Different Performances of Machine Learning Models to Classify Dysphonic and Non-Dysphonic Voices.

OBJECTIVE: To analyze the performance of 10 different machine learning (ML) classifiers for discrimination between dysphonic and non-dysphonic voices, using a variance threshold as a method for the selection and reduction of acoustic measurements used in the classifier. METHOD: We analyzed 435 samples of individuals (337 female and 98 male), with a mean age of 41.07 ± 13.73 years, of which 384 were dysphonic and 51 were non-dysphonic. From the sustained /ε/ vowel sample, 34 acoustic measurements were extracted, including traditional perturbation and noise measurements, cepstral/spectral measurements, and measurements based on nonlinear models. The variance method was used to select the best set of acoustic measurements. We tested the performance of the best-selected set with 10 ML classifiers using precision, sensitivity, specificity, accuracy, and F1-Score measurements. The kappa coefficient was used to verify the reproducibility between the two datasets (training and testing). RESULTS: The naive Bayes (NB) and stochastic gradient descent classifier (SGDC) models performed best in terms of accuracy, AUC, sensitivity, and specificity for a reduced dataset of 15 acoustic measures compared to the full dataset of 34 acoustic measures. SGDC and NB obtained the best performance results, with an accuracy of 0.91 and 0.76, respectively. These two classifiers presented moderate agreement, with a Kappa of 0.57 (SGDC) and 0.45 (NB). CONCLUSION: Among the tested models, the NB and SGDC models performed better in discriminating between dysphonic and non-dysphonic voices from a set of 15 acoustic measures.

Recognition of the Effect of Vocal Exercises by Fuzzy Triangular Naive Bayes, a Machine Learning Classifier: A Preliminary Analysis.

OBJECTIVES: Machine learning (ML) methods allow the development of expert systems for pattern recognition and predictive analysis of intervention outcomes. It has been used in Voice Sciences, mainly to discriminate between healthy and dysphonic voices. Parameter patterns of vocal acoustic analysis and vocal perceptual assessment can be evaluated by ML classifiers, such as the Fuzzy Triangular Naive Bayes (FTriangNB), after using techniques that improve the vocal quality of individuals with healthy or dysphonic voices. Thus, the goal of this study was to analyze the performance of the FTriangNB to detect patterns in the acoustic parameters and the auditory-perceptual assessment of 12 women with dysphonia and 12 vocally healthy women, after performing three vocal exercises (tongue trills, semi-occluded vocal tract exercise with a high-resistance straw - SOVTE, and over-articulation). METHODS: The FTriangNB classifier contained in the Fuzzy Class package was implemented in the data analysis software R Studio version 1.4.1106 for Macintosh. The confusion matrix was extracted, as well as the accuracy, the Kappa coefficient, and the class statistics. The final result was compared with those generated by FTriangNB with the same variables from the preapplication database of the exercises. RESULTS: The FTriangNB presented good accuracy (87.5%) and Kappa coefficient (81.3%), and showed almost perfect agreement after application of the exercises, while the results before the application of the exercises demonstrated accuracy without acceptable discrimination capacity (33.3%) and Kappa coefficient with a poor agreement (-6.67%). The Semioccluded Vocal Tract Exercises (SOVTE) with high strength straw presented with a sensitivity and Negative Predictive Value (NPV) of value 1 (one), and the over-articulation's specificity and Positive Predictive Value (PPV) also showed a value of 1 (one). CONCLUSIONS: The FTriangNB showed great accuracy in recognizing the effect of vocal exercises. Exploratory studies with larger samples using FTriangNB, as well as other Machine Learning classifiers should be further carried out for this purpose in the Voice Science to enable inferences.

Performance of Acoustic Measures for the Discrimination Among Healthy, Rough, Breathy, and Strained Voices Using the Feedforward Neural Network.

OBJECTIVE: To identify and evaluate the best set of acoustic measures to discriminate among healthy, rough, breathy, and strained voices. METHODS: This study used the vocal samples of the sustained /ε/ vowel from 251 patients with the vocal complaints, among which 51, 80, 63, and 57 patients exhibited healthy, rough, breathy, and strained voices, respectively. Twenty-two acoustic measures were extracted, and feature selection was applied to reduce the number of combinations of acoustic measures and obtain an optimal subset of measures according to the information gain attribute ranking algorithm. To classify signals as a function of predominant voice quality, a feedforward neural network was applied using a Levenberg-Marquardt supervised learning algorithm. RESULTS: The best results were obtained from 11 combinations, with each combination presenting six acoustic measures. Kappa indices ranged from 0.7527 to 0.7743, the overall hit rates are 81.67%-83.27%, and the hit rates of healthy, rough, breathy, and strained voices are 74.51%-84.31%, 78.75%-90.00%, 85.71%-98.41%, and 68.42%-82.46%, respectively. CONCLUSIONS: We obtained the best results from 11 combinations, with each combination exhibiting six acoustic measures for discriminating among healthy, rough, breathy, and strained voices. These sets exhibited good Kappa performance and a good overall hit rate. The hit rate varied between acceptable and good for healthy voices, acceptable and excellent for rough voices, good and excellent for breathy voices, and poor and good for strained voices.

