Brain imaging marker may predict depression risk years before symptoms

Clinical takeaway: A brain-based marker of altered emotional processing predicted depression-related symptoms four years later in adolescents without clinically significant symptoms at baseline, pointing to a potential new approach for identifying young people at higher risk before depression develops.

Depression often emerges during adolescence and early adulthood, but clinicians currently have no validated brain-based test that can reliably identify which asymptomatic young people will go on to develop symptoms. A new study suggests that altered processing of negative emotional cues may provide an early signal of vulnerability.

Researchers analyzed functional MRI data from 1,332 adolescents at age 19 while they viewed angry and neutral faces. Using deep learning, they identified a pattern suggesting that some individuals relied too heavily on preexisting negative emotional expectations when interpreting facial expressions, allowing those expectations to outweigh the visual information actually presented.

This pattern was associated with greater emotional symptoms at age 19. More importantly, among 725 participants without clinically relevant emotional symptoms at baseline, the computational markers predicted emotional symptoms four years later.

The signal also showed some specificity for depression. It was associated with genetic risk for major depressive disorder and was more prominent in an independent cohort of 134 young adults with major depression than in healthy controls, but not significantly different in patients with alcohol use disorder, anorexia nervosa, or bulimia nervosa.

The findings do not change current screening recommendations or support ordering neuroimaging to assess depression risk. But if validated in broader populations, this approach could eventually help identify adolescents at elevated risk before clinically significant symptoms emerge, creating an opportunity for closer monitoring and earlier preventive intervention.

Source: Lu H, et al. (2026 Aug 7) Sci Adv. Deep learning reveals a neurocomputational mechanism predicting depression risk in adolescents