Development of digital voice biomarkers and associations with cognition, cerebrospinal biomarkers, and neural representation in early Alzheimer's disease

Digital remote assessment of speech acoustics in cognitively unimpaired adults: feasibility, reliability and associations with amyloid pathology

BACKGROUND

Digital speech assessment has potential relevance in the earliest, preclinical stages of Alzheimer's disease (AD). We evaluated the feasibility, test-retest reliability, and association with AD-related amyloid-beta (Aβ) pathology of speech acoustics measured over multiple assessments in a remote setting.

METHODS

Fifty cognitively unimpaired adults (Age 68 ± 6.2 years, 58% female, 46% Aβ-positive) completed remote, tablet-based speech assessments (i.e., picture description, journal-prompt storytelling, verbal fluency tasks) for five days. The testing paradigm was repeated after 2-3 weeks. Acoustic speech features were automatically extracted from the voice recordings, and mean scores were calculated over the 5-day period. We assessed feasibility by adherence rates and usability ratings on the System Usability Scale (SUS) questionnaire. Test-retest reliability was examined with intraclass correlation coefficients (ICCs). We investigated the associations between acoustic features and Aβ-pathology, using linear regression models, adjusted for age, sex and education.

RESULTS

The speech assessment was feasible, indicated by 91.6% adherence and usability scores of 86.0 ± 9.9. High reliability (ICC ≥ 0.75) was found across averaged speech samples. Aβ-positive individuals displayed a higher pause-to-word ratio in picture description (B = -0.05, p = 0.040) and journal-prompt storytelling (B = -0.07, p = 0.032) than Aβ-negative individuals, although this effect lost significance after correction for multiple testing.

CONCLUSION

Our findings support the feasibility and reliability of multi-day remote assessment of speech acoustics in cognitively unimpaired individuals with and without Aβ-pathology, which lays the foundation for the use of speech biomarkers in the context of early AD.

Digital Neuropsychology beyond Computerized Cognitive Assessment: Applications of Novel Digital Technologies

Compared with other health disciplines, there is a stagnation in technological innovation in the field of clinical neuropsychology. Traditional paper-and-pencil tests have a number of shortcomings, such as low-frequency data collection and limitations in ecological validity. While computerized cognitive assessment may help overcome some of these issues, current computerized paradigms do not address the majority of these limitations. In this paper, we review recent literature on the applications of novel digital health approaches, including ecological momentary assessment, smartphone-based assessment and sensors, wearable devices, passive driving sensors, smart homes, voice biomarkers, and electronic health record mining, in neurological populations. We describe how each digital tool may be applied to neurologic care and overcome limitations of traditional neuropsychological assessment. Ethical considerations, limitations of current research, as well as our proposed future of neuropsychological practice are also discussed.

A novel speech analysis algorithm to detect cognitive impairment in a Spanish population

Objective

Early detection of cognitive impairment in the elderly is crucial for diagnosis and appropriate care. Brief, cost-effective cognitive screening instruments are needed to help identify individuals who require further evaluation. This study presents preliminary data on a new screening technology using automated voice recording analysis software in a Spanish population.

Method

Data were collected from 174 Spanish-speaking individuals clinically diagnosed as cognitively normal (CN, n = 87) or impaired (mild cognitive impairment [MCI], n = 63; all-cause dementia, n = 24). Participants were recorded performing four common language tasks (Animal fluency, alternating fluency [sports and fruits], phonemic "F" fluency, and Cookie Theft Description). Recordings were processed via text-transcription and digital-signal processing techniques to capture neuropsychological variables and audio characteristics. A training sample of 122 subjects with similar demographics across groups was used to develop an algorithm to detect cognitive impairment. Speech and task features were used to develop five independent machine learning (ML) models to compute scores between 0 and 1, and a final algorithm was constructed using repeated cross-validation. A socio-demographically balanced subset of 52 participants was used to test the algorithm. Analysis of covariance (ANCOVA), covarying for demographic characteristics, was used to predict logistically-transformed algorithm scores.

Results

Mean logit algorithm scores were significantly different across groups in the testing sample (p < 0.01). Comparisons of CN with impaired (MCI + dementia) and MCI groups using the final algorithm resulted in an AUC of 0.93/0.90, with overall accuracy of 88.4%/87.5%, sensitivity of 87.5/83.3, and specificity of 89.2/89.2, respectively.

Conclusion

Findings provide initial support for the utility of this automated speech analysis algorithm as a screening tool for cognitive impairment in Spanish speakers. Additional study is needed to validate this technology in larger and more diverse clinical populations.

Longitudinal observational cohort study: Speech for Intelligent cognition change tracking and DEtection of Alzheimer's Disease (SIDE-AD)

INTRODUCTION

There is emerging evidence that speech may be a potential indicator and manifestation of early Alzheimer's disease (AD) pathology. Therefore, the University of Edinburgh and Sony Research have partnered to create the Speech for Intelligent cognition change tracking and DEtection of Alzheimer's Disease (SIDE-AD) study, which aims to develop digital speech-based biomarkers for use in neurodegenerative disease.

METHODS AND ANALYSIS

SIDE-AD is an observational longitudinal study, collecting samples of spontaneous speech. Participants are recruited from existing cohort studies as well as from the National Health Service (NHS)memory clinics in Scotland. Using an online platform, participants record a voice sample talking about their brain health and rate their mood, anxiety and apathy. The speech biomarkers will be analysed longitudinally, and we will use machine learning and natural language processing technology to automate the assessment of the respondents' speech patterns.

ETHICS AND DISSEMINATION

The SIDE-AD study has been approved by the NHS Research Ethics Committee (REC reference: 23/WM/0153, protocol number AC23046, IRAS Project ID 323311) and received NHS management approvals from Lothian, Fife and Forth Valley NHS boards. Our main ethical considerations pertain to the remote administration of the study, such as taking remote consent. To address this, we implemented a consent process, whereby the first step of the consent is done entirely remotely but a member of the research team contacts the participant over the phone to consent participants to the optional, most sensitive, elements of the study. Results will be presented at conferences, published in peer-reviewed journals and communicated to study participants.

Spoken Language Alterations can Predict Phenoconversion in Isolated Rapid Eye Movement Sleep Behavior Disorder: A Multicenter Study

OBJECTIVE

This study assessed the relationship between speech and language impairment and outcome in a multicenter cohort of isolated/idiopathic rapid eye movement (REM) sleep behavior disorder (iRBD).

METHODS

Patients with iRBD from 7 centers speaking Czech, English, German, French, and Italian languages underwent a detailed speech assessment at baseline. Story-tale narratives were transcribed and linguistically annotated using fully automated methods based on automatic speech recognition and natural language processing algorithms, leading to the 3 distinctive linguistic and 2 acoustic patterns of language deterioration and associated composite indexes of their overall severity. Patients were then prospectively followed and received assessments for parkinsonism or dementia during follow-up. The Cox proportional hazard was performed to evaluate the predictive value of language patterns for phenoconversion over a follow-up period of 5 years.

RESULTS

Of 180 patients free of parkinsonism or dementia, 156 provided follow-up information. After a mean follow-up of 2.7 years, 42 (26.9%) patients developed neurodegenerative disease. Patients with higher severity of linguistic abnormalities (hazard ratio [HR = 2.35]) and acoustic abnormalities (HR = 1.92) were more likely to develop a defined neurodegenerative disease, with converters having lower content richness (HR = 1.74), slower articulation rate (HR = 1.58), and prolonged pauses (HR = 1.46). Dementia-first (n = 16) and parkinsonism-first with mild cognitive impairment (n = 9) converters had higher severity of linguistic abnormalities than parkinsonism-first with normal cognition converters (n = 17).

INTERPRETATION

Automated language analysis might provide a predictor of phenoconversion from iRBD into synucleinopathy subtypes with cognitive impairment, and thus can be used to stratify patients for neuroprotective trials. ANN NEUROL 2024;95:530-543.

