Abstract
Internalizing problems are associated with a wide range of adverse outcomes. While we have some understanding about risk factors (e.g., neurodevelopmental conditions), biological markers are not well understood. Here, we used deep learning to predict cross-sectional (N=14,523) and worsening longitudinal trajectories (N=10,540) of internalizing problems from measures of brain structure. A stratified cross-validation scheme was used, and performance was evaluated using the area under the receiving operating characteristic curve (AUC). The cross-sectional model performed well across the sample, reaching an AUC of 0.80 [95% CI: 0.71, 0.88]. For the longitudinal model, while performance was sub-optimal for predicting worsening trajectories in a sample of the general population (AUC=0.66 [0.65, 0.67]), good performance was reached across individuals with a neurodevelopmental condition (AUC=0.73 [0.70, 0.76]). Deep learning with features of brain structure is a promising avenue for biomarkers of internalizing problems, particularly for individuals who have a higher likelihood of experiencing difficulties.
Competing Interest Statement
AK has a patent for hollyTM (formerly Anxiety Meter) with royalties paid from Awake Labs. AK has received consulting fees from DNAStack and Shaftesbury. EA has received grants from Roche and Anavex, served as a consultant to Roche, Quadrant Therapeutics, Ono, and Impel Pharmaceuticals, has received in-kind support from AMO Pharma and CRA-Simons Foundation, received royalties from APPI and Springer, received an editorial honorarium from Wiley, and has a patent for hollyTM (formerly Anxiety Meter). The remaining authors have no potential conflicts of interest to report.
