Jinwoo Lee

News
0012Nov, 2025My second-authored paper was published at Journal of Child Psychology and Psychiatry!
 
Thumbnail of the journal-published paper “Individual differences in effects of stressful life events on childhood ADHD: genetic, neural, and familial contributions”
Thumbnail of the preprint “A more comprehensive and reliable analysis of individual differences with generalized random forest for high-dimensional data: validation and guidelines”.


I have spent much of my time to exploring methodologies that can better analyze individual differences and their underlying factors. About four years ago, I began studying the principles and applications of causal machine learning, exploring its potential to address these questions. As part of this effort, I contributed as the second author to the the paper that examined 1) how childhood stress shapes later ADHD symptoms, and 2) how such effects are jointly moderated by multiple modalities, including genetic vulnerability to psychiatric disorders, structural brain connectivity, and family environment.

Through this work, we were able not only to predict ‘individualized' exposure effects for each child but also to identify key moderators such as parental psychiatric history, the child’s genetic liability for ADHD, and white matter connectivity within cognitive control networks. These findings allowed us to discuss the possibility of early identification and prevention among vulnerable groups. From this process, I felt a certain optimism that psychological and psychiatric research could move from an average treatment–effect paradigm toward a more individual-centered framework.

Nevertheless, several challenges remain in achieving these goals. For instance, the generalized random forest (grf) used in this study sometimes exhibits instability, where results vary even with identical data and parameters due to random initialization. For these reasons, rather than stopping at applying existing methods to data, my colleagues and I have spent considerable time discussing how to enhance analytic rigor. We recently released a simulation-based validation and real-world tutorial example as a preprint, which is now under review.

Through such methodological elaboration, I hope that the subtle, individual-specific characteristics, often dismissed as “noise”, can be more systematically and meaningfully studied!









© 2025 Jinwoo Lee. All rights are reserved.