Data-driven Clustering of fMRI in Children with OCD and Subclinical Symptoms: Characterizing Cognitive Control Network Alterations and Predicting Outcomes of a Digital Cognitive Training

About This Grant

Project Summary: This K01 proposal aims to advance Dr. Dana Díaz's career through a mentored research project and training plan focused on data-driven, brain-based subtyping of children with obsessive-compulsive disorder (OCD) and subclinical obsessive-compulsive symptoms (sOCS) and its application to a novel cognitive control training (CT) intervention. OCD is a chronic, impairing condition affecting 1-2% of youth. Additionally, one in five children experience sOCS, which diminish quality of life and increase the risk of developing OCD and other psychiatric disorders. Impairments in cognitive control, the ability to flexibly adapt thoughts and behaviors, likely contribute to intrusive obsessions and compulsive urges. Middle childhood is a critical developmental period when cognitive control circuits develop, obsessive-compulsive symptoms fluctuate, and many children initiate treatment for OCD. Thus, interventions that enhance cognitive control during middle childhood may be key in improving or preventing symptoms before they become entrenched. CT has shown promise in improving cognitive control and symptom severity in pediatric ADHD, but it has yet to be tested in children with OCD or sOCS. Furthermore, little is known about the neural underpinnings of cognitive control deficits in pediatric OCD and sOCS, and nothing is known about neural predictors of CT response in these populations. Heterogeneity in OCD symptom presentation and still-developing neural networks for cognitive control in child patients may contribute to the mixed findings in pediatric OCD, which have largely relied on group-averaging approaches that do not account for individual differences. To address these limitations, this innovative project will employ data-driven machine learning (ML) clustering to identify unique subtypes of baseline brain activity patterns in cognitive control networks. Children aged 8-12 with OCD and sOCS will complete an fMRI cognitive control task prior to participating in an at-home, 4-week CT intervention. The following aims will be pursued: (Aim 1) Identify ML-defined subtypes of cognitive control network function in pediatric OCD; (Aim 2) Identify subtypes of cognitive control that distinguish between or cut across OCD and sOCS; (Aim 3) Determine which subtypes predict pre-to-post-CT improvements in cognitive control. This K01 project will help define treatment targets and provide pilot data for a future R01 grant. To ensure successful completion of these aims, Dr. Díaz will focus on the following training goals: (Goal 1) Gain expertise in neural substrates of cognitive control in children with OCD and sOCS; (Goal 2) Gain experience conducting clinical trials research for experimental medicine; and (Goal 3) Learn to leverage data-driven machine learning of fMRI data for personalized medicine. These training objectives will lay the foundation for Dr. Díaz's long-term career goals of becoming an independent clinical-translational researcher, using advanced statistical methods to elucidate developmentally sensitive mechanisms of childhood psychopathology, informing novel targets for treatment and prevention.