Consensus on rapid screening for prodromal Alzheimer's disease in China

Alzheimer's disease (AD) is a common cause of dementia, characterised by cerebral amyloid-β deposition, pathological tau and neurodegeneration. The prodromal stage of AD (pAD) refers to patients with mild cognitive impairment (MCI) and evidence of AD's pathology. At this stage, disease-modifying interventions should be used to prevent the progression to dementia. Given the inherent heterogeneity of MCI, more specific biomarkers are needed to elucidate the underlying AD's pathology. Although the uses of cerebrospinal fluid and positron emission tomography are widely accepted methods for detecting AD's pathology, their clinical applications are limited by their high costs and invasiveness, particularly in low-income areas in China. Therefore, to improve the early detection of Alzheimer's disease (AD) pathology through cost-effective screening methods, a panel of 45 neurologists, psychiatrists and gerontologists was invited to establish a formal consensus on the screening of pAD in China. The supportive evidence and grades of recommendations are based on a systematic literature review and focus group discussion. National meetings were held to allow participants to review, vote and provide their expert opinions to reach a consensus. A majority (two-thirds) decision was used for questions for which consensus could not be reached. Recommended screening methods are presented in this publication, including neuropsychological assessment, peripheral biomarkers and brain imaging. In addition, a general workflow for screening pAD in China is established, which will help clinicians identify individuals at high risk and determine therapeutic targets.

Early ultrasonic vocalization deficits and related thyroarytenoid muscle pathology in the transgenic TgF344-AD rat model of Alzheimer's disease

Background

Alzheimer's disease (AD) is a progressive neurologic disease and the most common cause of dementia. Classic pathology in AD is characterized by inflammation, abnormal presence of tau protein, and aggregation of β-amyloid that disrupt normal neuronal function and lead to cell death. Deficits in communication also occur during disease progression and significantly reduce health, well-being, and quality of life. Because clinical diagnosis occurs in the mid-stage of the disease, characterizing the prodrome and early stages in humans is currently challenging. To overcome these challenges, we use the validated TgF344-AD (F344-Tg(Prp-APP, Prp-PS1)19/Rrrc) transgenic rat model that manifests cognitive, behavioral, and neuropathological dysfunction akin to AD in humans.

Objectives

The overarching goal of our work is to test the central hypothesis that pathology and related behavioral deficits such as communication dysfunction in part manifest in the peripheral nervous system and corresponding target tissues already in the early stages. The primary aims of this study are to test the hypotheses that: (1) changes in ultrasonic vocalizations (USV) occur in the prodromal stage at 6 months of age and worsen at 9 months of age, (2) inflammation as well as AD-related pathology can be found in the thyroarytenoid muscle (TA) at 12 months of age (experimental endpoint tissue harvest), and to (3) demonstrate that the TgF344-AD rat model is an appropriate model for preclinical investigations of early AD-related vocal deficits.

Methods

USVs were collected from male TgF344-AD (N = 19) and wildtype (WT) Fischer-344 rats (N = 19) at 6 months (N = 38; WT: n = 19; TgF344-AD: n = 19) and 9 months of age (N = 18; WT: n = 10; TgF344-AD: n = 8) and acoustically analyzed for duration, mean power, principal frequency, low frequency, high frequency, peak frequency, and call type. RT-qPCR was used to assay peripheral inflammation and AD-related pathology via gene expressions in the TA muscle of male TgF344-AD rats (n = 6) and WT rats (n = 6) at 12 months of age.

Results

This study revealed a significant reduction in mean power of ultrasonic calls from 6 to 9 months of age and increased peak frequency levels over time in TgF344-AD rats compared to WT controls. Additionally, significant downregulation of AD-related genes Uqcrc2, Bace2, Serpina3n, and Igf2, as well as downregulation of pro-inflammatory gene Myd88 was found in the TA muscle of TgF344-AD rats at 12 months of age.

Discussion

Our findings demonstrate early and progressive vocal deficits in the TgF344-AD rat model. We further provide evidence of dysregulation of AD-pathology-related genes as well as inflammatory genes in the TA muscles of TgF344-AD rats in the early stage of the disease, confirming this rat model for early-stage investigations of voice deficits and related pathology.

