TLDR:

• Stanford Medicine investigators developed an AI model that can determine sex differences in brain scans with over 90% accuracy.
• The AI model identified key brain networks that differ between men and women and have implications for psychiatric and neurological disorders.

AI Finds Women and Men Differ with Respect to Brain Organization and Function

Stanford Medicine investigators have developed a new artificial intelligence model that can determine with over 90% accuracy whether MRI scans of human brain activity belong to a man or a woman. The findings suggest that sex differences play a crucial role in human brain development, aging, and neuropsychiatric conditions.

The AI model, called stDNN, utilizes deep learning techniques to analyze resting-state functional MRI (rsfMRI) brain images from the Human Connectome Project. It identified reliable sex differences in brain organization and performance, outperforming previous studies.

The model highlighted key brain networks, such as the default mode network (DMN) and the striatum and limbic network, which are associated with psychiatric disorders that have sex-specific outcomes.

Furthermore, the investigators used explainable AI (XAI) to predict cognitive abilities based on functional brain features that vary between men and women. They found that these differences have significant behavioral implications.

The team plans to make their AI model publicly available for researchers to use in studying brain connectivity and cognitive abilities. The results of the study lay the foundation for more targeted and personalized approaches in cognitive neuroscience research and clinical applications.