Current Approaches in Computational Psychiatry for the Data-Driven Identification of Brain-Based Subtypes

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Current psychiatric diagnostic systems face challenges in accurately defining patient groups and guiding effective treatments, leading to moderate intervention success and high relapse rates.
• The search for improved methods to stratify patients into homogeneous subgroups is crucial for better diagnosis, evaluation, and personalized treatment development.

Purpose of the Study:

• To review empirical brain-based subtyping studies in child, adolescent, and adult psychiatric populations published between 2019 and March 2022.
• To assess the current state of methodological exploration, validation, and clinical utility in brain-based subtyping research.

Main Methods:

• Systematic review of peer-reviewed literature on brain-based subtyping in psychiatry.
• Analysis of studies focusing on neuroimaging data for patient stratification.
• Evaluation of methodologies, validation strategies, and clinical relevance of identified subtypes.

Main Results:

• Significant methodological innovation in brain-based subtyping techniques.
• A lack of rigorous validation for reproducibility and clinical utility of discovered subtypes.
• Evidence of a disconnect between methodological advancements and clinical objectives, termed a 'collaboration crisis'.

Conclusions:

• Brain-based subtyping holds considerable promise for developing biologically grounded, targeted psychiatric treatments.
• Addressing the identified shortcomings in validation and clinical integration is essential for realizing the full potential of this approach.
• Strengthening collaboration between methodologists and clinicians is critical for advancing psychiatric nosology and treatment development.