Age-Related Changes in Speech and Voice: Spectral and Cepstral Measures

How ready is speech-to-text for psychological language research? Evaluating the validity of AI-generated English transcripts for analyzing free-spoken responses in younger and older adults

For the longest time, the gold standard in preparing spoken language corpora for text analysis in psychology was using human transcription. However, such standard comes at extensive cost, and creates barriers to quantitative spoken language analysis that recent advances in speech-to-text technology could address. The current study quantifies the accuracy of AI-generated transcripts compared to human-corrected transcripts across younger (n = 100) and older (n = 92) adults and two spoken language tasks. Further, it evaluates the validity of Linguistic Inquiry and Word Count (LIWC)-features extracted from these two kinds of transcripts, as well as transcripts specifically prepared for LIWC analyses via tagging. We find that overall, AI-generated transcripts are highly accurate with a word error rate of 2.50% to 3.36%, albeit being slightly less accurate for younger compared to older adults. LIWC features extracted from either transcripts are highly correlated, while the tagging procedure significantly alters filler word categories. Based on these results, automatic speech-to-text appears to be ready for psychological language research when using spoken language tasks in relatively quiet environments, unless filler words are of interest to researchers.

Factors influencing speech improvement following maxillary complete-arch implant-supported restorations: A retrospective study

OBJECTIVE

This study explored factors affecting speech improvement in patients with an edentulous maxilla after the delivery of a complete-arch implant-supported fixed dental prosthesis (IFDP).

MATERIALS AND METHODS

Patients who had received IFDP for edentulous maxilla were enrolled, and various potential speech improvement-related factors were considered, including patient demographics, anterior residual bone volume, preoperative facial features, preoperative acoustic parameters, and adaptation time. Acoustic analysis and perceptual ratings were used to assess three fricatives [s], [f], and [ɕ]. Correlation and regression analyses were conducted to assess the association between changes in fricatives and potential factors (α = .05).

RESULTS

The study included 50 patients (18 females and 32 males, aged 50.62 ± 15.71 years, range 19-76). Significant correlations were found among the change in the center of gravity (ΔCoG) of [s] and anterior residual bone volume, zygomatic implants number and proportion (p < .05). These correlations were largely mirrored in the perceptual score (ΔPS) changes. After controlling for age, sex, preoperative acoustic parameters, and adaptation time, the ΔCoG and ΔPS of fricatives were mainly correlated with the anterior residual bone volume, preoperative acoustic parameters, and adaptation time.

CONCLUSION

Speech improvements post-IFDP delivery are mainly related to preoperative speech characteristics, anterior residual bone volume, and adaptation time. The residual bone volume's impact on consonants varies with specific articulatory gestures. This study provides insights into forecasting speech outcomes following IFDP restoration and provides recommendations and methods for data collection in developing future prediction models.

Orofacial and cervical myofunctional intervention programmes for older adults: A scoping review

INTRODUCTION

Senescent stomatognathic systems undergo anatomical changes that are not always compensated by physiology due to associated factors such as poorly fitted dentures, dental changes, and clinical conditions that trigger sarcopenia. When these conditions are not properly addressed, they can lead to nutritional deficiencies. Hence, this study aimed to map orofacial and cervical myofunctional intervention programmes for older adults, evaluate the current research in this area, and formulate suggestions for further investigation.

METHODS

Scoping review is based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews checklist. The EMBASE, LILACS, PubMed/Medline, Scopus, Web of Science, and grey literature databases were systematically searched. Studies addressing therapeutic programmes, myofunctional exercises, and their influence on the stomatognathic system of older adults were included. Studies with populations younger than 60 years, residing in long-term care facilities, homes, or hospital care, and with associated comorbidities were excluded, as well as secondary studies.

RESULTS

A total of 3098 studies were retrieved. After applying the eligibility criteria, nine studies were eligible for this review. The nine studies were published in English between 2008 and 2021 in the United States, Japan, and South Korea. The participants' mean ages ranged from 64 to 81 years, with a preponderance of females. Six of the studies divided the sample into intervention and control groups. Two studies referred to programmes to promote oral function through facial expression and tongue muscle exercises, salivary gland massage, and swallowing biomechanics. Seven studies report the positive influence of myofunctional exercises on the stomatognathic system.

