Objective Acoustic Phonetic Speech Analysis Biology Essay

Speech damage frequently occurs in patients after intervention for caput and cervix malignant neoplastic disease. New intervention modes such as surgical Reconstruction or ( chemo ) radiation techniques aim at saving anatomical constructions that are correlated with address and swallowing. In randomised tests look intoing efficaciousness of assorted intervention modes or address rehabilitation, nonsubjective address analysis techniques may add to better address result appraisal. The end of the present survey is to look into the function of nonsubjective acoustic -phonetic analyses in a multidimensional address appraisal protocol. Speech recordings of 51 patients ( 6 months after rehabilitative surgery and postoperative radiation therapy for unwritten or oropharyngeal malignant neoplastic disease ) and of 18 control talkers were subjectively evaluated sing intelligibility, rhinal resonance, articulation, and patient-reported address result ( speech subscale of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Head and Neck 35 faculty ) . Acoustic-phonetic analyses were performed to cipher formant values of the vowels /a, I, u/ , vowel infinite, air force per unit area release of /k/ and spectral incline of /x/ . Intelligibility, articulation, and rhinal resonance were best predicted by vowel infinite and /k/ . Within patients, /k/ and /x/ differentiated tumour site and phase. Assorted nonsubjective address parametric quantities were related to speech jobs as reported by patients. Objective acoustic-phonetic analysis of address of patients is executable and contributes to further development of a address appraisal protocol.

Introduction

Tumors in the unwritten pit and oropharynx may ensue in harm of assorted anatomical constructions by tumour extension and intervention. Patients frequently report a reduced usage of lingua and perioral musculuss and address variety meats, such as the lips, lingua and veil, which often causes address trouble and other jobs such as those related to societal activities. These jobs

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

may finally hold a negative impact on health-related quality of life.1 Health-related quality of life significantly deteriorates during the first 6 months after completion of intervention, and may better by 12 months after intervention. Functionality of the caput and cervix country often remains below pretreatment level.2 Speech quality after intervention appears to be extremely dependent on tumour size and site.3-9 Patients who underwent intervention of larger tumor experienced more trouble with address than those with smaller tumor. Speech result after intervention for an unwritten tumor frequently consequences in articulation troubles due to weave loss, and construction change of assorted address variety meats, while jobs with speech production of patients treated for oropharyngeal malignant neoplastic disease frequently include rhinal resonance jobs due to velopharyngeal insufficiency. In the past decennaries, surgical possibilities of replacing damaged tissues in the unwritten pit and oropharynx by different flaps have increased aiming to forestall address and get downing damage. The preferable method of Reconstruction of larger defects in the unwritten pit or oropharynx is by free flaps. Free fasciocutaneous flaps are thin and fictile and are suited for Reconstruction of dynamic constructions, such as the lingua and pharynx.3-6 More late, organ saving protocols such as chemoradiation are introduced besides taking at bar of functional damage. However, a recent literature reappraisal reveals that both intervention modes, rehabilitative surgery and organ saving, still frequently result in address and get downing impairment.10 New radiation bringing techniques taking at saving anatomical constructions that are correlated with address and swallowing may lend to forestall long-run radiation-induced functional damage as may speech rehabilitation. Besides, new address rehabilitation attacks such as logopedic exercisings in an early phase before or during radiation therapy may better functional result. However, prospective randomized tests are needed to supply evidence-based effectivity of these attacks. Objective address analysis techniques may add to better address rating protocols and enable equal address result appraisal in clinical tests. Speech quality is most frequently assessed via subjective rating by hearers. Consequences obtained from subjective appraisals reveal correlativities between tumour phase, intelligibility and articulation: patients with a smaller tumor ( T2 ) have better intelligibility and articulation than patients with larger tumors ( T3-T4 ) . Nasal resonance and articulation of patients are significantly worse than in healthy individuals.9 Nasal resonance in patients treated for tumors in the oropharynx appears to be worse than in patients with unwritten tumors. This

