Recruiting

Using speech patterns to predict psychosis relapse risk

Plain-English translation of NCT07397975 on ClinicalTrials.gov ↗ · Source last updated February 2026 · Translation generated June 2026 · How we translate trials

What is this trial?

Researchers want to understand if analyzing the way you speak—through weekly recordings on your smartphone—can help predict whether someone with psychosis might experience a relapse. This study will collect speech data from about 360 people: some who have experienced psychosis and are at risk of relapse, and some who are healthy. The researchers will use artificial intelligence to analyze the recordings, but trained experts will also review the data to ensure accuracy.

Right now, doctors don't have a reliable way to predict when someone with psychosis might relapse. If speech patterns can serve as an early warning sign, it could help doctors step in sooner with support or treatment adjustments before a relapse happens.

Do you qualify?

You likely qualify if…

you are between 18 and 65 years old
you have been diagnosed with a psychotic disorder (such as schizophrenia, schizoaffective disorder, or bipolar disorder with psychotic features)
you are currently in a stable phase with only mild symptoms or fewer
you can speak and understand fluently one of these languages: English, German, French, Dutch, Czech, or Turkish
you have access to a smartphone and can use a mobile app
you are able to give informed consent and commit to weekly participation for one year

You likely don't qualify if…

you have severe speech disorders like aphasia or severe stuttering that would make voice recording difficult
you are not willing or able to understand what the study involves
you have significant untreated medical conditions affecting your heart, brain, kidneys, or lungs
you have active substance abuse or alcohol-related psychiatric problems

What participation looks like

Each week for one year, you would use a smartphone app to record yourself speaking—answering prompts or telling a short story. These recordings are securely sent to researchers who analyze them using computer software and human review. You'll be asked to do this consistently each week, but the study is observational, meaning the analysis won't directly change your current treatment. Researchers will compare speech patterns over time to see if changes might signal increased relapse risk.

AI-generated summary from trial data · Jun 14, 2026 · Not medical advice

Trial locations(7 sites)

Czechia

National Institute of Mental Health, Klecany

Ireland

Newcastle Hospital and RCSI University of Medicine and Health Sciences, Greystones, Wicklow

Netherlands

University Medical Center Groningen, Groningen

Norway

The Arctic University of Norway, Tromsø

Switzerland

University Hospital Geneva and University of Geneva, Geneva, Chêne-Bourg