CONCLUSION

The mapping concluded that therapeutic programmes, including educational actions, help prevent orofacial myofunctional disorders, and improve the functionality of the stomatognathic system.

Cepstral Measurements: A Comparison of Results Between Singing and Non-Singing Individuals

OBJECTIVE

To analyze and compare cepstral measurements in singing and non-singing men and women to understand if vocal adaptations of singers reflect greater cepstral measurement results.

METHOD

The study included 91 vocally healthy individuals, consisting of 60 erudite or popular singers (30 males, 30 females) and 31 non-singers (16 males, 15 females). The Cepstral Peak Prominence (CPP) and Cepstral Peak Prominence-Smoothed (CPPS) measures were carried out using the sustained /a/ vowel recorded with a unidirectional condenser microphone associated with a sound card on the computer. The Praat software (version 6.1.16) was used to extract the CPP and CPPS measurements. The intra and intergroup values obtained for each sex were compared using the t-test with a 5% significance level.

RESULTS

There was a CPP (P = 0,000) and CPPS (P = 0,000 and P = 0,001) value difference between the sexes in both groups. Cepstral measurements showed no statistically significant difference between singing and non-singing participants (male: P = 0,778 and P = 0,622; female: P = 0,622 and P = 0,460).

CONCLUSION

Men presented higher CPP and CPPS values than women, which may be related to the presence of the physiological posterior glottic cleft in women. The knowledge of the cepstral values of singers shows that singing specificities may not reflect in all vocal evaluations.

Normative Values of Cepstral Peak Prominence Measures in Typical Speakers by Sex, Speech Stimuli, and Software Type Across the Life Span

PURPOSE

The purpose of this study was to determine normative values for cepstral peak prominence measures across the life span as a function of sex using clinically relevant stimuli (/ɑ/, /i/, and two sentences of The Rainbow Passage) and two commonly used software types: Praat (Version 6.0.50) and Analysis of Dysphonia in Speech and Voice (ADSV).

METHOD

One hundred fifty speakers (75 males, 75 females; evenly distributed into three age groups) without voice disorders aged 18-91 years were recorded via headset microphone in a sound-treated booth. Cepstral measures were analyzed using common analysis methods in Praat and ADSV by sex, stimuli, and software type. Kruskal-Wallis tests and post hoc Mood's Median tests for significant factors were performed on cepstral measures to assess the effects of age group, sex, stimuli, and software type.

RESULTS

The results revealed statistically significant effects of sex, stimuli, and software type on cepstral measures, but no statistical effect of age group on cepstral values. Females had lower average cepstral values compared to males. Across stimuli, the highest average cepstral measure was found for sustained /ɑ/, followed by sustained /i/, and then of the two sentences of The Rainbow Passage. Average cepstral measures in Praat were higher than those from ADSV.

CONCLUSIONS

The current work did not find a statistical effect of age group on cepstral values; thus, normative cepstral values were reported by sex, stimuli, and software type. Future work should examine the applicability of these normative values for discriminating speakers with and without voice disorders.

Acoustic Measures of Voice and Physiologic Measures of Autonomic Arousal During Speech as a Function of Cognitive Load in Older Adults

OBJECTIVES/HYPOTHESIS

The purpose of this study was to determine the relationships among cognitive loading, autonomic arousal, and acoustic measures of voice in healthy older adults.

STUDY DESIGN

Prospective and observational.

METHODS

Twelve healthy older adults (six females) produced a sentence containing an embedded Stroop task in each of two cognitive load conditions: congruent and incongruent. Three physiologic measures of autonomic arousal (pulse volume amplitude, pulse period, and skin conductance response amplitude) and four acoustic measures of voice (cepstral peak prominence, low-to-high spectral energy ratio, fundamental frequency, and sound pressure level) were analyzed in each cognitive load condition.