difference is due to the oropharyngeal country that is involved in the divider between the unwritten and rhinal pit. In instance of neglecting velar closing, air flights through the olfactory organ, which consequences in hypernasal features of speech.11 Objective measurings of address quality are less frequently performed. Acoustic-phonetic analysis of the speech signal appeared to distinguish between healthy talkers and glossectomy patients.12 Acoustic-phonetic analyses besides revealed that patients who underwent partial resection of the lingua have deviant formant values for vowels, particularly for /i/.12, 13 A survey utilizing a nasometer revealed that address of patients after Reconstruction with big flaps had worse rhinal resonance scores.5 They besides reported that patients with resections of more than half of the soft roof of the mouth had more rhinal resonance than patients with smaller resections of the soft roof of the mouth. The purpose of this survey is to obtain more penetration in phonetic-acoustic address features of patients after microvascular rehabilitative surgery for unwritten or oropharyngeal malignant neoplastic disease sing formant values of the vowels /a, I, u/ , and the velar consonants /k/ and /x/ . The 2nd purpose is to look into the cogency of nonsubjective phonetic-acoustic address parametric quantities. The consequences contribute to further development of a multidimensional address appraisal protocol that can be used in future prospective tests on efficaciousness of assorted intervention modes and rehabilitation for caput and cervix malignant neoplastic disease.

Patients and Methods

Patients

Patients underwent intervention for advanced unwritten or oropharyngeal squamous cell carcinoma with microvascular soft tissue transportation for the Reconstruction of surgical defects. Surgery consisted of composite resections including deletion of the primary tumor with en axis ipsilateral or bilateral cervix dissection. In instance of oropharyngeal carcinomas a paramedian mandibular swing attack was used. Defects were reconstructed by a microvascular

fasciocutaneous flap ; no flap failures were observed. Patients received postoperative radiation therapy in instance of advanced ( T3-T4 ) tumor, positive or close surgical borders, multiple lymph node metastases and excess nodal spread. The primary site received a dosage of 56-66 Gy in entire ( 2 Gy per fraction, 5 times per hebdomad ) , depending on surgical borders. The nodal countries received a sum of 46-66 Gy ( 2 Gy per fraction, 5 times a hebdomad ) . Exclusion standards were inability to take part in functional trials, trouble pass oning in Dutch and age above 75 old ages. Fifty-one patients between 23 and 73 old ages ( mean: 53.8 old ages, SD: 8.7 old ages ) were included in the survey after obtaining written informed consent, every bit good as 18 gender- and age-matched controls ( table 1 ) .

Speech Assessment

Patients ( 6 months after intervention ) and controls read aloud a text with an approximative length of 60 s. The distance between lips and mike ( Sennheiser MKE 212 to 213 ) was 30 centimeter. Speech recordings were conducted in a soundproof cabin. For each talker the entering degree was adjusted to optimise signal-to-noise ratio. The recorded address was digitalized with Cool Edit PRO 1.2 ( Adobe Systems Incorporated, San Jose, Calif. , USA ) , with a 22- kilohertz sample frequence and 16-bit declaration.

Subjective Speech Evaluation

Perceptual rating of address quality comprised evaluations on intelligibility, articulation and rhinal resonance by two address diagnosticians. To enable subjective address rating, a computing machine plan was developed to execute blinded randomized hearing experiments and to automatically hit intelligibility, nasality, and articulation. Intelligibility was scored utilizing a 10-point graduated table, where 1 represents the worst mark and 10 represents the best mark and 6 is merely sufficient. Articulation and rhinal resonance were judged utilizing a 4-point graduated table, runing from normal to progressively deviant address quality. Interrater understanding for subjective appraisal of intelligibility ranged from 40 to 90 % . Intrarater understanding for repeated address fragments of articulation and rhinal resonance was high, with 100 % equal tonss between the evaluations. Patient-reported address result was assessed by the address subscale ( including 3 points ) of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Head and Neck 35 faculty. The tonss were linearly transformed to a graduated table of 0-100, with a higher mark bespeaking a higher degree of address problems.14

Acoustic-Phonetic Analysiss

In the present survey, the vowels /a, I, u/ ( the cardinal vowels in Dutch ) and velar consonants were used as study stuff. Vowels are -compared to consonants- comparatively easy to place in the address signal, and easier to analyse acoustically.

