Early language patterns in children may predict depression and anxiety years ahead

Research suggests that analysing children’s speech for subtle linguistic habits could enable earlier detection of depression and anxiety, offering a new tool for mental health screening before traditional symptoms emerge.

Children’s words may offer an early warning of depression or anxiety long before the first obvious symptoms appear, according to research highlighted in Nature Mental Health and summarised by the National Institutes of Health. In a study of 204 children aged nine to 13, investigators found that the way young people described stressful experiences was more revealing than the specific events themselves, with language patterns predicting later internalising problems up to six years ahead.

The Stanford-led team followed children who did not have a diagnosed mood or anxiety disorder at the outset and recorded about 30 minutes of speech from each during interviews about stressful moments. Using natural language processing, the researchers examined grammar, sentence structure and word choice rather than simply judging emotional content. Their analysis suggested that language style carried more predictive value than baseline clinical assessments, including expert ratings of how severe a child’s stress appeared to be.

What stood out most was not dramatic storytelling but the small linguistic habits that can reflect thought patterns. The study linked later mental health difficulties with heavier use of function words such as prepositions and conjunctions, which may signal repetitive thinking and rumination. It also found that children who later developed depression or anxiety were more likely to use absolutist language such as “always” and “never”, along with more first-person singular words such as “I”, “me” and “my”. By contrast, children who stayed mentally well tended to speak in a more outward-looking way and referred more often to other people and everyday activities.

That pattern fits with a broader body of research showing that language can carry mental health clues across age groups. Studies in adults have found that tone, pronoun use and time-focused wording can be associated with depressive or anxious symptoms, while other work has shown that computational language analysis can help distinguish age and mental health status. Separate research has also suggested that language collected through smartphones may be a practical way to monitor changes in mood over time without requiring constant direct questioning.

For parents, paediatricians and schools, the appeal of this approach is clear: it could help identify vulnerable children before problems become entrenched. Researchers cited by the National Institute of Mental Health have argued that earlier detection could reduce both suffering and the long-term financial burden linked with paediatric mental illness in the United States. But experts also stress that any automated screening tool would need strong privacy safeguards and should support, not replace, clinical judgement. In practice, the promise lies in using children’s language as one more signal among many, not as a diagnosis on its own.