A groundbreaking study conducted by Stanford Medicine investigators has led to the development of a new artificial intelligence model that can accurately determine whether brain scans belong to a man or a woman with over 90% accuracy. The research, published in the Proceedings of the National Academy of Sciences, sheds light on reliable sex differences in the human brain.
Vinod Menon, Ph.D., the senior author of the study, highlights the significance of understanding these sex differences in addressing neuropsychiatric conditions that impact men and women differently. According to Menon, sex plays a crucial role in brain development, aging, and neuropsychiatric disorders.
The AI model identified key brain networks, such as the default mode network, striatum, and limbic network, that help distinguish male and female brains. This breakthrough comes at a time when previous research has struggled to connect sex with concrete brain differences due to the complexity of brain structures and inconsistencies in identifying sex indicators.
Using deep neural networks and dynamic MRI scans, the AI model achieved superior performance in detecting sex differences across datasets. The success of this model suggests that sex is a robust determinant of human brain organization. Additionally, “Explainable AI” was used to identify brain networks crucial for the model’s decisions when classifying male and female brains.
The team also developed sex-specific models to predict cognitive performance based on unique functional brain features for each sex. These findings have significant implications for understanding neuropsychiatric disorders and can be applied to various aspects of brain connectivity and cognitive abilities.
In a move to further research efforts in this area, the AI model will be publicly available for researchers to explore brain differences related to learning impairments, social functioning, and other cognitive abilities. This development marks a crucial step forward in understanding the complexities of the human brain and how sex influences its organization.
