Photo Credit: iStock.com/PeopleImages

Open-source analysis of clinical interview speech revealed distinct vocal patterns linked to schizophrenia symptom severity, supporting passive digital phenotyping. 

A study published in the June 2025 issue of Frontiers in Psychiatry about speech as a clinically relevant marker for schizophrenia symptom severity, yet current approaches to digital phenotyping often rely on closed-source platforms that can increase the burden for both clinicians and individuals.  

Researchers conducted a retrospective study to assess clinical interview recordings from standard trial procedures as a reliable source of speech for assessing schizophrenia symptom severity using Positive and Negative Syndrome Scale (PANSS) scores.  

They analyzed 825 PANSS interview recordings from 212 participants enrolled in a Phase 2 clinical trial for schizophrenia. Speech features were derived using open-source code. Mixed effects models were applied, adjusting for demographic variables and time to examine the association between speech characteristics and PANSS scores.  

The results showed that calculated speech characteristics were strongly linked to schizophrenia symptom severity. Positive symptoms correlated with increased speech quantity, faster rate, and shorter, less varied pauses. In contrast, negative symptoms were linked to reduced speech, slower rate, and longer, more varied pauses.  

Investigators concluded that speech captured during routine clinical interviews could be effectively processed using open-source methods to support digital phenotyping and symptom assessment in schizophrenia. 

Source: frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2025.1571647/full