RESULTS

A logistic regression model was used to predict the cognitive load condition using participant as a categorical predictor and the four acoustic measures and three autonomic measures as continuous predictors. Skin conductance response amplitude and pulse volume amplitude were both predictive of cognitive load; however, no acoustic measures of voice were statistically significant predictors of cognitive load for older adults.

CONCLUSIONS

These findings support the idea that increased cognitive load is associated with increased autonomic nervous system activity in older adults. The lack of changes in acoustic measures of voice with increased cognitive load may result from age-related changes in vocal quality and speech subsystems.

[Current methods of acoustic analysis of voice: a review]

Acoustic analysis of the voice, as an objective, quantitative, non-invasive and reproducible method for the evaluation of voice quality, can be used to detect and analyze the acoustic characteristics of normal, artistic or pathological voice. With the development of medicine, physics, statistics, and artificial intelligence technology, there are new advances in the study of voice acoustic analysis, especially in terms of acoustic parameters. In addition, artificial neural networks can be used to perform complex multi-parameter analysis, which greatly improves the efficiency of acoustic analysis. This paper provides an overview of the methods of acoustic analysis and its latest development.

Proteomic analysis reveals that aging rabbit vocal folds are more vulnerable to changes caused by systemic dehydration

BACKGROUND

Older adults are more prone to develop systemic dehydration. Systemic dehydration has implications for vocal fold biology by affecting gene and protein expression. The objective of this study was to quantify vocal fold protein changes between two age groups and hydration status, and to investigate the interaction of age and hydration status on protein expression, which has not been investigated in the context of vocal folds before. Comparative proteomics was used to analyze the vocal fold proteome of 6.5-month-old and > 3-year-old rabbits subjected to water ad libitum or water volume restriction protocol.

RESULTS

Young and older adult rabbits (n = 22) were either euhydrated (water ad libitum) or dehydrated by water volume restriction. Dehydration was confirmed by body weight loss of - 5.4% and - 4.6% in young and older groups, respectively, and a 1.7-fold increase of kidney renin gene expression in the young rabbits. LC-MS/MS identified 2286 proteins in the rabbit vocal folds of young and older adult rabbits combined. Of these, 177, 169, and 81 proteins were significantly (p ≤ 0.05) affected by age, hydration status, or the interaction of both factors, respectively. Analysis of the interaction effect revealed 32 proteins with opposite change patterns after dehydration between older and young rabbit vocal folds, while 31 proteins were differentially regulated only in the older adult rabbits and ten only in the young rabbits in response to systemic dehydration. The magnitude of changes for either up or downregulated proteins was higher in the older rabbits. These proteins are predominantly related to structural components of the extracellular matrix and muscle layer, suggesting a disturbance in the viscoelastic properties of aging vocal fold tissue, especially when subjected to systemic dehydration.

CONCLUSIONS

Water restriction is a laboratory protocol to assess systemic dehydration-related changes in the vocal fold tissue that is translatable to human subjects. Our findings showed a higher number of proteins differentially regulated with a greater magnitude of change in the vocal folds of older adult rabbits in the presence of systemic dehydration compared to younger rabbits. The association of these proteins with vocal fold structure and biomechanical properties suggests that older human subjects may be more vulnerable to the effects of systemic dehydration on vocal function. The clinical implications of these protein changes warrant more investigation, but age should be taken into consideration when evaluating vocal treatment recommendations that interfere with body fluid balance.

Bright Voice Quality and Fundamental Frequency Variation in Non-binary Speakers

OBJECTIVES

1) To investigate if vocal variation produced by assigned-female-at-birth (AFAB) non-binary people differed from vocal variation produced by cisgender (cis) participants. Cue values produced by non-binary participants were predicted to differ from those values produced by cisgender participants. 2) To determine if previous subjective assessments of bright voice quality in AFAB non-binary participants were quantifiable, and if so, if non-binary and cisgender participants differed in their voice quality production.

STUDY DESIGN

A quantitative comparative research design.

METHODS

Phonetic and statistical analyses of continuous speech samples produced by AFAB non-binary and cisgender participants. Vocal cues were mean fundamental frequency (F0) and bright voice quality, measured by cepstral peak prominence-smoothed and spectral slope, with speaker gender as the predictor.

