Acoustic, Myoelectric, and Aerodynamic Parameters of Euphonic and Dysphonic Voices: A Systematic Review of Clinical Studies

: Background: At present, there is no clinical consensus on the concept of normal and dysphonic voices. For many years, the establishment of a consensus on the terminology related to normal and pathological voices has been studied, in order to facilitate the communication between professionals in the ﬁeld of the voice. Aim: systematically review the literature to compare and learn more precisely the measurable and objective characteristics of the acoustic, aerodynamic and surface electromyographic parameters of the normal and dysphonic voices. Methods: The PRISMA 2020 methodology was used as a review protocol together with the PICO procedure to answer the research question through six databases. Results: In total, 467 articles were found. After duplicate records were removed from the selection, the inclusion and exclusion criteria were applied and 19 articles were eligible. A qualitative synthesis of the included studies is presented in terms of their methodology and results. Conclusions: Studying the acoustic, aerodynamic, and electromyographic parameters with more precision, in both normal and dysphonic voices, will allow health professionals working in the ﬁeld of voice (speech therapy, otorhinolaryngology, phoniatrics, etc.) to establish a diagnostic and detailed consensus of the vocal pathology, enhancing the communication and generalization of results worldwide. voice diagnostic


Introduction
To enable rehabilitation that adapts to the needs of the patient, a functional evaluation of the voice is necessary, which requires an otorhinolaryngological medical diagnosis and a speech therapy evaluation [1].
Through the functional evaluation of the voice, the patients' vocal behavior should be observed by analyzing aspects related to anatomy and physiology and the negative vocal habits they present, thereby determining the severity of the disorder and establishing a prognosis and/or a diagnosis [2].
With regards to speech therapy, a detailed anamnesis of the vocal pathology and a postural examination should be carried out. In addition, a subjective assessment of the voice (i.e., auditory perceptual analysis) should be performed, preferably by a speech therapist trained in alterations of the voice; as a complementary evaluation to an objective assessment of the voice through acoustic, aerodynamic, and surface electromyographic analyses, among others. For this reason, being able to perform an exhaustive evaluation of a majority of the parameters and structures that are related to vocal emission, helps and facilitates the professionals who work with it in a clinical manner (speech therapists, doctors, otolaryngologists . . . ) to provide an adequate diagnosis of the vocal pathology.
When describing an individual's voice, there are terminological differences. For professionals working to develop, describe and/or rehabilitate human voices, it is necessary to be able to recognize and agree on terms and vocal parameters [3].
At present, there is no clinical consensus on the concept of normal voice and dysphonic voice. For many years, the establishment of a consensus on the terminology related to normal and pathological voices has been studied, in order to facilitate the communication between professionals in the field of the voice. Most recently, a comprehensive review of the worldwide healthy adult voice baseline parameters has been performed, in an attempt to address this lack of consensus [2].
When referring to the euphonic or "normal voice", Jackson-Menaldi [4] takes into account the presence of a pleasant vocal timbre, a volume which is appropriate to the vocal needs, and a tone which is appropriate to the sex and age of the patient.
In contrast, according to Lee [5], dysphonia appears when the quality of the voice, the tone, the volume, the resonance and/or the duration differ or are inappropriate for the age, gender, cultural background, or geographic location of a person.
During the last 40 years, the clinical application of acoustic analysis has been developing, but the acoustic differentiation between healthy and pathological voices is still inaccurate [6,7].
Accordingly, it appears relevant to describe the objective parameters of the normal and dysphonic voices. Taking this into consideration, the objective of the present systematic review was to evaluate the conditions for a normal and a dysphonic voice in adult patients from the measurement of acoustic, aerodynamic, and myoelectric characteristics of their voices.

Materials and Methods
The present systematic review was carried out following the PRISMA 2020 (Preferred Reporting Items for Systematic reviews and Meta-Analysis) statement guidelines [8].

