JMIR Publications

ABSTRACT

Background:

Monitoring symptoms of bipolar disorder (BD) is a challenge faced by mental health services. Speech patterns are crucial in assessing current experiences, emotions, thought patterns of people with BD. Natural Language Processing (NLP), acoustic signal processing and Ecological Momentary Assessment (EMA) may support ongoing BD assessment within a mHealth framework.

Objective:

Jointly using both acoustic and NLP-based features from the speech of people with BD, we built an app-based tool testing its feasibility and performance to remotely assess the individual clinical status.

Methods:

We carried out a pilot, observational study, sampling adults diagnosed with BD to explore the relationship between selected speech features and symptom severity, as assessed by the Young Mania and Montgomery-Åsberg Depression Rating scales, and to test their potential to remotely assess mental health status. Leveraging a digital health tool embedded in a mobile app, which records and processes speech, participants self-administered verbal performance tasks. Both NLP-based and acoustic features were extracted, testing associations with mood states and exploiting machine learning approaches based on random forest models.

Results:

Small-to-moderate correlations between speech features and symptom severity were uncovered, though with gender-based differences in predictive capability. Higher latency time and increased silences, as well as vocal perturbations, correlated to depressive symptomatology. Pressure of speech and lower voice instability were detected for manic symptoms. However, a higher contribution of NLP-based and conversational, rather than prosodic features, was uncovered, especially for predictive models.

Conclusions:

Remotely collected speech patterns, underlying both linguistic and acoustic features, are associated with symptom severity levels, somehow discriminating clinical conditions of people with BD when assessing their mood states. Multimodal, smartphone-integrated, digital EMA can provide a powerful tool for clinical purposes, remotely complementing standard mental health evaluation.