Abstract
Background: Bipolar disorder (B.D.) defines mood disorders that lead to an imbalance in emotional mood and substantially impact the well-being of patients with B.D. Antidepressants are commonly prescribed in patients with B.D. to improve the severity of depressive symptoms. However, several symptoms are associated with long-term antidepressant use, such as metabolic, psychological， cardiological, and sexual problems. Thus, it is essential to manage the use of antidepressants in patients with B.D.
Goal of Study: The study aimed to predict antidepressant use status in patients with B.D. with parameters related to childhood trauma (C.T.) and the severity of B.D. symptoms.
Methods: We utilized publicly available open data from an fMRI database to examine individuals with B.D. These individuals provided signed consent forms for the sharing of their anonymized data. The study included twenty patients with B.D., with fifteen of them being prescribed antidepressants and the remaining five not receiving such medication.
Results: Our primary findings indicated that the RF machine learning model accurately predicted antidepressant status at a rate of 67%. Additionally, the NN machine learning model achieved a prediction accuracy of 50%.
Conclusion: The study’s conclusion highlighted that the RF machine learning model could predict antidepressant usage status in patients with B.D. above the chance level (60%). In contrast, the NN model did not achieve predictions above chance level.
https://www.qeios.com/read/Z4LP8D
https://doi.org/10.32388/Z4LP8D
you can review paper.
Kadir Uludağ. (2024). Predicting Antidepressant Use in Patients with Bipolar Disorder. Qeios. doi:10.32388/Z4LP8D.