A new combination rule for Spatial Decision Support Systems for epidemiology.

BACKGROUND: Decision making in the health area usually involves several factors, options and data. In addition, it should take into account technological, social and spatial aspects, among others. Decision making methodologies need to address this set of information , and there is a small group of them with focus on epidemiological purposes, in particular Spatial Decision Support Systems (SDSS). METHODS: Makes uses a Multiple Criteria Decision Making (MCDM) method as a combining rule of results from a set of SDSS, where each one of them analyzes specific aspects of a complex problem. Specifically, each geo-object of the geographic region is processed, according to its own spatial information, by an SDSS using spatial and non-spatial data, inferential statistics and spatial and spatio-temporal analysis, which are then grouped together by a fuzzy rule-based system that will produce a georeferenced map. This means that, each SDSS provides an initial evaluation for each variable of the problem. The results are combined by the weighted linear combination (WLC) as a criterion in a MCDM problem, producing a final decision map about the priority levels for fight against a disease. In fact, the WLC works as a combining rule for those initial evaluations in a weighted manner, more than a MCDM, i.e., it combines those initial evaluations in order to build the final decision map. RESULTS: An example of using this new approach with real epidemiological data of tuberculosis in a Brazilian municipality is provided. As a result, the new approach provides a final map with four priority levels: "non-priority", "non-priority tendency", "priority tendency" and "priority", for the fight against diseases. CONCLUSION: The new approach may help public managers in the planning and direction of health actions, in the reorganization of public services, especially with regard to their levels of priorities.

Methylation profile of the ADRB3 gene and its association with lipid profile and nutritional status in adults.

BACKGROUND: Defects in DNA methylation have been shown to be associated with metabolic diseases such as obesity, dyslipidemia, and hypercholesterolemia. To analyze the methylation profile of the ADRB3 gene and correlate it with lipid profile, lipid intake, and oxidative stress based on malondialdehyde (MDA) and total antioxidant capacity (TAC), homocysteine and folic acid levels, nutritional status, lifestyle, and socioeconomic variables in an adult population. A cross-sectional epidemiological study representative of the East and West regions of the municipality of João Pessoa, Paraíba state, Brazil, enrolled 265 adults of both genders. Demographic, lifestyle, and socioeconomic questionnaires and a 24-h recall questionnaire were applied by trained interviewers' home. Nutritional and biochemical evaluation (DNA methylation, lipid profile, MDA, TAC, homocysteine and folic acid levels) was performed. RESULTS: DNA hypermethylation of the ADRB3 gene, analyzed in leukocytes, was present in 50% of subjects and was associated with a higher risk of being overweight (OR 3.28; p = 0.008) or obese (OR 3.06; p = 0.017), a higher waist-hip ratio in males (OR 1.17; p = 0.000), greater intake of trans fats (OR 1.94; p = 0.032), higher LDL (OR 2.64; p = 0.003) and triglycerides (OR 1.81; p = 0.031), and higher folic acid levels (OR 1.85; p = 0.022). CONCLUSIONS: These results suggest that epigenetic changes in the ADRB3 gene locus may explain the development of obesity and non-communicable diseases associated with trans-fat intake, altered lipid profile, and elevated folic acid. Because of its persistence, DNA methylation may have an impact in adults, in association with the development of non-communicable diseases. This study is the first population-based study of the ADRB3 gene, and the data further support evaluation of ADRB3 DNA methylation as an effective biomarker.

Decrease of the DNA methylation levels of the ADRB3 gene in leukocytes is related with serum folate in eutrophic adults.

