JMIR Publications

ABSTRACT

Background:

Machine learning-based facial and vocal measurements have demonstrated relationships with schizophrenia diagnosis and severity. Here, we determine their accuracy when acquired through automated assessments conducted remotely through smartphones. Demonstrating utility and validity of remote and automated assessments conducted outside of controlled experimental settings can facilitate scaling such measurement tools to aid in risk assessment and tracking of treatment response in difficult to engage populations.

Objective:

We aim to assess the accuracy of these facial and vocal markers through remote assessments and compare them with traditional clinical assessments of schizophrenia severity.

Methods:

Measurements of facial and vocal characteristics including facial expressivity, vocal acoustics, and speech prevalence were assessed in 20 schizophrenia patients over the course of 2 weeks in response to two classes of prompts previously utilized in experimental laboratory assessments: evoked prompts, where subjects are guided to produce specific facial expressions and phonations, and spontaneous prompts, where subjects are presented stimuli in the form of emotionally evocative imagery and asked to freely respond. Facial and vocal measurements were assessed in relation to schizophrenia symptom severity using the Positive and Negative Syndrome Scale.

Results:

Vocal markers including speech prevalence, vocal jitter, fundamental frequency, and vocal intensity demonstrated specificity as markers of negative symptom severity while measurement of facial expressivity demonstrated itself as a robust marker of overall schizophrenia severity.

Conclusions:

Established facial and vocal measurements, collected remotely in schizophrenia patients via smartphones in response to automated task prompts, demonstrated accuracy as markers of schizophrenia severity. Clinical implications are discussed.