not in other domains (e.g., average level of cognitive abilities
combined with high levels of socioemotional problems).
3. To identify predictors of neurobehavioral functioning in children with NF1. Based on prior research and
our preliminary study, we will examine biological factors (i.e., sex, heritability status, and the presence of
plexiform neurofibromas, optic glioma, or T2-hyperintensities) and demographic factors (e.g., family
socioeconomic status, race/ethnicity, nationality). Based on theoretical views and our preliminary data, we
hypothesize more neurobehavioral problems (e.g., lower mean level, a higher likelihood for slower development,
classification as more problematic profiles) in children with inherited NF1 (vs. de novo NF1 mutation), plexiform
neurofibromas, optic glioma, discrete thalamic T2-hyperintensities, or low socioeconomic status.
Study Design: We will use integrative data analysis (IDA)11 to combine existing neurobehavioral datasets of
individuals with NF1 across 13 sites (approximate n = 2183, ages 2-18). These datasets include comprehensive
neuropsychological assessments with common measures of cognitive functioning, academic achievement,
behavioral and socioemotional problems as well as biological and demographic variables. To estimate the
developmental trajectories of neurobehavioral outcomes (Aim 1) and predictors of trajectories (Aim 3), we will
employ time-varying effects modeling (TVEM) to analyze cross-sectional data with age as the time variable given
the lack of available multi-wave longitudinal data. TVEM is an advanced nonparametric statistical approach
providing more precise estimates than traditional growth curve modeling. We will address Aim 2 from a variable-
centered approach using correlational and regression analysis and from a person-centered approach using latent
profile analysis (LPA).
Innovation: The proposed research represents the most extensive and novel study of neurobehavioral functioning
in NF1 to date. It overcomes critical barriers in the field (e.g., small sample size of published NF1 neurobehavioral
studies) by using IDA to create the first “big data” of NF1 neuropsychological assessments. Big data is important
not only for increased statistical power and reproducibility but it also opens up research questions and statistical
methods (such as those in the proposed project, e.g., TVEM, LPA) that are not possible with smaller datasets.
Impact: We address an Area of Emphasis for the 2020 Neurofibromatosis Research Program: Non-Tumor
Manifestations. This study will be critical for guiding future research and patient management and maximizing
the clinical impact of future intervention studies by informing the NF1 field about 1) neurobehavioral
developmental trajectories to help identify sensitive periods of intervention, 2) association of neurobehavioral
functioning across domains and phenotypic subpopulations, 3) biological and demographic predictors of
neurobehavioral outcomes to understand who is more likely to have neurobehavioral problems, and 4) to stimulate
collaborative research efforts in the NF1 scientific community.
Keywords: neurofibromatosis type 1, neurobehavioral phenotype, cognitive function, academic achievement,
learning disabilities, behavioral problems, socioemotional problems, integrative data analysis
Status | Finished |
---|---|
Effective start/end date | 9/1/21 → 8/31/22 |
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.