Eligibility Criteria
The eligibility criteria were established by following the PICO model [9], as follows. population/problem (P): adult patients with or without vocal pathology (normal/dysphonic voice); intervention (I): procedures for the determination of the healthy/pathological state of the voice, comprising the evaluation of acoustic, myoelectric, and/or aerodynamic objective measures; comparison/control (C): comparison of such objective measures in the previously described population; outcome (O): measurable outcomes from voice diagnostic tests.
Clinical studies that addressed the aforementioned PICO parameters were eligible for the present systematic review. The date or language of the publications were not considered as inclusion/exclusion criteria.
The exclusion criteria were the following: descriptive, evaluation, or treatment studies of organic vocal pathology (except nodules and polyps); studies on therapy or rehabilitation in benign pathology of the vocal cords; articles on voice disorders not related to vocal overexertion (puberphonia, presbyphonia, head and neck cancer, laryngectomy, etc.); works on the voice in neurodegenerative diseases or feminization of the voice; voice rehabilitation techniques and/or methods; studies on evaluation and/or rehabilitation of the singing voice; childhood dysphonia; gastroesophageal reflux; phonosurgery; and systematic reviews.
Voice disorders derived from vocal muscle tension and organic pathology such as nodules and polyps were included, which are usually the most frequent in the vocal rehabilitation of the speech therapy clinic.

Search Strategy
The database search, study screening process, data extraction, and risk of bias analysis were performed by two independent researchers. In case of doubt or disagreement, a third investigator was consulted.

Study Selection Process
Study records resulting from the search process were exported to Mendeley Desktop 1.19.4 reference manager software (Elsevier, AMS, The Netherlands) to manually check for duplicates. After discarding duplicates, an initial screening of the reference titles and abstracts was performed, following the aforementioned inclusion and exclusion criteria. The studies that did not fulfill such criteria upon the screening of their title and abstract were discarded. Studies that did meet the criteria where subsequently assessed for eligibility by a full-text screening.

Data Extraction
The process of data extraction was divided into three categories: extraction of variables for study characteristics, methodology, and results. Variables for study characteristics comprised the authors and year of publication. Variables for study methodology were the sample size and its characteristics (vocal healthy/pathological state, gender, age), study type, technique/s for the registry and evaluation of vocal parameters, and statistical analysis. Lastly, results variables included acoustic measures, aerodynamic data, myoelectric activity registries, and comparisons performed.

Quality Assessment
For the quality assessment, "The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement" [10] was used for the included observational stud-ies. To evaluate the quality of the case series included in the present review, the "Tool for evaluating the methodological quality of case reports and case series" from "Methodological quality and synthesis of case series and case reports" [11] was followed. The full texts of the studies were screened and evaluated for each of the parameters included in the quality assessment tools.

Search Results and Study Selection
The study selection process is shown in Figure 2. The database searches identified 467 records: seven from Scopus, 66 from Cochrane Library, 178 from Medline, 188 from Web of Science, and 28 from Embase. The search performed in SciELO produced no results. Duplicates were discarded using the reference manager software, resulting in 384 records. From the resulting records, 367 were excluded upon screening of the title and abstract. Two additional eligible studies were found upon screening the references of the resulting studies. The resulting 19 papers were evaluated by full-text screening, and all of them were eligible for qualitative synthesis. Table 1 shows the summary of the methodology used in the 19 studies included in the review. In most of them, a sample made up of two groups was observed: an experimental group made up of people with vocal pathology and a control group made up of people without vocal pathology [12][13][14][15][16][17][18][19][20][21][22][23][24][25]. Another three were made up of a group of volunteers with vocal pathology [26][27][28] and two more were composed of volunteers without vocal pathology [29,30]. In all of the observational studies, the groups were established in a non-randomized controlled manner. The sample size of the included studies ranged from 10 [26] to 387 [15] participants, except for one study [20] which included a sample of 1410 participants due to the use of a previous database.

