A New AI Model that can Distinguish Between Male and Female Brain Activity

Researchers at Stanford Medicine have developed a new artificial intelligence model that can distinguish between male and female brain activity scans with over 90% accuracy.

The results settle a long-running debate over whether there are discernible sex differences in the human brain and raise the possibility that treating neuropsychiatric disorders that affect men and women differently may depend on an understanding of these differences.

The Proceedings of the National Academy of Sciences published the findings.

A key motivation for this study is that sex plays a crucial role in human brain development, in aging, and in the manifestation of psychiatric and neurological disorders. Identifying consistent and replicable sex differences in the healthy adult brain is a critical step toward a deeper understanding of sex-specific vulnerabilities in psychiatric and neurological disorders.”

Vinod Menon PhD, Professor, Study Senior Author and Director, Psychiatry and Behavioral Sciences, Stanford Cognitive and Systems Neuroscience Laboratory

The senior author of the study is Menon. Senior research scientist Srikanth Ryali, PhD, and academic staff researcher Yuan Zhang, PhD, are the lead authors.

The limbic and striatal networks, which are involved in learning and reward response, as well as the default mode network, a brain system that aids in processing self-referential information, are the "hotspots" that most assisted the model in differentiating between the male and female brains.

The researchers pointed out that this research does not address whether sex-related differences develop early in life, whether hormonal variations or the distinct social situations that men and women may be more likely to experience are the causes of sex-related differences.

Uncovering Brain Differences

Scientists have long disagreed on the degree to which a person's sex influences the structure and function of their brain. The hormone cocktail human brains are exposed to, especially during early development, puberty, and aging, is known to be influenced by the sex chromosomes people are born with.

However, researchers have long struggled to link sex to specific differences in the human brain. Men’s and women's brain structures are generally similar, and prior studies looking at the interactions between different brain regions have also mostly failed to find consistent brain markers of sex.

In order to conduct a more potent analysis than has previously been used, Menon and his team used access to numerous large datasets and recent developments in artificial intelligence in their current study. They started by developing a deep neural network model that can categorize data from brain imaging: The model began to “notice” what minute patterns could aid in differentiating between a male and female brain as the researchers showed it brain scans and explained that it was looking at a brain.

This model's superior performance over those in earlier studies can be partly attributed to its use of a deep neural network for dynamic MRI scan analysis. This method captures the complex interactions between various brain regions. The model was able to identify whether a brain scan belonged to a woman or a man nearly all the time when it was tested on about 1,500 brain scans.

Due to the model's success, it appears that there are detectable sex differences in the brain, they have simply not been identified with any degree of reliability. The findings are particularly convincing because they account for numerous confounding variables that can impede studies of this type, as evidenced by its excellent performance across a variety of datasets, including brain scans from several locations in the US and Europe.

This is a very strong piece of evidence that sex is a robust determinant of human brain organization.”

Vinod Menon PhD, Professor, Study Senior Author and Director, Psychiatry and Behavioral Sciences, Stanford Cognitive and Systems Neuroscience Laboratory

Making Predictions

Up until recently, a model similar to the one Menon's team used would assist researchers in classifying brains into distinct groups, but it would not reveal the mechanism underlying the classification. But now, a tool known as “explainable AI” is available to researchers, which can sort through enormous volumes of data and explain how a model makes decisions.

Menon and his colleagues used explainable AI to determine which brain networks were most crucial to the model's ability to determine whether a brain scan was from a man or a woman. They discovered that the model frequently relied on the limbic, striatal, and default mode networks to make decisions.

Subsequently, the team contemplated whether they could develop an alternative model that could forecast an individual's performance on specific cognitive tasks by utilizing functional brain features that vary between the sexes. They created sex-specific cognitive ability models: Two models were found to be effective in predicting cognitive performance: one for men only, and another for women only. The results suggest that behavioral consequences of functional brain differences between genders are substantial.

Menon says, “These models worked really well because we successfully separated brain patterns between sexes. That tells me that overlooking sex differences in brain organization could lead us to miss key factors underlying neuropsychiatric disorders.”

Although Menon says the model can be used to answer questions about how almost any aspect of brain connectivity might relate to any kind of cognitive ability or behavior, the team applied their deep neural network model to questions about sex differences. He and his group intend to release their model for use by any researcher on a public basis.

Our AI models have very broad applicability. A researcher could use our models to look for brain differences linked to learning impairments or social functioning differences, for instance - aspects we are keen to understand better to aid individuals in adapting to and surmounting these challenges.”

Vinod Menon PhD, Professor, Study Senior Author and Director, Psychiatry and Behavioral Sciences, Stanford Cognitive and Systems Neuroscience Laboratory