BACKGROUND: DNA methylation has been evidenced as a potential epigenetic mechanism related to various candidate genes to development of obesity. Therefore, the objective of this study was to evaluate the DNA methylation levels of the ADRB3 gene by body mass index (BMI) in a representative adult population, besides characterizing this population as to the lipid profile, oxidative stress and food intake. METHODS: This was a cross-sectional population-based study, involving 262 adults aged 20-59 years, of both genders, representative of the East and West regions of the municipality of João Pessoa, Paraíba state, Brazil, in that were evaluated lifestyle variables and performed nutritional, biochemical evaluation and DNA methylation levels of the ADRB3 gene using high resolution melting method. The relationship between the study variables was performed using analyses of variance and multiple regression models. All results were obtained using the software R, 3.3.2. RESULTS: From the stratification of categories BMI, was observed a difference in the average variables values of age, waist-to-height ratio, waist-to-hip ratio, waist circumference, triglycerides and intake of trans fat, which occurred more frequently between the categories "eutrophic" and "obesity". From the multiple regression analysis in the group of eutrophic adults, it was observed a negative relationship between methylation levels of the ADRB3 gene with serum levels of folic acid. However, no significant relation was observed among lipid profile, oxidative stress and food intake in individuals distributed in the three categories of BMI. CONCLUSIONS: A negative relationship was demonstrated between methylation levels of the ADRB3 gene in eutrophic adults individuals with serum levels of folic acid, as well as with the independent gender of BMI, however, was not observed relation with lipid profile, oxidative stress and variables of food intake. Regarding the absence of relationship with methylation levels of the ADRB3 gene in the categories of overweight, mild and moderate obesity, the answer probably lies in the insufficient amount of body fat to initiate inflammatory processes and oxidative stress with a direct impact on methylation levels, what is differently is found most of the times in exacerbated levels in severe obesity.

Influence of the C677T Polymorphism of the MTHFR Gene on Oxidative Stress in Women With Overweight or Obesity: Response to a Dietary Folate Intervention.

The C677T polymorphism of the methylenetetrahydrofolate reductase gene (MTHFR) is related to folate metabolism and can alter the levels of biochemical markers.Objective: Investigate the influence of the MTHFR C677T polymorphism on the effects of a dietary folate intervention on oxidative stress in women with overweight or obesity.Methods: Forty-eight adult women with overweight or obesity were subjected to a 24-hour dietary recall, anthropometric measurements, biochemical analysis, and genotyping of the MTHFR C677T polymorphism. They were allocated by convenience sampling to 2 groups, which received 300 g of folate-rich vegetables containing 191 µg/d (Group 1) (n = 24) or 95 µg/d (Group 2) (n = 24) of folate for 8 weeks.Results: The dietary intervention increased the serum folic acid levels in the 2 analyzed groups. The intervention with 191 µg/d of folate led to relevant results in terms of homocysteine levels (p = 0.0005) and total antioxidant capacity (p = 0.0261); the effect was larger among carriers of the TT genotype.Conclusions: The study demonstrated the beneficial effect of folate intake in terms of a TAC elevation for the CC and TT genotypes of the MTHFR C677T polymorphism, an increase in folic acid levels for all genotypes, and a reduction in the Hcy levels for the TT genotype in response to an intervention consisting of an intake of 191 µg/d of folate supplied by vegetables.

Effect of a diet containing folate and hazelnut oil capsule on the methylation level of the ADRB3 gene, lipid profile and oxidative stress in overweight or obese women.

BACKGROUND: Studies of genes that play an important role in the development of obesity are needed, especially studies focusing on genes that regulate food intake and affect nutrient metabolism. For example, the beta-3 adrenergic receptor (ADRB3) responds to noradrenaline and mediates lipolysis in adipocytes. METHODS: This was a controlled intervention study involving 40 overweight and obese adult women in which food intake, anthropometric measurements, biochemical analyses, and methylation levels of the ADRB3 gene were evaluated before and after intervention. The individuals were randomized into four groups: group 1 (G1) received 300 g of vegetables and legumes containing on average 191 µg/day of folate and 1 hazelnut oil capsule; group 2 (G2) received 300 g of vegetables and legumes containing on average 191 µg/day of folate and 1 placebo capsule; group 3 (G3) received 300 g of vegetables and legumes containing on average 90 µg/day of folate and 1 hazelnut oil capsule; and individuals in group 4 (G4) were only followed-up and maintained their regular dietary habits. Statistical analysis was performed using analysis of variance (ANOVA), Student's t test and simple regression, using STATA 13 software. RESULTS: In the total sample, after the intervention, the women classified as overweight and obese did not present weight loss, and there was a reduction in the methylation levels of the ADRB3 gene and malondialdehyde, as well as an increase in high-density lipoprotein cholesterol and total antioxidant capacity. CONCLUSIONS: The beneficial effect of the intake of a hazelnut capsule on the methylation levels of the ADRB3 gene was demonstrated for the first time. TRIAL REGISTRATION: ClinicalTrials.gov, NCT 02846025.

BMI, overweight status and obesity adjusted by various factors in all age groups in the population of a city in Northeastern Brazil.

