Michael A. Powell

This study investigates the heritability of human brain connectomes using a novel method, finding that heritability is insignificant after accounting for network variability.

The paper uses generative network models to analyze bilateral symmetry in a Drosophila larval brain connectome, identifying significant differences in connection probabilities and proposing adjusted definitions of symmetry.

The paper demonstrates how overfitting through hyperparameter tuning can lead to non-generalizable results in machine learning, despite cross-validation, and suggests safeguards to prevent this.

The paper uses diffusion MRI to analyze white matter architecture and shows it can predict social, health, and cognitive attributes across individuals.