Activity within brain networks appears to differ between boys and girls
People Images/Getty Images
Artificial intelligence can now distinguish the brain patterns of 9- to 10-year-old boys and girls according to their sex and even gender, but not everyone is convinced of the accuracy of the results.
Although the prevalence of symptoms such as pain, headaches and cardiac disease differs between men and women, little is known about neurological differences, especially among children.
To learn more, Elvisha Damala and her colleagues at the Feinstein Institute for Medical Research in New York analyzed thousands of sets of magnetic resonance imaging (MRI) data from more than 4,700 children, roughly equal in gender, all between the ages of 9 and 10, who were participating in the Adolescent Brain Cognitive Development Project.
Sex was defined based on “anatomical, physiological, genetic and hormonal structures at birth,” while gender was determined based on “an individual’s attitudinal, emotional and behavioral characteristics.”
Parents weren’t asked directly about their thoughts about their child’s gender, but were assessed with a series of questions, such as how often their child imitates male or female characters on TV or in movies, whether they wanted to be a girl or a boy, whether they said they disliked their genitals, etc. All these questions were weighted equally and combined into a score.
A separate score was created from questions that asked the children themselves, such as whether they felt like a boy or a girl.
The researchers did not disclose the different genders the children identified as, or how many of the children had a gender that was different from their own gender. “We thought of gender as a continuum, not a binary,” Damala said. “We did not limit our analysis to gender categories, so we cannot comment on how many children had a gender that was different from their own gender.”
The researchers first looked at the relationship between brain networks and sex, and then looked at the relationship between these networks and sex for each assigned sex. They found that sex and gender differences were associated with distinct patterns of functional connectivity, a measure of communication between distant brain regions.
Gender was associated with connectivity between the visual cortex, which controls movement, and the limbic system, a group of deep brain structures involved in regulating emotion, behavior, motivation, and memory. These networks were “important in distinguishing participants based on their gender,” Damala said.
Gender-related networks were widespread throughout the cerebral cortex (the outer layer of the brain that is also associated with memory, movement, sensation and problem solving), both when using gender scores constructed from responses to parental questions and when using separate scores constructed by asking questions of the children themselves.
“In assigned females, sex mapped to networks involved in attention, emotion processing, motor control, and higher-order thinking,” Damala says. “In assigned males, the same relationships existed, but there were additional networks involved in higher-order thinking and visual processing. Although there was some overlap between sex- and gender-related brain networks, they were very distinct from each other.”
Once the researchers trained an AI model on some of the MRI data, it was able to identify a child’s gender based on brain connectivity patterns in other datasets. It could also predict gender, but this was far less accurate than predicting gender, and was based solely on the gender reported by the parents, not the child themselves.
A better understanding of how brain activity patterns differ by sex could help scientists learn more about conditions that affect boys and girls at different rates, such as ADHD, Damala said.
The findings could also have implications for how human brain research is conducted, she says: “This shows that sex and gender need to be considered separately in biomedical research. This applies to how data is collected, how it is analyzed, and how results are interpreted and communicated,” Damala says.
But Ragini Verma of the University of Pennsylvania says the study tells us very little about the neurological basis of gender: Because of the study’s large sample size, the team could only find signals of brain activity patterns that likely differ between the sexes, but “the variability in gender prediction is based on low precision,” she says.
topic:
