Search

PURPOSE: Mental health screening is recommended by the US Preventive Services Task Force for all patients in areas where treatment options are available. Still, it is estimated that only 4% of primary care patients are screened for depression. The goal of this study was to evaluate the efficacy of machine learning technology (Kintsugi Voice, v1, Kintsugi Mindful Wellness, Inc) to detect and analyze voice biomarkers consistent with moderate to severe depression, potentially allowing for greater compliance with this critical primary care public health need. METHODS: We performed a cross-sectional study from February 1, 2021 to July 31, 2022 to examine =25 seconds of free-form speech content from English-speaking samples captured from 14,898 unique adults in the United States and Canada. Participants were recruited via social media, provided informed consent, and their voice biomarker results were compared with a self-reported Patient Health Questionnaire-9 (PHQ-9) at a cut-off score of 10 (moderate to severe depression). RESULTS: From as few as 25 seconds of free-form speech, machine learning technology was able to detect vocal characteristics consistent with an increased PHQ-9 =10, with a sensitivity of 71.3 (95% CI, 69.0-73.5) and a specificity of 73.5 (95% CI, 71.5-75.5). CONCLUSIONS: Machine learning has potential utility in helping clinicians screen patients for moderate to severe depression. Further research is needed to measure the effectiveness of machine learning vocal detection and analysis technology in clinical deployment. [ABSTRACT FROM AUTHOR]

Introduction: The 2024 Voice AI Symposium, presented by the Bridge2AI-Voice Consortium, featured deep-dive educational workshops conducted by experts from diverse fields to explore the latest advancements in voice biomarkers and artificial intelligence (AI) applications in healthcare. Through five workshops, attendees learned about topics including international standardization of vocal biomarker data, real-world deployment of AI solutions, assistive technologies for voice disorders, best practices for voice data collection, and deep learning applications in voice analysis. These workshops aimed to foster collaboration between academia, industry, and healthcare to advance the development and implementation of voice-based AI tools. Methods: Each workshop featured a combination of lectures, case studies, and interactive discussions. Transcripts of audio recordings were generated using Whisper (Version 7.13.1) and summarized by ChatGPT (Version 4.0), then reviewed by the authors. The workshops covered various methodologies, from signal processing and machine learning operations (MLOps) to ethical concerns surrounding AI-powered voice data collection. Practical demonstrations of AI-driven tools for voice disorder management and technical discussions on implementing voice AI models in clinical and non-clinical settings provided attendees with hands-on experience. Results: Key outcomes included the discussion of international standards to unify stakeholders in vocal biomarker research, practical challenges in deploying AI solutions outside the laboratory, review of Bridge2AI-Voice data collection processes, and the potential of AI to empower individuals with voice disorders. Additionally, presenters shared innovations in ethical AI practices, scalable machine learning frameworks, and advanced data collection techniques using diverse voice datasets. The symposium highlighted the successful integration of AI in detecting and analyzing voice signals for various health applications, with significant advancements in standardization, privacy, and clinical validation processes. Discussion: The symposium underscored the importance of interdisciplinary collaboration to address the technical, ethical, and clinical challenges in the field of voice biomarkers. While AI models have shown promise in analyzing voice data, challenges such as data variability, security, and scalability remain. Future efforts must focus on refining data collection standards, advancing ethical AI practices, and ensuring diverse dataset inclusion to improve model robustness. By fostering collaboration among researchers, clinicians, and technologists, the symposium laid a foundation for future innovations in AI-driven voice analysis for healthcare diagnostics and treatment. [ABSTRACT FROM AUTHOR]

IntroductionArtificial intelligence (AI) models trained on audio data may have the potential to rapidly perform clinical tasks, enhancing medical decision-making and potentially improving outcomes through early detection. Existing technologies depend on limited datasets collected with expensive recording equipment in high-income countries, which challenges deployment in resource-constrained, high-volume settings where audio data may have a profound impact on health equity.MethodsThis report introduces a novel protocol for audio data collection and a corresponding application that captures health information through guided questions.ResultsTo demonstrate the potential of Voice EHR as a biomarker of health, initial experiments on data quality and multiple case studies are presented in this report. Large language models (LLMs) were used to compare transcribed Voice EHR data with data (from the same patients) collected through conventional techniques like multiple choice questions. Information contained in the Voice EHR samples was consistently rated as equally or more relevant to a health evaluation.DiscussionThe HEAR application facilitates the collection of an audio electronic health record (“Voice EHR”) that may contain complex biomarkers of health from conventional voice/respiratory features, speech patterns, and spoken language with semantic meaning and longitudinal context–potentially compensating for the typical limitations of unimodal clinical datasets.

IntroductionThe 2024 Voice AI Symposium, hosted by the Bridge2AI-Voice Consortium in Tampa, FL, featured two keynote speeches that addressed the intersection of voice AI, healthcare, ethics, and law. Dr. Rupal Patel and Dr. Nicholson Price provided insights into the advancements, applications, and challenges of AI-driven voice tools in healthcare. The symposium aimed to advance cross-disciplinary collaboration and establish frameworks for the ethical use of AI technologies in healthcare.MethodsThe keynote speeches, delivered on May 1st and 2nd, were 30 minutes each, followed by 10-minutes Q&A sessions. The audio was recorded and transcribed using Whisper (v7.13.1). Content summaries were generated with the aid of ChatGPT (v4.0), and the authors reviewed and edited the final transcripts to ensure accuracy and clarity.ResultsDr. Rupal Patel’s keynote, “Reflections and New Frontiers in Voice AI”, explored the potential of voice AI for early detection of health conditions, monitoring disease progression, and promoting non-invasive global health management. She highlighted innovative uses beyond traditional applications, such as examining menopause-related symptoms. Dr. Nicholson Price’s keynote, “Governance for Clinical Voice AI”, addressed the regulatory and ethical challenges posed by AI in healthcare. He stressed the need for context-aware systems and dynamic legal frameworks to address liability and accountability.ConclusionsThe 2024 Voice AI Symposium highlighted the transformative potential of voice AI for early detection, health monitoring, and reducing healthcare disparities. It also underscored the importance of dynamic governance to address the ethical and regulatory challenges of deploying AI in clinical settings.

Introduction: The 2024 Voice AI Symposium, hosted by the Bridge2AI-Voice Consortium in Tampa, FL, featured two keynote speeches that addressed the intersection of voice AI, healthcare, ethics, and law. Dr. Rupal Patel and Dr. Nicholson Price provided insights into the advancements, applications, and challenges of AI-driven voice tools in healthcare. The symposium aimed to advance cross-disciplinary collaboration and establish frameworks for the ethical use of AI technologies in healthcare. Methods: The keynote speeches, delivered on May 1st and 2nd, were 30 minutes each, followed by 10-minutes Q&A sessions. The audio was recorded and transcribed using Whisper (v7.13.1). Content summaries were generated with the aid of ChatGPT (v4.0), and the authors reviewed and edited the final transcripts to ensure accuracy and clarity. Results: Dr. Rupal Patel's keynote, "Reflections and New Frontiers in Voice AI", explored the potential of voice AI for early detection of health conditions, monitoring disease progression, and promoting non-invasive global health management. She highlighted innovative uses beyond traditional applications, such as examining menopause-related symptoms. Dr. Nicholson Price's keynote, "Governance for Clinical Voice AI", addressed the regulatory and ethical challenges posed by AI in healthcare. He stressed the need for context-aware systems and dynamic legal frameworks to address liability and accountability. Conclusions: The 2024 Voice AI Symposium highlighted the transformative potential of voice AI for early detection, health monitoring, and reducing healthcare disparities. It also underscored the importance of dynamic governance to address the ethical and regulatory challenges of deploying AI in clinical settings. [ABSTRACT FROM AUTHOR]