Severity Classification Using Dynamic Time Warping-Based Voice Biomarkers for Patients With COVID-19: Feasibility Cross-Sectional Study

Background

In Japan, individuals with mild COVID-19 illness previously required to be monitored in designated areas and were hospitalized only if their condition worsened to moderate illness or worse. Daily monitoring using a pulse oximeter was a crucial indicator for hospitalization. However, a drastic increase in the number of patients resulted in a shortage of pulse oximeters for monitoring. Therefore, an alternative and cost-effective method for monitoring patients with mild illness was required. Previous studies have shown that voice biomarkers for Parkinson disease or Alzheimer disease are useful for classifying or monitoring symptoms; thus, we tried to adapt voice biomarkers for classifying the severity of COVID-19 using a dynamic time warping (DTW) algorithm where voice wavelets can be treated as 2D features; the differences between wavelet features are calculated as scores.

Objective

This feasibility study aimed to test whether DTW-based indices can generate voice biomarkers for a binary classification model using COVID-19 patients' voices to distinguish moderate illness from mild illness at a significant level.

Methods

We conducted a cross-sectional study using voice samples of COVID-19 patients. Three kinds of long vowels were processed into 10-cycle waveforms with standardized power and time axes. The DTW-based indices were generated by all pairs of waveforms and tested with the Mann-Whitney U test (α<.01) and verified with a linear discrimination analysis and confusion matrix to determine which indices were better for binary classification of disease severity. A binary classification model was generated based on a generalized linear model (GLM) using the most promising indices as predictors. The receiver operating characteristic curve/area under the curve (ROC/AUC) validated the model performance, and the confusion matrix calculated the model accuracy.

Results

Participants in this study (n=295) were infected with COVID-19 between June 2021 and March 2022, were aged 20 years or older, and recuperated in Kanagawa prefecture. Voice samples (n=110) were selected from the participants' attribution matrix based on age group, sex, time of infection, and whether they had mild illness (n=61) or moderate illness (n=49). The DTW-based variance indices were found to be significant (P<.001, except for 1 of 6 indices), with a balanced accuracy in the range between 79% and 88.6% for the /a/, /e/, and /u/ vowel sounds. The GLM achieved a high balance accuracy of 86.3% (for /a/), 80.2% (for /e/), and 88% (for /u/) and ROC/AUC of 94.8% (95% CI 90.6%-94.8%) for /a/, 86.5% (95% CI 79.8%-86.5%) for /e/, and 95.6% (95% CI 92.1%-95.6%) for /u/.

Conclusions

The proposed model can be a voice biomarker for an alternative and cost-effective method of monitoring the progress of COVID-19 patients in care.

Harnessing acoustic speech parameters to decipher amyloid status in individuals with mild cognitive impairment

Alzheimer's disease (AD) is a neurodegenerative condition characterized by a gradual decline in cognitive functions. Currently, there are no effective treatments for AD, underscoring the importance of identifying individuals in the preclinical stages of mild cognitive impairment (MCI) to enable early interventions. Among the neuropathological events associated with the onset of the disease is the accumulation of amyloid protein in the brain, which correlates with decreased levels of Aβ42 peptide in the cerebrospinal fluid (CSF). Consequently, the development of non-invasive, low-cost, and easy-to-administer proxies for detecting Aβ42 positivity in CSF becomes particularly valuable. A promising approach to achieve this is spontaneous speech analysis, which combined with machine learning (ML) techniques, has proven highly useful in AD. In this study, we examined the relationship between amyloid status in CSF and acoustic features derived from the description of the Cookie Theft picture in MCI patients from a memory clinic. The cohort consisted of fifty-two patients with MCI (mean age 73 years, 65% female, and 57% positive amyloid status). Eighty-eight acoustic parameters were extracted from voice recordings using the extended Geneva Minimalistic Acoustic Parameter Set (eGeMAPS), and several ML models were used to classify the amyloid status. Furthermore, interpretability techniques were employed to examine the influence of input variables on the determination of amyloid-positive status. The best model, based on acoustic variables, achieved an accuracy of 75% with an area under the curve (AUC) of 0.79 in the prediction of amyloid status evaluated by bootstrapping and Leave-One-Out Cross Validation (LOOCV), outperforming conventional neuropsychological tests (AUC = 0.66). Our results showed that the automated analysis of voice recordings derived from spontaneous speech tests offers valuable insights into AD biomarkers during the preclinical stages. These findings introduce novel possibilities for the use of digital biomarkers to identify subjects at high risk of developing AD.