# Delineating neurobiological heterogeneity in internalizing symptoms using machine learning and deep phenotyping

> **NIH NIH R00** · VANDERBILT UNIVERSITY · 2021 · $237,045

## Abstract

Symptom-based classification approaches based on the DSM 5 are often not supported by
epidemiological, genetic, and clinical neuroimaging research and may impede the advancement of
interventions that target the pathophysiological mechanisms underlying mental health disorders. A novel
alternative to classifying psychopathology based on presenting clinical symptoms is to identify
neurobiologically-informed biotypes. Individuals are clustered according to shared patterns of brain dysfunction
using data-driven machine learning techniques to reveal the heterogeneous biological mechanisms that
underlie comorbid disorders. Internalizing symptoms often first begin during development, suggesting that this
is a critical period of vulnerability. Additionally, strong sex differences are found in anxiety and depressive
symptoms, starting in adolescence. Thus, studies are needed that examine sex differences in the
neurobiological mechanisms associated with internalizing symptoms during development. The purpose of the
current study is to uncover the neurobiological heterogeneity associated with internalizing symptoms in youth.
During the K99 phase, Aim 1 will use machine-learning techniques to delineate patterns of neurobiological
heterogeneity among youth with anxiety and depressive disorders using multimodal neuroimaging data from a
large community-based sample of over 1,200 youth studied as part of the Philadelphia Neurodevelopmental
Cohort (PNC; Training phase). We will test these heterogeneous patterns on a hold-out sample from the same
cohort to examine the model’s validity (Validation phase). While the PNC provides an ideal dataset for
developing a model, it does not have paradigms relevant to fear and anxiety that would allow us to identify
important phenotypic differences between biotypes. Thus, Aim 2 will evaluate the generalizability of this model
in an independent sample collected during the R00 phase, and further characterize these biotypes using
pertinent measures related to error and reward processing. Finally, Aim 3 will investigate how sex differences
in brain development associate with heterogeneous neural patterns in internalizing symptoms. Dr. Kaczkurkin’s
long-term goal is to establish an independent research program where she will use advanced multi-modal
neuroimaging techniques to study the mechanisms underlying internalizing disorders in youth. This study will
provide a unique opportunity to capitalize on the PNC database at the University of Pennsylvania to develop a
well-validated model while also collecting a refined independent dataset, which will provide Dr. Kaczkurkin with
the training and experience needed to transition to an independent research career.

## Key facts

- **NIH application ID:** 10227979
- **Project number:** 5R00MH117274-05
- **Recipient organization:** VANDERBILT UNIVERSITY
- **Principal Investigator:** Antonia Kaczkurkin
- **Activity code:** R00 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2021
- **Award amount:** $237,045
- **Award type:** 5
- **Project period:** 2018-07-01 → 2023-06-30

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10227979

## Citation

> US National Institutes of Health, RePORTER application 10227979, Delineating neurobiological heterogeneity in internalizing symptoms using machine learning and deep phenotyping (5R00MH117274-05). Retrieved via AI Analytics 2026-05-25 from https://api.ai-analytics.org/grant/nih/10227979. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*