Marcella La Cioppa According to a recent study, artificial intelligence can identify when brain scans come from a woman or a man. An important discovery for finding a solution to certain neuropsychiatric conditions
Researchers at Stanford Medicine have recently unveiled a groundbreaking artificial intelligence model that can distinguish whether brain activity scans come from a woman or a man with more than 90% accuracy. Published in the Proceedings of the National Academy of Sciences, this study not only contributes to the long-standing debate on the existence of sexual differences in the human brain but also emphasizes the importance of understanding these differences for addressing neuropsychiatric conditions that disproportionately affect women and children.
Key motivation for the study
The underlying motivation for this research stems from the critical role that sex plays in brain development, aging, and the manifestation of psychiatric and neurological disorders. Identifying consistent and replicable sexual differences in the healthy adult brain is essential for a deeper understanding of sex-specific vulnerabilities in psychiatric and neurological disorders.
The “hotspots” that significantly aided the AI model in distinguishing male from female brains include the default mode network, involved in processing self-referential information, and the striatum and limbic network, which are crucial for learning and response to rewards.
Unveiling brain differences
The extent to which an individual’s sex influences the organization and functioning of their brain has been a controversial topic among scientists. While the sexual chromosomes we are born with determine the hormone cocktail our brains are exposed to during key developmental stages, researchers have struggled to link sex to concrete differences in the human brain. Traditional studies failed to provide consistent brain markers of sex, as brain structures tend to look similar in men and women, and previous research into how brain regions work together was inconclusive.
Leveraging recent advancements in artificial intelligence and access to large datasets, Menon and his team developed a deep neural network model that surpasses previous studies in classifying brain imaging data. This model’s superior performance is partly due to its analysis of dynamic MRI scans, capturing the intricate interaction between different brain regions.
Tested on approximately 1,500 brain scans, the model could almost always accurately determine the sex of the brain being scanned, offering compelling evidence that sex-based differences in the brain exist but were not reliably detected before.
Potential applications and cognitive performance models
The success of this AI model opens the door to a wide range of applications. Researchers can now explore brain differences related to learning disorders or social functioning differences, aiming to better understand and help individuals overcome these challenges. Additionally, the team developed sex-specific models for predicting cognitive task performance, finding that functional brain characteristics varying between sexes have significant behavioral implications.
These findings underscore the vast applicability of artificial intelligence in uncovering and understanding the subtle yet significant differences in brain organization and function between sexes, potentially revolutionizing the approach to diagnosing and treating neuropsychiatric disorders.
Source: PNAS