OBJECTIVE: In Brazil, demographic, socioeconomic and epidemiological changes over time have led to a transition in nutritional standards, resulting in a gradual reduction of malnutrition and an increased prevalence of overweight and obese individuals, similar to the situation in developed countries in previous decades. This study assessed the body mass index (BMI) and the prevalence of an overweight status and obesity, adjusted for various factors, in a population in northeastern Brazil including all age groups. METHODS: This is a cross-sectional population-based epidemiological study using single sampling procedure composed of levels. Given the heterogeneity of the variable "income" and the relationship between income, prevalence of diseases and nutrition, a stratified sampling on blocks in the first level was used. In this, city districts were classified by income into 10 strata, according to information obtained from IBGE. A systematic sampling was applied on randomly selected blocks in order to choose the residences that would be part of the sample (second level), including 1165 participants from all age groups. RESULTS AND DISCUSSION: The prevalence of an overweight status or obesity was adjusted for demographic, socioeconomic and lifestyle variables. When the Chi-square test was applied, a relationship was observed between the prevalence of an overweight status or obesity and the age group, gender, educational level and income of the participants. Regarding lifestyle parameters, only smoking was associated with the prevalence of an overweight status or obesity, in both adults and in the total sample. The results for the following groups were significant (p < 0.05): the age group from 20 to 59 years, when the individual presented an educational level greater than or equal to high school; and the age group ≥ 60 years, when the individual was female. It is noteworthy that educational level and being female were significant in adjusting for the total population as major factors influencing an increased BMI, followed by the variables physical activity and family income. CONCLUSIONS: The adjusted results justify the adoption of intervention and prevention policies to combat these clinical conditions for the study population as a whole, particularly directed toward adults with higher education level as well as elderly females.

Association between waist-to-height ratio, isolated and combined morbidities and C-reactive protein in the elderly: a clinical-epidemiological study.

The aim of this study was to analyze the association between waist-to-height ratio (WHtR) and C-reactive protein (CRP) in the elderly (considering their most prevalent morbidities and lifestyles), to investigate the relationship between this anthropometric index and the presence of the most prevalent morbidities (isolated or combined), and to identify which morbidities (analyzed individually) would have greater associations with WHtR. This cross-sectional population-based epidemiological study of a stratified sampling comprised 170 elderly individuals between 60 and 90 years of age (both genders). Home visits were used to administer questionnaires and to perform anthropometric measurements and blood collection. The mean patient age was younger than 70 years, with women comprising the majority (69.41%) and with 90% of the patients presenting with inadequate WHtR. Hypertension was the most prevalent morbidity in this cohort (58.52%), and when analyzed in combination, hypertension plus obesity were the most frequently diagnosed morbidities (17.65%). Obesity, which was among the most prevalent comorbidities, was the only comorbidity combined with WHtR (p = 0.0019). Individuals with no morbidities had lower mean WHtR values compared with individuals with at least one morbidity (p = 0.0075). In the multiple linear regression model, it was identified that when individuals had one or more of the most prevalent comorbidities, the mean WHtR increased by 0.0415 (p = 0.0065). A correlation between WHtR and CRP (p = 0.0379) was also verified. Based on the relationships observed between WHtR (isolated or in combination, data unpublished) and CRP among the elderly, WHtR may represent a screening tool because it is a simple and effective anthropometric index.

Association between obesity and calcium:phosphorus ratio in the habitual diets of adults in a city of Northeastern Brazil: an epidemiological study.

BACKGROUND: Low calcium:phosphorus ratios (Ca:P ratio) in habitual diet have been observed worldwide, and it has been shown to be harmful to the bone health of the population. However, no study associating this ratio with obesity was found. Thus, considering that the intake of calcium and phosphorus will generate a ratio between them, which may be associated with obesity, this research seeks at evaluating the relation between obesity and the Ca:P ratio in the habitual diet of adults. METHODS: Cross-sectional population-based epidemiological study with stratified and systematic sampling. The sample was composed of 506 adults, aged between 18 and 60 years, of both genders. Information on socioeconomic and demographic conditions was obtained through questionnaires completed during home visits, where anthropometric and dietary evaluations were also conducted. RESULTS: In the habitual diet consumed by the study subjects, a Ca:P ratio above the median of 0.57 reduced the risk of central obesity based on waist-to-height ratio (WHtR) (OR: 0.61; 95% CI: 0.41 - 0.92). Habitual dietary intake of calcium (OR: 0.65; 95% CI: 0.43 - 0.97) and dairy products (OR: 0.56; 95% CI: 0.37 - 0.84) above the median value (485.4 mg and 0.9 servings, respectively) was found to be a protective factor related to central obesity based on WHtR. CONCLUSIONS: Values above the median for the Ca:P ratio found in the habitual diet were negatively associated with central obesity based on WHtR. In addition, calcium and dairy consumption were negatively associated with central obesity based on WHtR. Therefore, higher Ca:P ratios contributed to a lower prevalence of central obesity.

Online training evaluation in VR simulators using Gaussian Mixture Models.

A new approach to evaluate training in simulators based on virtual reality is proposed. This approach uses Gaussian Mixture Models (GMM) for modeling and classification of the simulation in pre-defined classes of training.