The Role of Machine Learning in Diagnosing Bipolar Disorder: Scoping Review

Area of Science:

• Neuroscience
• Computer Science
• Medical Informatics

Background:

• Bipolar disorder (BD) is a significant cause of global morbidity and mortality, reducing life expectancy by 9-17 years.
• Misdiagnosis as depression is common, complicating treatment for approximately 60% of BD patients.
• Machine learning (ML) offers advanced techniques for improved BD diagnosis.

Purpose of the Study:

• To systematically review machine learning algorithms utilized for the detection and diagnosis of bipolar disorder and its subtypes.

Main Methods:

• A systematic scoping review following PRISMA-ScR guidelines.
• Searches conducted across Google Scholar, ScienceDirect, and PubMed, with backward screening of references.
• Data extraction and narrative synthesis performed by independent investigators after title/abstract and full-text screening.

Main Results:

• 33 articles were included from 573 retrieved studies.
• Clinical data (58%) was most common, with classification models (55%) being the most used ML approach.
• Magnetic resonance imaging data (34%) was prevalent for classification, achieving up to 98% accuracy.

Conclusions:

• Machine learning models show promise in diagnosing bipolar disorder across diverse patient groups.
• Further research can enhance ML applications for clinical decision support in mental healthcare.
• This review highlights the growing role of ML in psychiatric diagnostics.