AI Can Detect a Child’s Sex From Brain Scans—But Gender Is More Complicated
The Study: Mapping Sex and Gender in the Young Brain
Neuroscientists have long debated whether—and how—sex and gender manifest in the brain. A new study using artificial intelligence (AI) adds fuel to this discussion, revealing that:
✔ An AI can predict a child’s sex (assigned at birth) with high accuracy based on brain connectivity patterns.
✔ Gender identity is also reflected in brain activity—but the AI struggled to predict it reliably.
✔ Sex and gender map onto distinct (but overlapping) neural networks, suggesting they shape the brain in different ways.
How the Research Was Conducted
Led by Elvisha Dhamala (Feinstein Institutes for Medical Research), the team analyzed MRI scans from 4,700 children (ages 9-10) in the Adolescent Brain Cognitive Development (ABCD) study.
- Sex was defined biologically (anatomy, hormones, genetics at birth).
- Gender was assessed via surveys—parents and children answered questions about:
- Self-perception (“Do you feel like a boy/girl?”)
- Behavior (“Do they imitate male/female TV characters?”)
- Discomfort with assigned sex (“Do they dislike their genitals?”)
Key Finding:
- Sex-linked brain patterns were strongest in regions controlling vision, movement, and emotion (limbic system).
- Gender-linked patterns were more widespread, involving attention, problem-solving, and sensory processing (cerebral cortex).
What the AI Revealed—And Where It Fell Short
1. Predicting Sex: Highly Accurate
- The AI model, trained on brain connectivity data, could correctly identify a child’s sex with significant reliability.
- This aligns with prior research showing structural differences between male and female brains.
2. Predicting Gender: Less Clear
- The AI could not reliably predict gender based on brain scans alone.
- It performed slightly better when using parents’ reports of gender expression (vs. children’s self-reports).
- Criticism: Some experts argue the gender findings may be statistical noise due to the large sample size.
“Gender is complex—it’s not just ‘male’ or ‘female.’ Reducing it to brain connectivity oversimplifies identity.”
— Dr. Ragini Verma, University of Pennsylvania
Why This Matters
1. Implications for Mental Health
Conditions like ADHD, depression, and anxiety differ in prevalence between sexes. Understanding neural differences could lead to better-targeted treatments.
2. Rethinking Neuroscience Research
- Most studies only account for sex, not gender—potentially missing key variables.
- Dhamala’s team urges future research to measure both separately.
3. The Limits of AI in Gender Science
- Gender is shaped by culture, upbringing, and personal experience—not just biology.
- Brain scans alone can’t capture this complexity, highlighting the need for multidisciplinary approaches.
Unanswered Questions
❓ How do transgender or nonbinary children’s brains compare? (The study didn’t categorize gender identity.)
❓ Do these patterns change during puberty?
❓ Could societal expectations influence gender-related brain differences?
Final Takeaways
🔹 Sex differences in the brain are detectable by AI—but gender is far more nuanced.
🔹 Neuroscience must evolve to study sex and gender separately.
🔹 AI has limits—identity can’t be reduced to brain scans alone.
“The brain is not destiny. It’s a dynamic canvas shaped by biology, experience, and self-perception.”