Neural network model identifies distinct brain organization patterns in women and men

A new study by Stanford Medicine researchers unveils a new artificial intelligence model that was more than 90 percent successful in determining whether scans of brain activity came from a woman or a man.

The results, published on February 19 in the Proceedings of the National Academy of Sciences, help resolve a long-term controversy over whether reliable sex differences exist in the human brain and suggest that understanding these differences may be key to combating neuropsychiatric diseases that affect women and men differently.

“One of the main motivations 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,” said Vinod Menon, Ph.D., professor of Psychiatry and Behavioral Sciences and Director. from the Stanford Cognitive and Systems Neuroscience Laboratory.

“Identifying consistent and reproducible sex differences in the healthy adult brain is a crucial step toward a deeper understanding of sex-specific vulnerabilities in psychiatric and neurological disorders.”

Menon is the lead author of the study. The lead authors are principal investigator Srikanth Ryali, Ph.D., and academic researcher Yuan Zhang, Ph.D.

The “hot spots” that most helped the model distinguish male from female brains include the default mode network, a brain system that helps us process self-referential information, as well as the striatum and limbic network, which are involved in learning and how we respond to rewards. .

The investigators noted that this work does not determine whether sex-related differences occur early in life or whether they may be driven by hormonal differences or the different societal circumstances that men and women are more likely to experience.

Discover brain differences

The extent to which a person’s gender affects the way their brain is organized and functions has long been a subject of controversy among scientists. Although we know which sex chromosomes we are born with help determine the cocktail of hormones our brains are exposed to – particularly during early development, puberty and aging – researchers have long had doubts. struggling to connect sex to concrete differences in the human brain.

Brain structures tend to be much the same in men and women, and previous research examining how brain regions work together has also failed to find consistent brain indicators of gender.

In their current study, Menon and his team took advantage of recent advances in artificial intelligence, as well as access to several large data sets, to pursue more powerful analysis than previously used.

First, they created a system that learned to classify brain imaging data: As the researchers showed the model brain scans and told it it was looking at a male or female brain, the model began to ” notice” what subtle patterns might help him. make the difference.

This model demonstrated superior performance than previous studies, in part because it used a deep neural network that analyzes dynamic MRI scans. This approach captures the complex interaction between different brain regions. When the researchers tested the model on about 1,500 brain scans, they were almost always able to determine whether the scan came from a woman or a man.

The success of the model suggests that detectable sex differences exist in the brain, but they simply have not been reliably detected before. The fact that this worked so well in different data sets, including brain scans from multiple sites in the United States and Europe, makes the results particularly compelling because they control for many of the confounds that can harm studies of this type. .

“This is very strong evidence that sex is an important determinant of human brain organization,” Menon said.

Make predictions

Until recently, a model like the one used by Menon’s team would help researchers sort brains into different groups, but would not provide information on how the sorting was accomplished. Today, however, researchers have access to a tool called “explainable AI,” which can sift through large amounts of data to explain how a model’s decisions are made.

Using explainable AI, Menon and his team identified which brain networks were most important in determining whether a brain scan came from a man or a woman. They found that the model most often looked to the default mode network, the striatum, and the limbic network to make the call.

The team then wondered if they could create another model that could predict participants’ performance in certain cognitive tasks, based on functional characteristics of the brain that differ between women and men. They developed sex-specific models of cognitive abilities: one model effectively predicted cognitive performance in men but not women, and another in women but not men.

The results indicate that functional brain characteristics that vary by sex have significant behavioral implications.

“These models worked very well because we were able to separate brain patterns between the sexes,” Menon said. “This tells me that neglecting sex differences in brain organization could lead us to miss key factors that cause neuropsychiatric disorders.”

While the team applied its deep neural network model to questions about gender differences, Menon says the model can be applied to answer questions about how almost any aspect of brain connectivity could be linked to any type of cognitive ability or behavior. He and his team plan to make their model publicly available so any researcher can use it.

“Our AI models have very broad applicability,” Menon said. “A researcher could use our models to look for brain differences related to learning disabilities or differences in social functioning, for example – aspects we want to better understand to help individuals adapt and overcome these challenges .”