IntroductionThe 2024 Voice AI Symposium, presented by the Bridge2AI-Voice Consortium, featured deep-dive educational workshops conducted by experts from diverse fields to explore the latest advancements in voice biomarkers and artificial intelligence (AI) applications in healthcare. Through five workshops, attendees learned about topics including international standardization of vocal biomarker data, real-world deployment of AI solutions, assistive technologies for voice disorders, best practices for voice data collection, and deep learning applications in voice analysis. These workshops aimed to foster collaboration between academia, industry, and healthcare to advance the development and implementation of voice-based AI tools.MethodsEach workshop featured a combination of lectures, case studies, and interactive discussions. Transcripts of audio recordings were generated using Whisper (Version 7.13.1) and summarized by ChatGPT (Version 4.0), then reviewed by the authors. The workshops covered various methodologies, from signal processing and machine learning operations (MLOps) to ethical concerns surrounding AI-powered voice data collection. Practical demonstrations of AI-driven tools for voice disorder management and technical discussions on implementing voice AI models in clinical and non-clinical settings provided attendees with hands-on experience.ResultsKey outcomes included the discussion of international standards to unify stakeholders in vocal biomarker research, practical challenges in deploying AI solutions outside the laboratory, review of Bridge2AI-Voice data collection processes, and the potential of AI to empower individuals with voice disorders. Additionally, presenters shared innovations in ethical AI practices, scalable machine learning frameworks, and advanced data collection techniques using diverse voice datasets. The symposium highlighted the successful integration of AI in detecting and analyzing voice signals for various health applications, with significant advancements in standardization, privacy, and clinical validation processes.DiscussionThe symposium underscored the importance of interdisciplinary collaboration to address the technical, ethical, and clinical challenges in the field of voice biomarkers. While AI models have shown promise in analyzing voice data, challenges such as data variability, security, and scalability remain. Future efforts must focus on refining data collection standards, advancing ethical AI practices, and ensuring diverse dataset inclusion to improve model robustness. By fostering collaboration among researchers, clinicians, and technologists, the symposium laid a foundation for future innovations in AI-driven voice analysis for healthcare diagnostics and treatment.

Clinicians use the patient's voice intuitively to evaluate general health and frailty. Voice is an emerging health indicator but has been scarcely studied in the context of frailty. This study explored voice parameters as possible predictors of frailty in older adults. Fifty-three participants over 70 years old were recruited from rehabilitation wards at a tertiary medical center. Participants' frailty was assessed using Rockwood frailty index and they were classified as most-frail (n = 33, 68%) or less-frail (n = 20, 32%). Participants were recorded counting from 1 to 10 and backwards using a smartphone recording application. The following voice biomarkers were derived: peak and average volume, peak/average volume ratio, pauses' total length, and pause length standard deviation. The most-frail group had a higher peak volume/average volume ratio (p = 0.03) and greater variance in lengths of pauses between speech segments (p = 0.002). These parameters indicate greater speech irregularity in the most-frail, compared to the less-frail. The most-frail group also had a longer total duration of pauses (p = 0.02). No statistically significant difference was found in peak and average volume (p = 0.75 and 0.39). Most-frail participants' speech had different characteristics, compared to participants in the less-frail group. This is a first step to developing an AI-based frailty assessment tool that can assist in identifying our most vulnerable patients. [ABSTRACT FROM AUTHOR]

Recent research has achieved a great classification rate for separating healthy people from those with Parkinson's disease (PD) using speech and the voice. However, these studies have primarily treated early and advanced stages of PD as equal entities, neglecting the distinctive speech impairments and other symptoms that vary across the different stages of the disease. To address this limitation, and improve diagnostic precision, this study assesses the selected acoustic features of dysphonia, as they relate to PD and the Hoehn and Yahr stages, by combining various preprocessing techniques and multiple classification algorithms, to create a comprehensive and robust solution for classification tasks. The dysphonia features extracted from the three sustained Korean vowels /아/(a), /이/(i), and /우/(u) exhibit diversity and strong correlations. To address this issue, the analysis of variance F-Value feature selection classifier from scikit-learn was employed, to identify the topmost relevant features. Additionally, to overcome the class imbalance problem, the synthetic minority over-sampling technique was utilized. To ensure fair comparisons, and mitigate the influence of individual classifiers, four commonly used machine learning classifiers, namely random forest (RF), support vector machine (SVM), k-nearest neighbor (kNN), and multi-layer perceptron (MLP), were employed. This approach enables a comprehensive evaluation of the feature extraction methods, and minimizes the variance in the final classification models. The proposed hybrid machine learning pipeline using the acoustic features of sustained vowels efficiently detects the early and mid-advanced stages of PD with a detection accuracy of 95.48%, and with a detection accuracy of 86.62% for the 4-stage, and a detection accuracy of 89.48% for the 3-stage classification of PD. This study successfully demonstrates the significance of utilizing the diverse acoustic features of dysphonia in the classification of PD and its stages. [ABSTRACT FROM AUTHOR]

