AI can tell a person’s sex from brain scans with 90 per cent accuracy

Are the brains of men and women so different? A new way of investigating this question has concluded that they are – but it takes Artificial intelligence (AI) to differentiate them.

The question is whether we can measure the differences between men and women. the brain has long been controversial, with previous research showing conflicting results.

One problem is that men’s brains are slightly larger than women’s, perhaps because their bodies are generally larger, and some previous studies compared the size of different small brain regions. Unable to adjust overall brain volume. However, even before doing so, clear results have not emerged. “Once you correct for brain size, the results are quite variable,” says Vinod Menon at Stanford University in California.

To address the question differently, Menon’s team used a relatively new method called dynamic functional connectivity fMRI. This involves recording people’s brain activity while they lie in an active MRI scanner and tracking changes in how the activity of different regions varies in sync with each other.

The researchers designed an AI to analyze such brain-scanning data, which they trained on the results of nearly 1,000 young adults from an existing US database called the Human Connectome Project, to tell the AI ​​which individuals were male. And which women were there? In this analysis, the brain was divided into 246 different regions.

After this training process, the AI ​​was about 90 percent accurate in distinguishing between another set of brain scanning data of the same 1,000 men and women.

More importantly, the AI ​​was just as effective at distinguishing between brain scans of men and women from two different brain scanning datasets than it had seen before. They consisted of about 200 people between the ages of 20 and 35, both from the United States and Germany.

“What we bring to the table is a more rigorous study, with replication and generalization to other samples,” Menon says. None of the individuals included in the training or testing data were transgender.

“Their replication in samples that are completely independent of the Human Connectome Project makes me feel more confident about their results,” says. Camille Williams at the University of Texas at Austin.

The next question is whether the AI ​​will be as accurate when tested on larger additional sets of brain scanning results. “Only time will tell how it works with other datasets,” says Menon.

If confirmed, the findings could help us understand the reasons why certain medical conditions or forms of neural diversity differ by sex, Menon says. Attention Deficit Hyperactivity Disorder.

“If we don’t develop these sex-specific models, we will miss important aspects of the factors that make a difference. [for example]males with autism versus control males and females with autism versus control females,” says Menon.