Jeung-Hyun Lee bio photo

Jeung-Hyun Lee

Stress & Computational psychiatry, Decision-making neuroscience

Email Twitter LinkedIn Github

Projects

Memory bias and greater preference for smoking-associated contexts in smokers under acute stress

image-center

  • Manuscript in preparation
  • Poster presented at the 2022 Annual meeting of Society of Biological Psychiatry (New Orleans, LA)
  • Like most addictive substances, nicotine dependence and smoking addiction is also known for its high relapse rate. Here, stress is a huge risk factor of relapse and increased craving to smoke. Referring to the phrase of “chasing the first high”, this project aimed to test the role of acute stress bringing up the drug-related memory, especially the rewarding ones and that would contribute to increased craving. By investigating the process of which context memories are retrieved and preferred in episodes associated with cigarettes, we hope the findings to provide a deeper understanding of stress-induced relapse.


Identifying phenotypes for body-image satisfaction and healthy outcome for obestiy: using machine-learning approach

image-center

  • Manuscript in preparation
  • Team: Mina Kwon, Rose Chang, Jaeyeong Yang
  • We recruited over 1,000 participants from 16 branches nationwide with 365mc clinic, which is a liposuction hospital located in South Korea. This project aimed to investigate the significant variables that predict successful weight loss or maintenance. The battery we applied includes computational tasks and models on mobile applications and traditional psychometric measures.


Contribution of Early-Life Stress and Genetic factors to the young brains: a machine-learning approach with ABCD dataset

  • Team: Hyeonjin Kim
  • Related paper: Joo et al (2022)
  • There has been increasing evidence highlighting the gene x environment interaction in understanding the individual difference that mediates associations between early life stress (ELS) and later development of psychopathology. In this project, we used elastic net to distinguish significant predictors of psychopathology development in different ELS groups. Different data from the ABCD dataset including fMRI task results, clinical diagnosis, and interview answers were included as predictors.

Acute stress alters social discounting rate: a model-based fMRI study

image-center

  • Manuscript in preparation
  • Team: Kunil Kim (LSD lab, 1st author)
  • This project investigated 1) if cortisol is indeed a prosocial hormone that promotes sharing with close and distant others among males and 2) disentagle different roles of parts in medial prefrontal cortex (mPFC) to compuate values of onself and close others with fMRI and behavioral modeling.
  • Poster presented at the 2022 Society for Neuroscience (San Diego, CA).


From image to emotion: Multi-label image classification based on the emotions represented in images

  • Project paper for 2021 MLVU course (SNU)
  • Expressing oneself with images is prevalent online these days, perhaps even more than texts Millions of new images are posted on social network services everyday. Enterprises are eager to collect information from such posts and try to follow the preference trend. As a part of 2021 Machine Learning for Visual Understanding course (SNU, Instructor: Joonseok Lee), our team developed a multi-label image classification utilizing triplet loss embeddings on the Pittsburgh advertisement image database.
  • Team: Youngeun Choi, Junghyun Ryu