Assess the Utility of a Speech-based Machine Learning Algorithm to Predict Treatment Response to Psychiatric Interventions

This observational cohort study (n=200) will assess whether a speech-based machine learning algorithm can predict treatment response to psychiatric interventions, specifically repetitive transcranial magnetic stimulation (TMS) and Spravato (esketamine) nasal spray.

Conducted by Psyrin, this study focuses on individuals diagnosed with major depressive disorder (MDD), post-traumatic stress disorder (PTSD), bipolar disorder (BD), generalised anxiety disorder (GAD), or obsessive-compulsive disorder (OCD). Participants, aged 18-68, will record 12-minute speech samples before treatment, daily during treatment, immediately after, and four weeks later. Researchers will analyse speech patterns and compare them with clinical outcomes, measured by symptom severity changes on the Clinical Global Impression scale. The study aims to develop a non-invasive tool to help doctors predict which treatment is more likely to be effective for individual patients.

The study will take place over 12 months at two locations in the United States (California and New York). It will also assess whether speech assessments cause any distress to participants. If successful, this research could improve personalised treatment approaches and optimise healthcare resource allocation.

Status Not yet recruiting
Results Published No
Start date 01 February 2025
End date 01 January 2026
Phase Not Applicable
Design Open
Type Observational
Generation Second
Participants 200
Sex All
Age 18- 68
Therapy No

Trial Details

This is a research study looking at whether the way people speak can help predict how well they'll respond to certain mental health treatments. The Main Goal: The researchers want to see if computer analysis of a person's speech patterns can predict whether they'll respond well to two specific treatments: TMS (Transcranial Magnetic Stimulation) and Spravato (a nasal spray medication). They're focusing on people with depression, bipolar disorder, OCD, anxiety, and PTSD. How It Works: 200 people with these conditions will participate in the study.Participants will record themselves speaking for about 12 minutes, responding to six different prompts.They'll do these recordings before treatment starts, daily during treatment, right after treatment ends, and again four weeks later. Doctors will track how well people are doing using various questionnaires and rating scales The researchers will look for connections between speech patterns and treatment success. The study will last 12 months. What Makes Someone a "Treatment Success": The study considers treatment successful if a person's symptoms improve significantly (specifically, a 2-point or greater reduction on a clinical rating scale (called Clinical Global Impression) and stays improved during the follow-up period (4-weeks). Why This Matters: If successful, this research could lead to a simple, non-invasive way to help doctors predict which treatments might work best for different patients. This could help people get the most effective treatment more quickly and help healthcare providers use their resources more efficiently. Safety Consideration The researchers will also check whether doing the speech assessments causes any distress to participants, making sure the evaluation process itself is safe and comfortable.

Trial Number NCT06823024

Data attribution

A large set of the trials in our database are sourced from ClinicalTrials.gov (CTG). We have modified these post to display the information in a more clear format or to correct spelling mistakes. Our database in actively updated and may show a different status (e.g. completed) if we have knowledge of this update (e.g. a published paper on the study) which isn't reflected yet on CTG. If a trial is not sourced from CTG, this is indicated on this page and you can follow the link to the alternative source of information.