When artificial intelligence succeeds in distinguishing the brains of women from those of men

Alexandre Bozio 3:16 p.m., February 27, 2024

American researchers from Stanford University have succeeded in developing artificial intelligence capable of discerning female and male brains based on the analysis of brain MRIs.

The algorithm notably used certain brain networks linked to emotions. 

For a long time, scientists who had tackled the question of the difference in the brain between men and women had found themselves at an impasse.

Of course, we knew that there were certain disparities such as the size and constitution of the brain, but we did not know whether the male brain really worked differently from that of a woman.

Neurologists from the American University of Stanford have worked on the subject using artificial intelligence and the answer is clear: on the physiological level of brain activity, sexual dimorphism, namely all the differences more or less marked morphological differences between male and female individuals of the same species is clearly present. 

>> READ ALSO

- What happens in our brain when we sleep?

90% success

During this study, published in the journal Proceedings of the National Academy of Sciences (Pnas), the algorithm trained by the researchers was able to differentiate almost without fail whether the brain MRIs that The ones presented to him were those of a man or a woman.

To achieve this result, the scientists first showed the artificial intelligence nearly 800 MRI scans of different regions of the brain, telling it each time whether it belonged to a male or female person. 

Next comes the testing phase.

Another 200 MRIs were shown to the AI, which in 90% of cases managed to classify the image in the correct category.

"It's quite interesting. Without the scientists having specified a type of activity within specific brain regions, the algorithm seems to have succeeded in identifying a brain signature of biological sex", detailed neuroscientist Christophe Rodo, during of an interview with our colleagues from Le

Figaro

. 

>> READ ALSO

 - We have found the area of ​​the brain which is responsible for our attention

Emotions as an element of differentiation

But how did the AI ​​manage to differentiate the gender of the subjects so successfully?

The researchers were notably able to question the algorithm on its internal reasoning and it emerged that it primarily looked at the activity of three types of neural networks closely linked to the processing of emotions, namely the striatum which regulates the motivation and impulses, but also the limbic system which plays a major role in memory and the control of emotions.

Finally, the neural network on which AI has relied the most is the one called "default mode" (MPD) by scientists, which refers to a part of the brain that is activated when it is at rest, but active, especially during dreams. 

From now on, the researchers hope that this study can be used for medical purposes and in particular to improve the treatment of neurological diseases, even if the spectrum of the subjects, all aged between 20 and 35 years old, does not allow a perfectly representative vision.