Manifestation of depression in speech overlaps with characteristics used to represent and recognize speaker identity.

The sound of a person's voice is commonly used to identify the speaker. The sound of speech is also starting to be used to detect medical conditions, such as depression. It is not known whether the manifestations of depression in speech overlap with those used to identify the speaker. In this paper, we test the hypothesis that the representations of personal identity in speech, known as speaker embeddings, improve the detection of depression and estimation of depressive symptoms severity. We further examine whether changes in depression severity interfere with the recognition of speaker's identity. We extract speaker embeddings from models pre-trained on a large sample of speakers from the general population without information on depression diagnosis. We test these speaker embeddings for severity estimation in independent datasets consisting of clinical interviews (DAIC-WOZ), spontaneous speech (VocalMind), and longitudinal data (VocalMind). We also use the severity estimates to predict presence of depression. Speaker embeddings, combined with established acoustic features (OpenSMILE), predicted severity with root mean square error (RMSE) values of 6.01 and 6.28 in DAIC-WOZ and VocalMind datasets, respectively, lower than acoustic features alone or speaker embeddings alone. When used to detect depression, speaker embeddings showed higher balanced accuracy (BAc) and surpassed previous state-of-the-art performance in depression detection from speech, with BAc values of 66% and 64% in DAIC-WOZ and VocalMind datasets, respectively. Results from a subset of participants with repeated speech samples show that the speaker identification is affected by changes in depression severity. These results suggest that depression overlaps with personal identity in the acoustic space. While speaker embeddings improve depression detection and severity estimation, deterioration or improvement in mood may interfere with speaker verification.

The Beginning of the AI-Enabled Preventative PAP Therapy Era: A First-in-Human Proof of Concept Interventional Study.

Positive Airway Pressure (PAP) therapy is the most common and efficacious treatment for Obstructive Sleep Apnea (OSA). However, it suffers from poor patient adherence due to discomfort and may not fully alleviate all adverse consequences of OSA. Identifying abnormal respiratory events before they have occurred may allow for improved management of PAP levels, leading to improved adherence and better patient outcomes. Our previous work has resulted in the successful development of a Machine-Learning (ML) algorithm for the prediction of future apneic events using existing airflow and air pressure sensors available internally to PAP devices. Although researchers have studied the use of ML for the prediction of apneas, research to date has focused primarily on using external polysomnography sensors that add to patient discomfort and has not investigated the use of internal-to-PAP sensors such as air pressure and airflow to predict and prevent respiratory events. We hypothesized that by using our predictive software, OSA events could be proactively prevented while maintaining patients' sleep quality. An intervention protocol was developed and applied to all patients to prevent OSA events. Although the protocol's cool-down period limited the number of prevention attempts, analysis of 11 participants revealed that our system improved many sleep parameters, which included a statistically significant 31.6% reduction in Apnea-Hypopnea Index, while maintaining sleep quality. Most importantly, our findings indicate the feasibility of unobtrusive identification and unique prevention of each respiratory event as well as paving the path to future truly personalized PAP therapy by further training of ML models on individual patients.

Applications of Speech Analysis in Psychiatry.

ABSTRACT: The need for objective measurement in psychiatry has stimulated interest in alternative indicators of the presence and severity of illness. Speech may offer a source of information that bridges the subjective and objective in the assessment of mental disorders. We systematically reviewed the literature for articles exploring speech analysis for psychiatric applications. The utility of speech analysis depends on how accurately speech features represent clinical symptoms within and across disorders. We identified four domains of the application of speech analysis in the literature: diagnostic classification, assessment of illness severity, prediction of onset of illness, and prognosis and treatment outcomes. We discuss the findings in each of these domains, with a focus on how types of speech features characterize different aspects of psychopathology. Models that bring together multiple speech features can distinguish speakers with psychiatric disorders from healthy controls with high accuracy. Differentiating between types of mental disorders and symptom dimensions are more complex problems that expose the transdiagnostic nature of speech features. Convergent progress in speech research and computer sciences opens avenues for implementing speech analysis to enhance objectivity of assessment in clinical practice. Application of speech analysis will need to address issues of ethics and equity, including the potential to perpetuate discriminatory bias through models that learn from clinical assessment data. Methods that mitigate bias are available and should play a key role in the implementation of speech analysis.

High Frequency-Low Amplitude Oscillometry: Continuous Unobtrusive Monitoring of Respiratory Function on PAP Machines.

Oscillometry or Forced Oscillation Technique, traditionally used in intermittent clinical measurements, has recently gained substantial attention from its application as a continuous monitoring tool for large and small airways. However, low frequency (<8 Hz) continuous oscillometry faces high breathing noise, and hence requires high oscillation amplitudes to maintain an acceptable signal-to-noise ratio. Therefore, PAP machines that utilize low frequency oscillometry do so intermittently to distinguish airway patency several seconds after a breathing pause has occurred. We hypothesized that high frequency and low amplitude (HFLA) oscillometry may be as sensitive and applicable for monitoring upper airway patency to distinguish between central and obstructive apnea and hypopnea events, and for monitoring respiratory impedance. An inline oscillometry prototype device was developed and connected to commercial PAP machines to test whether oscillometry at 17, 43, and 79 Hz are as sensitive to airway patency as oscillometry at 4 Hz. Analysis of 11 patients with 171 apneas and hypopneas showed that all frequency oscillometry inputs were equally sensitive in distinguishing between central and obstructive apneas, while 17 Hz and 43 Hz oscillometry were most sensitive in distinguishing between central and obstructive hypopneas. Observations during normal breathing also showed the same periodicity and cross-correlation between impedance measurements from HFLA oscillometry compared to 4 Hz. Our findings provide an unobtrusive means of distinguishing airway patency during sleep and a means of continuous monitoring of respiratory function, with the potential for detection and prediction of developing respiratory diseases and significantly richer context for data analytics.