Study Methodology
Regarding the parameters analyzed in the different studies (Table 2), three categories can be established: acoustic measurements, aerodynamic parameters, and parameters of myoelectric activity.
Within the section on acoustic measurements, a small section has been included referring to the auditory perceptual analysis assessment scales and instruments used in the different articles.
Regarding the instruments and materials used in the selected studies, a series of microphones, hardware, and software, both for the vocal register and for the aerodynamic register were used. This is summarized in Table 3.

Study Results
A summary of the results of the included studies in terms of acoustic and aerodynamic parameters is presented in Tables 4 and 5, respectively.                With regards to the results obtained related to the acoustic parameters assessed in the selected studies, (Table 4), only those studies in which acoustic data appear (13 articles) have been included. In all of the studies, there are data on the average fundamental frequency (in Hz). In some, the maximum and minimum frequencies are also included. In two works [15,27] only disturbance parameters are studied: PPQ, APQ, and shimmer; and HNR and APQ, respectively.
Two studies [16,17] assessed the variation coefficient of the fundamental frequency (CV F 0 ), providing its values in % and not in Hz. The intensity is only studied in 7 of the 13 studies that analyze acoustic parameters.
The same happens with the acoustic disturbance parameters, which are studied in 9 of the 13 cases and the results are provided as a percentage, except in one of them [23] which reports shimmer results in dB.
Regarding the aerodynamic parameters assessed by the selected studies (Table 5), a great variability is observed. The most studied parameters were the maximum phonation time (MPT) in six of the nine articles and the subglottic pressure (SGP) in five of the nine studies.
In the only work included in the review which studied the relationship between acoustic and electromyographic parameters [19], data obtained in the different phonatory tasks (vowel /ε/ and counting at normal and maximum intensity) are reflected together with the different muscle groups studied: suprahyoid and infrahyoid muscles.
The aforementioned tables do not include data obtained in the subjective and quasi-objective assessment of the voice (VPSS, GRBAS scale, CAPE-V, V-RQOL, etc.), since the objective of this study was to assess the acoustic, aerodynamic, and electromyographic parameters.

Quality Assessment
The quality assessment of the observational studies included in the present review is presented in Tables 6 and 7. Overall, studies presented a clear structure, and the key data regarding the methodology and results were reported. Titles were appropriate, but often failed to indicate the study design. However, in all cases, the abstract presented an adequate summary of the methodology and main findings. Introductions presented a clear background and rationale for the investigation and indicated the specific aims of the studies. Generally, data regarding the methodology used was enough to allow for replication (i.e., inclusion criteria and variables were clearly defined, and statistical analyses were described), but none of the studies indicated how the sample size was calculated. With reference to the results, outcome data and the description of the main results were generally structured and clear, except data regarding the participants (i.e., characteristics, missing data, and drop-out rates, when applicable). Lastly, studies generally presented an adequate discussion: by describing, interpreting, and generalizing the main results, whilst addressing the possible limitations of the study design. Two other observational studies were included in the present review [15,20], but were not assessed using the STROBE statement, since they were based on voice recordings and not human subjects.
The last study included in the review [26] was a case series and was independently assessed for quality and risk of bias parameters using a specific tool for case series [11], as previously mentioned. Overall, the exposures and outcomes were adequately ascertained, and the cases were described with sufficient detail to allow for replication. However, the selection method for the participants was unclear.

