One of the most fundamental beliefs in our society is that men and women are distinct. Recent times have seen this as one of the most frequently asked questions. Society is fascinated by the existence of brain differences between men and women. Each new discovery is the subject of heated debates, in which scientific knowledge is frequently mixed with preconceived notions.
There are many people, both inside and outside of the scientific community, who believe brain differences are small and few. They point out, with good reason, that we don't know much about brain differences and their possible relationships to temperament, cognitive abilities, or behavior.
There are also people outside and within the scientific community who believe the contrary. These differences are not insignificant and, even though they may seem small, add them all up and you get two types of brains: male or female.
These results, which have been supported by a multitude of scientific studies, show that machine learning algorithms (AAA), are capable of predicting sex based on brain scan images. They do this with a success rate of over 90%.
Are these two types of brains the final proof?
It is possible. It seems plausible to believe that these algorithms are capable of predicting gender with high accuracy. They can also identify a set features that distinguish the brains of men and women clearly and reliably. It is almost impossible.
But, as is often the case, it's not always easy.
First, and regardless of how famous the figure of 90% may be, many studies show that the efficacy is much lower at around 70%. In one study, the efficacy was only 51.1%.
These predictions are based on neuroanatomical features and have not considered the size differences between men's and women's skulls. This is because the variable artificially increases neuroanatomical differences' size. The ability of algorithms to predict gender can actually be reduced by as much as 60% when they are controlled using statistical procedures.
Different algorithms can identify sex with great accuracy. However, it is not possible to assume that all of them have identified the same set characteristic features capable of distinguishing the brains of men and women.
Although they show a similar efficiency in their predictions, various studies have identified distinct brain areas that are characteristic of male and female brains.
These studies also showed that brain characteristics of North American men and women are different in other regions.
A recent study has shown that even though a single sample was used, different algorithms still base their predictions from different brain areas. The result is that the same brain can be classified by different algorithms as "male" and "female".
The five algorithms used in this study had a range of 86 to 90% for correctly classifying cases, but 75% of the cases were correctly classified by them all. If skull size differences were considered, it was 38%!
The same study revealed that cases correctly classified using a single algorithm do not have the same brain characteristics. The brain patterns of men and women were not similar to those of random women.
When analyzed in detail, it is impossible to conclude that there are only two types of brains. However, it seems that each algorithm has its own type of types.
Additionally, although algorithms can identify the sex of an individual based on brain characteristics, it is not possible to predict the person's characteristics by knowing their sex. Your brain.
This doesn't mean there aren't differences in the brains and brains of men and women, or that algorithms can't be used to characterize them.
These and other findings suggest that there are differences at the group level. However, they do not result from divergences between two types brains (one "typically masculine" and one "typically feminine" reproduced in most men. and women, respectively.
We need to reexamine how we interpret and study these differences.
This article was published in "The Conversation".