Break stereotypes: Brain models are not a panacea

Machine learning has helped scientists understand how the brain produces complex human traits, revealing patterns

But this only works if the model represents everyone, whereas previous research has shown that models cannot represent everyone; There are always some people who don't fit into this model for any model

In a study published Aug.

Abigail Greene, MD, says that for models to work their best, they need to be applicable to any particular individual

Greene and her colleagues are interested in how the model provides more precise psychiatric features, which they believe can be achieved

Greene said: "Both of these advances will have an impact on

To understand the failure of the model, Greene and her colleagues first trained models that could use patterns in brain activity to predict a person's score on various cognitive tests

"We found consistency — the same individuals being misclassified

Next, they wanted to see if these similar misclassifications could be explained by differences in the brains of these individuals

Ultimately, they concluded that the models didn't just reflect cognitive abilities

This means that researchers need to think more carefully about what a given test really measures and, therefore, what

Taking these measures, the researchers say, will result in models that better reflect the cognitive structures

Todd Constable, professor of radiology and biomedical imaging at Yale School of Medicine and senior author of the study, said: "One day, you just need to provide different models