Discussion
The multifactorial analysis of the voice, both objectively and subjectively, is of great interest for the study and evaluation of vocal disorders from a global perspective.
In studies where the sample was divided into two groups, only 10 specified the type of vocal pathology [12,13,[16][17][18]20,[24][25][26]28]. In addition, none of the studies specified how the calculation of the sample size or the sampling procedure was performed. Both the distribution by groups (gender and number) and the age ranges observed in the included studies were quite heterogeneous, except in two of the included works [24,25], where the same number of participants appeared in both groups. This heterogeneity may hinder the interpretation of the studies' results since both age and gender could act as factors influencing voice quality and its characteristics. Other factors such as ethno-cultural features or traits could also present a certain specificity in terms of the characteristics of the voice, as suggested by Lee et al. [31].
In all of the included works, the phonatory tasks performed to evaluate the vocal register were indicated, except in one study [27]. The most commonly used was the emission of a sustained vowel /a/ at a comfortable frequency and intensity, although in some cases the minimum and maximum intensity possible was also requested. Other types of phonatory tasks that were carried out were: counting numbers (from 20 to 30) [19], repeating words [17], phrases [13,14,17,29], spontaneous conversation [17] and reading aloud [15]. In two works [19,28], the sustained vowel /ε/ was requested instead of /a/.
Regarding the results of the included studies in terms of acoustic parameters, the vocal disturbance parameters (jitter, shimmer, HNR, etc.) were higher and statistically significant differences were observed in people with pathological voices compared to people with healthy or non-dysphonic voices [12,16,23,24], the shimmer being one of the significant factors in the discrimination between subjects with and without vocal pathology [17].
On the other hand, the aerodynamic measurements reflected, to a higher degree, the presence of vocal pathology than the acoustic measurements [26]. Subglottic pressure has an important role in normal voice production and direct importance in the evaluation of laryngeal function [32]. Statistically significant differences were observed in subglottic pressure between the group with vocal pathology compared to the group without vocal pathology [22] and between genders [30]. Men with vocal pathology were also found to have higher subglottic pressure and glottic resistance compared to women with vocal pathology and shorter maximum phonation time [18].
In addition, one study indicated that the maximum phonation time and the glottic efficiency index are significant factors in the discrimination between subjects with and without vocal pathology [17].
According to another study [13], the use of aerodynamic measurements was a strong predictor to differentiate between groups with vocal pathology and without pathology.
They also indicate that these should be used as an adjunct to therapy and not just as a diagnostic tool.
Analyzing the selected studies, a great variability was observed in almost all of the aspects observed, i.e., the sample size and characteristics, etc. However, what is most striking is the number of different protocols, hardware, and software used among the studies, hindering the generalization and establishment of acoustic, aerodynamic, and electromyographic normative parameters, as well as the comparison between the results obtained by the different studies and, in turn, defining what is understood by a normal voice and a pathological voice.
Nevertheless, an effort has been made to propose standardized protocols for the instrumental assessment of vocal function, such as the one presented by Patel et al. [33], developed by an American Speech-Language-Hearing Association expert panel. In this proposal, however the lack of evidence and consensus in the field is acknowledged, as described in the present manuscript.

Conclusions
In accordance with the objective of the present study, which aimed to evaluate the conditions of the normal and dysphonic voices in adults from acoustic, aerodynamic, and surface electromyographic parameters, it can be highlighted that the results from the included studies show differences between both voices.
The most studied acoustic parameters were frequency and intensity, which tend to decrease with the presence of dysphonia and disturbance parameters (jitter, shimmer, and HNR) that tend to increase in vocal pathology. Regarding the aerodynamic measurements, maximum phonation time (MPT) and subglottic pressure (SGP) were the most analyzed, with MPT being generally higher in people without vocal pathology and SGP higher in people with dysphonia. The existence of differences in surface electromyographic parameters cannot be determined due to the small number of articles that discuss this topic.
The small number of articles, the variability of the populations analyzed, the variety of instruments used, and aspects examined related to the voice, hinder the establishment of normative criteria to objectively define what is a normal voice and a dysphonic voice.
Studying the acoustic, aerodynamic, and electromyographic parameters with more precision, both of the normal and dysphonic voice, will allow health professionals working in the field of voice (speech therapy, otorhinolaryngology, phoniatrics, etc.) to establish a diagnostic and detailed consensus of the vocal pathology in question, favoring the communication and generalization of the results worldwide.
Therefore, more studies are needed to deepen the development and implementation of study protocols. This would favor the homogeneity of the records in order to be able to generate the necessary data to establish thresholds of normality and pathology, as well as allow the comparison of results between different study groups.