RESULTS

At the group level, non-binary participants produced intermediate F0 values - significantly lower than the cis women's and significantly higher than the cis men's. Individually, the majority of non-binary participants produced mean F0 in this intermediate range. Non-binary participants produced significantly less negative spectral slope and higher cepstral peak prominence-smoothed, indicative of a brighter, more resonant voice quality. Individual-level results indicated that vocal training and vocal tract physiology did not fully account for the results found.

CONCLUSION

Participants' agency, particularly their motivation to alter vocal output to avoid being misgendered, has an effect on the AFAB non-binary participants' F0 production and potentially their voice quality. The majority of AFAB non-binary participants uniquely produced the cue combination of intermediate F0 and bright voice quality.

The Acoustic Dissection of Cough: Diving Into Machine Listening-based COVID-19 Analysis and Detection

OBJECTIVES

The coronavirus disease 2019 (COVID-19) has caused a crisis worldwide. Amounts of efforts have been made to prevent and control COVID-19's transmission, from early screenings to vaccinations and treatments. Recently, due to the spring up of many automatic disease recognition applications based on machine listening techniques, it would be fast and cheap to detect COVID-19 from recordings of cough, a key symptom of COVID-19. To date, knowledge of the acoustic characteristics of COVID-19 cough sounds is limited but would be essential for structuring effective and robust machine learning models. The present study aims to explore acoustic features for distinguishing COVID-19 positive individuals from COVID-19 negative ones based on their cough sounds.

METHODS

By applying conventional inferential statistics, we analyze the acoustic correlates of COVID-19 cough sounds based on the ComParE feature set, i.e., a standardized set of 6,373 acoustic higher-level features. Furthermore, we train automatic COVID-19 detection models with machine learning methods and explore the latent features by evaluating the contribution of all features to the COVID-19 status predictions.

RESULTS

The experimental results demonstrate that a set of acoustic parameters of cough sounds, e.g., statistical functionals of the root mean square energy and Mel-frequency cepstral coefficients, bear essential acoustic information in terms of effect sizes for the differentiation between COVID-19 positive and COVID-19 negative cough samples. Our general automatic COVID-19 detection model performs significantly above chance level, i.e., at an unweighted average recall (UAR) of 0.632, on a data set consisting of 1,411 cough samples (COVID-19 positive/negative: 210/1,201).

CONCLUSIONS

Based on the acoustic correlates analysis on the ComParE feature set and the feature analysis in the effective COVID-19 detection approach, we find that several acoustic features that show higher effects in conventional group difference testing are also higher weighted in the machine learning models.

Voice, Speech, and Clinical Aspects During Pregnancy: A Longitudinal Study

BACKGROUND

Pregnancy involves anatomical, physiological, and metabolic changes in a woman's body. However, the effects of these changes on the voice remains unclear, particularly regarding the clinical characteristics.

OBJECTIVES

We aimed to evaluate changes in vocal and speech acoustic measures and the relationship between them and clinical aspects in women during pregnancy.

METHOD

A prospective, longitudinal study was carried out with 41 low risk, adult, pregnant women, followed for prenatal care. Demographic and anthropometric data as well as lifestyle habits and health conditions were collected. Voice recordings of sustained vowels, and automatic and spontaneous speech were held over each trimester and analyzed by PRAAT®to evaluate acoustic, aerodynamic, and articulatory measures.

RESULTS

There were no changes in fundamental frequency, jitter, shimmer, and harmony to noise ratio during pregnancy. Maximum phonation time (MPT), pause rate, and pause duration reduced at the end of pregnancy. MPT was lower in sedentary pregnant women. The fundamental frequency peak rate was higher in eutrophic participants and lower in the third trimester in women with BMI ≥25 kg/m2. Pause rate was higher in pregnant women with BMI ≥25 kg/m2. There was no relationship between sleep quality, reflux, and vocal symptoms and acoustic and aerodynamic measures.

CONCLUSIONS

Differences were shown in MPT and temporal pause measurements during pregnancy. Acoustic measurements did not change. There was a relationship between acoustic and aerodynamic measures and clinical variables (BMI, physical activity, and body mass gain).