The goal of this article is to present and compare recent approaches which use speech and voice analysis as biomarkers for screening tests and monitoring of some diseases. The article takes into account metabolic, respiratory, cardiovascular, endocrine, and nervous system disorders. A selection of articles was performed to identify studies that assess voice features quantitatively in selected disorders by acoustic and linguistic voice analysis. Information was extracted from each paper in order to compare various aspects of datasets, speech parameters, methods of applied analysis and obtained results. 110 research papers were reviewed and 47 databases were summarized. Speech analysis is a promising method for early diagnosis of certain disorders. Advanced computer voice analysis with machine learning algorithms combined with the widespread availability of smartphones allows diagnostic analysis to be conducted during the patient's visit to the doctor or at the patient's home during a telephone conversation. Speech analysis is a simple, low-cost, non-invasive and easy-toprovide method of medical diagnosis. These are remarkable advantages, but there are also disadvantages. The effectiveness of disease diagnoses varies from 65% up to 99%. For that reason it should be treated as a medical screening test and should be an indication of the need for classic medical tests. [ABSTRACT FROM AUTHOR]

Over the past decades, many machine-learning- and artificial-intelligence-based technologies have been created to deduce biometric or bio-relevant parameters of speakers from their voice. These voice profiling technologies have targeted a wide range of parameters, from diseases to environmental factors, based largely on the fact that they are known to influence voice. Recently, some have also explored the prediction of parameters whose influence on voice is not easily observable through data-opportunistic biomarker discovery techniques. However, given the enormous range of factors that can possibly influence voice, more informed methods for selecting those that may be potentially deducible from voice are needed. To this end, this paper proposes a simple path-finding algorithm that attempts to find links between vocal characteristics and perturbing factors using cytogenetic and genomic data. The links represent reasonable selection criteria for use by computational by profiling technologies only, and are not intended to establish any unknown biological facts. The proposed algorithm is validated using a simple example from medical literature—that of the clinically observed effects of specific chromosomal microdeletion syndromes on the vocal characteristics of affected people. In this example, the algorithm attempts to link the genes involved in these syndromes to a single example gene (FOXP2) that is known to play a broad role in voice production. We show that in cases where strong links are exposed, vocal characteristics of the patients are indeed reported to be correspondingly affected. Validation experiments and subsequent analyses confirm that the methodology could be potentially useful in predicting the existence of vocal signatures in naïve cases where their existence has not been otherwise observed. [ABSTRACT FROM AUTHOR]

Background: The application of polygenic risk scores (PRSs) in major depressive disorder (MDD) detection is constrained by its simplicity and uncertainty. One promising way to further extend its usability is fusion with other biomarkers. This study constructed an MDD biomarker by combining the PRS and voice features and evaluated their ability based on large clinical samples. Methods: We collected genome-wide sequences and utterances edited from clinical interview speech records from 3,580 women with recurrent MDD and 4,016 healthy people. Then, we constructed PRS as a gene biomarker by p value-based clumping and thresholding and extracted voice features using the i-vector method. Using logistic regression, we compared the ability of gene or voice biomarkers with the ability of both in combination for MDD detection. We also tested more machine learning models to further improve the detection capability. Results: With a p -value threshold of 0.005, the combined biomarker improved the area under the receiver operating characteristic curve (AUC) by 9.09% compared to that of genes only and 6.73% compared to that of voice only. Multilayer perceptron can further heighten the AUC by 3.6% compared to logistic regression, while support vector machine and random forests showed no better performance. Conclusion: The addition of voice biomarkers to genes can effectively improve the ability to detect MDD. The combination of PRS and voice biomarkers in MDD detection is feasible. This study provides a foundation for exploring the clinical application of genetic and voice biomarkers in the diagnosis of MDD. [ABSTRACT FROM AUTHOR]

Background: The application of polygenic risk scores (PRSs) in major depressive disorder (MDD) detection is constrained by its simplicity and uncertainty. One promising way to further extend its usability is fusion with other biomarkers. This study constructed an MDD biomarker by combining the PRS and voice features and evaluated their ability based on large clinical samples.Methods: We collected genome-wide sequences and utterances edited from clinical interview speech records from 3,580 women with recurrent MDD and 4,016 healthy people. Then, we constructed PRS as a gene biomarker by p value-based clumping and thresholding and extracted voice features using the i-vector method. Using logistic regression, we compared the ability of gene or voice biomarkers with the ability of both in combination for MDD detection. We also tested more machine learning models to further improve the detection capability.Results: With a p-value threshold of 0.005, the combined biomarker improved the area under the receiver operating characteristic curve (AUC) by 9.09% compared to that of genes only and 6.73% compared to that of voice only. Multilayer perceptron can further heighten the AUC by 3.6% compared to logistic regression, while support vector machine and random forests showed no better performance.Conclusion: The addition of voice biomarkers to genes can effectively improve the ability to detect MDD. The combination of PRS and voice biomarkers in MDD detection is feasible. This study provides a foundation for exploring the clinical application of genetic and voice biomarkers in the diagnosis of MDD.

