Brain Connectome Mapping of Complex Human Traits and Their Polygenic Architecture Using Machine Learning

Area of Science:

• Neuroscience
• Genetics
• Psychiatry

Background:

• Mental disorders and individual characteristics are complex traits with largely unknown neuronal bases.
• Their genetic architectures are highly polygenic and overlapping, leading to heterogeneous phenotypic expression and clinical overlap.
• Brain network analysis offers a noninvasive method to study biological heterogeneity but faces challenges in clinical applications for mental health and genetics.

Purpose of the Study:

• To investigate the predictive power of brain connectomics for individual differences in cognitive abilities and mental health traits.
• To explore the relationship between functional magnetic resonance imaging (fMRI)-based brain connectivity and traits like educational attainment, fluid intelligence, depression, anxiety, and neuroticism.
• To assess the potential of brain network analysis in understanding the genetic underpinnings of these complex traits.

Main Methods:

• A machine learning approach was used to predict individual scores for educational attainment, fluid intelligence, depression, anxiety, and neuroticism.
• Functional magnetic resonance imaging (fMRI)-based static and dynamic temporal synchronization between large-scale brain network nodes were utilized.
• Data from 10,343 healthy individuals from the UK Biobank were analyzed, including predictions for polygenic scores related to neuroticism and schizophrenia.

Main Results:

• High prediction accuracy was achieved for age and sex, validating the biological sensitivity of connectome-based features.
• Above chance-level prediction accuracy was observed for educational attainment and fluid intelligence, primarily associated with negative static brain connectivity in frontal and default mode networks.
• Prediction accuracy for depression, anxiety, neuroticism, and polygenic scores was at chance level, indicating limited predictive power for these traits.

Conclusions:

• Brain connectomics show potential for predicting certain individual characteristics like educational attainment and fluid intelligence.
• The findings suggest distinct neural mechanisms underlying cognitive abilities versus mental health traits.
• This study provides a benchmark for future research integrating brain connectomics, genetics, and individual/mental health traits.