Data-driven symptom dimensions reveal familial patterns in bipolar disorder

Area of Science:

• Psychiatry
• Genetics
• Biostatistics

Background:

• Bipolar disorder (BD) is a heritable psychiatric condition with complex genetic heterogeneity.
• Traditional diagnostic categories may not fully capture the multidimensional nature of BD symptoms.
• Identifying biologically relevant subtypes is crucial for understanding BD's genetic underpinnings.

Purpose of the Study:

• To evaluate if data-driven dimensions of bipolar disorder exhibit familial aggregation.
• To explore potential genetic influences on these dimensions.
• To assess the utility of dimensional phenotypes in BD research.

Main Methods:

• Principal component analysis (PCA) was used to derive latent dimensions from 21 clinical variables in two independent cohorts (N=368 and N=1356).
• Familial similarity was quantified in the PCA-derived space and compared between relatives and unrelated BD subjects.
• Mixed-effects models assessed the relationship between degree of relatedness and familial similarity on latent dimensions.

Main Results:

• The first two principal components, representing mood episode frequency and age of illness onset, were consistent across both cohorts.
• Relatives showed greater clinical phenotype similarity than unrelated BD cases in both cohorts.
• Significant familial similarity was observed for age of onset (PC2) in both cohorts and for episode frequency (PC1) in one cohort.

Conclusions:

• Latent clinical dimensions in BD, particularly mood episode recurrence and age of onset, aggregate within families.
• These findings suggest that dimensional phenotypes may reflect underlying genetic liability in bipolar disorder.
• Data-driven dimensions offer promising, genetically informative targets for future BD research and neurobiological studies.