Research Focus

Mechanisms of Repetitive Negative Thinking and Contrast Avoidance

Repetitive negative thinking—such as rumination and worry—is a common, harmful pattern that amplifies cognitive, physical, and interpersonal difficulties and elevates risk for affective disorders (major depression, bipolar disorder, and anxiety). Many individuals find it difficult to disengage from these thoughts, yet the mechanisms that sustain them remain unclear. The Contrast Avoidance Model proposes that some people are especially sensitive to sharp emotional downturns—negative emotional contrasts—and use repetitive negative thinking to avoid abrupt shifts. Paradoxically, this strategy can also produce positive emotional contrasts when outcomes exceed expectations. My work aims to: (1) empirically test the model’s validity, (2) examine its applicability to bipolar-spectrum and related affective disorders, and (3) design and validate mechanism-targeted interventions that directly reduce contrast-avoidant processes.

Representative Research

• Kim, H., McInnis, M. G., & Sperry, S. H. (2024). An initial test of the Contrast Avoidance Model in bipolar spectrum disorders. Journal of Psychiatric Research.

• Kim, H., & Newman, M. G. (2023). Worry and Rumination Enhance Positive Emotional Contrast Based on the Framework of the Contrast Avoidance Model. Journal of Anxiety Disorders.

• Kim, H., & Newman, M. G. (2022). Avoidance of negative emotional contrast from worry and rumination: An application of the contrast avoidance model. Journal of Behavioral and Cognitive Therapy.

• Newman, M. G., Schwob, J. T., Rackoff, G. N., Van Doren, N., Shin, K. E., & Kim, H. (2022). The naturalistic reinforcement of worry from positive and negative emotional contrast: Results from a momentary assessment study within social interactions. Journal of Anxiety Disorders.

• Newman, M. G., Rackoff, G. N., Zhu, Y., & Kim, H. (2022). A transdiagnostic evaluation of contrast avoidance across generalized anxiety disorder, major depressive disorder, and social anxiety disorder. Journal of Anxiety Disorders.

• Kim, H. & Newman, M. G. (2019). The paradox of relaxation-induced anxiety and mediation effects of contrast avoidance in generalized anxiety disorder and major depressive disorder. Journal of Affective Disorders.

• Newman, M. G., Cho S., & Kim H. (2017). Worry and generalized anxiety disorder: A review. Reference module in Neuroscience and Biobehavioral Psychology.

Clinical AI for Affective Science

Suicidal ideation emerges from dynamic interactions among depressive symptoms, yet conventional prediction often treats symptoms as independent and misses these dependencies. I address this gap with the Network-Augmented Machine Learning Utility (NAMU; “tree” in Korean), which integrates network-derived features (e.g., symptom centrality, pairwise edge strength) into machine-learning models to enhance both interpretability and predictive performance. Using PHQ-9 data from a nationally representative U.S. cohort (NHANES; N = 44,922), I engineered 37 network features from eight depressive symptoms and achieved strong performance, clarifying how specific symptoms, and their interconnections, contribute to suicide risk. Building on this foundation, I am extending NAMU to bipolar-spectrum populations and assessing cross-cultural generalizability with Korean national data (KNHANES). In parallel, I am preparing EHR-based implementations (IMPACT grant under consideration) to move toward deployable, clinically actionable decision support. This program of work illustrates how combining symptom-network analytics with machine learning can advance early detection and targeted prevention in psychiatric care.

Representative Research

• Kim, H., Yocum, A., McInnis, M., & Sperry, S. H. (2025, September). Predicting suicidal ideation from depression screening data: A network‑augmented machine learning approach. Manuscript under review at Nature Mental Health.

• Kim, H., Choi, K. (2025, November). Inferring Suicidal Ideation Without Direct Questions: A Network-Augmented Machine-Learning Analysis of Depression Screening Data. Manuscript invited for submission at Nature Health.

Person-Specific Dynamics in Affective Disorders

Affective disorders vary not only across groups but, crucially, within individuals over time. My work characterizes person-specific (idiographic) symptom patterns and dynamics, linking them to personalized assessment and care. Methodologically, I integrate intensive longitudinal designs, time-varying models, and network-based analytics to: (1) identify individual variations in symptom presentation and interrelations; (2) map time-varying trajectories and tipping points that signal worsening or recovery; and (3) develop and validate personalized intervention strategies that leverage these idiographic signatures. Group-level (nomothetic) differences by sex or life stage are examined as moderators of these person-specific processes to inform targeted resource allocation and timely clinical decision-making.

Representative Research

• Kim. H. (2024). Sex differences in age-varying trends of depressive symptoms, substance use, and their associations among South Korean adults: A Time-Varying Effect Modeling (TVEM) analysis of a nationwide sample. Journal of Affective Disorders. 

• Kim. H., McInnis, M. G., & Sperry. S. H. (2024). Longitudinal dynamics between anxiety and depression in bipolar spectrum disorders. Journal of Psychopathology and Clinical Science, 133(2), 129-139.

• Jo, D., & Kim, H. (2023). Network analysis of depressive symptoms in South Korean adults: Similarities and differences between women and men. Current Psychology, 1-12.

• Kim, H., & Duval, E. R. (2022). Social anxiety and depression symptoms are differentially related in men and women. Current Psychology. 1-12.

• Kim, H., Rackoff, G. (Co-first author), Fitzsimmons-Craft, E., Shin, K., Zainal, N., Schwob, T., Eisenberg, D., Wilfley, D., Taylor, C., & Newman, M. (2022). College mental health before and during the COVID-19 pandemic: Results from a nationwide survey. Cognitive Therapy and Research.

Methodologies for Affective Disorder Research

My program develops standardized, reproducible, and scalable methods to elicit, measure, and compare affective processes implicated in mood and anxiety disorders. Beyond building potent stimuli, I emphasize procedure-level standardization (to minimize experimenter effects), measurement efficiency (short forms with preserved validity), and generalizability across clinical populations.

Representative Research

• Sperry, S. H., Smith, J. L., Sandorffy, B. L., Murphy, V. A., Kim, H., Van Rheenen, T. E., Dodd, A. (2025). Development and validation of a short version of the Hypomanic Attitudes and Positive Predictive Inventory (HAPPI): Measuring extreme appraisals of internal states and bipolar risks, Journal of Affective Disorders.

• Kim, H., & Newman, M. G. (2024). Development and validation of novel worry and rumination induction methods, Under review at the Jounral of Anxiety Disorders.

• Kandemir, B., Kim, H., Newman, M.G., Adams, R.B., Li, J., & Wang, J.Z. Demographic differences and biases in affect evoked by visual features. (2023). In Wang, J. Z., & Adams, R. B. (Eds.), Modeling visual aesthetics, emotion, and artistic style. New York, NY: Springer.

• Kim, H., Lu, X., Costa, M., Kandemir, B., Adams, R. B., Li, J., Wang, J. Z., & Newman, M.G. (2018). Development and validation of Image Stimuli for Emotion Elicitation (ISEE), Psychiatry Research.