Korean researchers to use AI to predict psychiatric drug-induced weight gain

A research team at Chung-Ang University Hospital will develop a novel artificial intelligence (AI) algorithm aimed at predicting the likelihood of weight gain and the effectiveness of obesity medications in patients taking psychiatric drugs.

Psychiatric medications such as antipsychotics, antidepressants, and mood stabilizers are known to cause side effects, including significant weight gain and metabolic alterations. However, individual susceptibility to weight gain from the same drug varies considerably, and until now, predicting which patients are more likely to gain weight and which anti-obesity drugs might be effective has been nearly impossible.

To address this gap, the team, Professors Kim Sun-mi of the Department of Psychiatry and Lee Hye-jun of the Department of Family Medicine, evaluated weight-related side effects after 24 weeks of administering psychiatric medications to patients with depression, bipolar disorder, or schizophrenia.

Among those who were overweight or obese, the study introduced an additional 24-week course of anti-obesity drugs while maintaining psychiatric treatment. During this period, the researchers monitored changes in body measurements, body composition, lifestyle habits, blood markers, and psychological states.

Building on these clinical data, the team will collaborate with Professor Kim Young-bin of the Department of Artificial Intelligence at Chung-Ang University, to conduct a joint study using deep learning techniques. The AI model will be trained to recognize patterns in how specific psychiatric and anti-obesity drugs both individually and in combination affect weight, obesity, and metabolic parameters.

The ultimate goal is to develop a personalized AI algorithm capable of predicting weight gain risk and treatment efficacy, enabling more precise and patient-specific therapeutic strategies.

“This algorithm will provide crucial support in medical decision-making by helping physicians select the most appropriate anti-obesity therapy early in the treatment process,” Professor Lee said. “It could also prevent obesity-related complications, including metabolic syndrome, cardiovascular disease, and cancer.”

Through interdisciplinary collaboration between medicine and engineering, the team hopes this AI algorithm will allow clinicians to tailor psychiatric treatment plans more safely and effectively from the outset, Lee added.

The results were published in PLOS ONE.