Sleepiness is a major public and personal health issue. Measuring sleepiness in patients' everyday living conditions would represent a significant advancement in managing them. Voice is a signal linked to numerous health dimensions, including sleepiness. However, previous research has focused on short-term subjective sleepiness, using corpora with questionable medical validity and without measuring the specificity of identified voice biomarkers. In this article, we estimate different symptoms and long-term sleepiness-related syndromes in hypersomnolent patients. To achieve this, we have developed machine learning models that identify biomarkers that are sensitive and specific to sleepiness, reaching classification performances (Unweighted Average Recall) above 75%. Importantly, we only used statistical functions (decorrelation, Principal Component Analysis, Sensitivity test, linear classifier) so that this model remains simple and explainable to collaborating clinicians. We then leverage this explainability to identify specific vocal and speech manifestations for each type of sleepiness. By combining objective measures and the analysis of vocal characteristics, our approach provides a comprehensive understanding of long-term sleepiness and enhances patient care and management. This research holds great potential for advancing the field of digital health and contributing to improved well-being for individuals affected by sleepiness-related conditions. • Voice allows for the estimation of symptoms and syndromes related to sleepiness. • Subjective sleepiness interferes with the acoustic quality of voice. • Physiological sleepiness interferes preferentially with reading quality. [ABSTRACT FROM AUTHOR]

Voice technology has grown tremendously in recent years and using voice as a biomarker has also been gaining evidence. We demonstrate the potential of voice in serving as a deep phenotype for Parkinson's Disease (PD), the second most common neurodegenerative disorder worldwide, by presenting methodology for voice signal processing for clinical analysis. Detection of PD symptoms typically requires an exam by a movement disorder specialist and can be hard to access and inconsistent in findings. A vocal digital biomarker could supplement the cumbersome existing manual exam by detecting and quantifying symptoms to guide treatment. Specifically, vocal biomarkers of PD are a potentially effective method of assessing symptoms and severity in daily life, which is the focus of the current research. We analyzed a database of PD patient and non-PD subjects containing voice recordings that were used to extract paralinguistic features, which served as inputs to machine learning models to predict PD severity. The results are presented here and the limitations are discussed given the nature of the recordings. We note that our methodology only advances biomarker research and is not cleared for clinical use. Specifically, we demonstrate that conventional machine learning models applied to voice signals can be used to differentiate participants with PD who exhibit little to no symptoms from healthy controls. This work highlights the potential of voice to be used for early detection of PD and indicates that voice may serve as a deep phenotype for PD, enabling precision medicine by improving the speed, accuracy, accessibility, and cost of PD management. [ABSTRACT FROM AUTHOR]

Depression is a common mental health problem leading to significant disability world wide. Depression is not only common but also commonly co-occurs with other mental and neurological illnesses. Parkinson's Disease gives rise to symptoms directly impairing a person's ability to function. Early diagnosis and detection of depression can aid treatment, but diagnosis typically requires an interview with a health provider or structured diagnostic questionnaire. Thus, unobtrusive measures to monitor depression symptoms in daily life could have great utility in screening depression for clinical treatment. Vocal biomarkers of depression are a potentially effective method of assessing depression symptoms in daily life, which is the focus of the current research. We have a database of 921 unique patients with Parkinson's disease and their self assessment of whether they felt depressed or not. Voice recordings from these patients were used to extract paralinguistic features, which served as inputs to machine-learning and deep learning techniques to predict depression. The results are presented here and the limitations are discussed given the nature of the recordings which lack language content. Our models achieved accuracies as high as 0.77 in classifying depressed and non-depressed subjects accurately using their voice features and PD severity. We found depression and severity of Parkinson's Disease had a correlation coefficient of 0.3936, providing a valuable feature when predicting depression from voice. Our results indicate a clear correlation between feeling depressed and the severity of the Parkinson's disease. Voice may be an effective digital biomarker to screen for depression among patients suffering from Parkinson's Disease.