Table 1. Overview of gender, tumour site and phase of 51 patients included in the survey

n ( % )

Gender

Male 28 ( 55 )

Female 23 ( 45 )

Tumour site

Oral pit 21 ( 41 )

Oropharynx 30 ( 59 )

T-classification

2 26 ( 51 )

3-4 25 ( 49 )

Vowel formant analyses proved to be valid steps of address quality in patients with aberrant address arising from unwritten malignant neoplastic disease or other beginnings in earlier studies.15, 16 Vowel individuality ( or its spectral colour ) is characterized by acoustic correlatives and is chiefly determined by its formants. Broadly talking, the first formant frequence ( F1 ) is associated with ‘height ‘ , that is, the grade of gap of the vocal piece of land, whereas the 2nd formant frequence ( F2 ) is associated with the anterior-posterior lingua position.17 Ploting the vowels /a, I, u/ onto a graphical F1-F2 representation shows the vowel infinite ( more specifically the vowel trigon ) . The vertices of the vowel trigon represent the most drawn-out places. The country of the vowel trigon is a step for the sum of decrease in the vowel system and can ( officially ) be measured in footings of Hz 2 ( see besides figure 1 ) .18

Figure 1. Vowel infinite of male ( bluish ) and female ( tap ) patients ( fat lines ) and of controls ( thin lines ) .

In add-on to vowels, the velar consonants /k/ and /x/ were acoustic-phonetically analyzed, because earlier research revealed that patients with an unwritten or oropharyngeal tumors frequently have troubles with the production of velar address sounds. Speech raters frequently mistook /k/ for /x/.9, 11 For /k/ the continuance of air force per unit area release ( the alleged stop consonant ) as a per centum of the entire continuance ( short silent period of force per unit area edifice + the force per unit area release ) was measured and used as outcome step. For /x/ the spectral incline was used as outcome step. For each selected address sound ( /a, one, u, K, x/ ) , two acoustic realisations were segmented from running address and were acoustic-phonetically analyzed utilizing the address processing package

Praat version 4.0.28.19 Since the acoustic realisation of certain speech sounds may depend on its context, we took different phonological contexts around the mark address sounds into history, in order to better generalisation. A spectrograph functioned as a ocular representation of the address signal, which facilitated acknowledgment of phonemes in the address signal and facilitated precise extraction of phonemes from running address. Spectral and acoustic address analyses were automatically performed utilizing scripts19

Statistical Analysis

Cogency of nonsubjective address analyses was tested by agencies of univariate Pearson correlativity coefficients between the subjective address ratings of intelligibility, articulation and rhinal resonance and nonsubjective parametric quantities ( formants of the vowels /a, I, u/ , size of the vowel infinite, spectral incline of /x/ and continuance of force per unit area release of /k/ ) . To obtain insight into the function of nonsubjective parametric quantities in foretelling subjective address rating, multivariate arrested development analyses were performed. For intelligibility and self-assessments by patients, a additive arrested development was used, while for articulation and rhinal resonance, logistic arrested development was performed on a binary graduated table [ normal ( score 0 ) vs. pervert ( scores 1-3 ) ] . Mann-Whitney trials were performed alternatively of T trials due to skewed informations and were used to find the cogency of the nonsubjective address parametric quantities sing known group differences: patients versus controls, smaller ( T2 ) versus larger ( T3-T4 ) tumor, and tumour location ( unwritten vs. oropharyngeal ) .

Consequences

The two formants of two realizations of each vowel were averaged because review of formant values of the two realizations of one vowel revealed that there were no important differences. For the velar address sounds /k/ and /x/ , nevertheless, larger differences were found which made utilizing the mean inappropriate. Therefore, /k/1 and /k/2 and /x/1 and /x/2 are analyzed individually and described in the consequences.

Objective versus Subjective Speech Assessment

Univariate correlativities between subjective ( self- ) ratings and nonsubjective parametric quantities reveal that evaluations on intelligibility and articulation are significantly related to nonsubjective analyses of /k/ , the 2nd formant of /i/ , and formant infinite ( table 2 ) .

Table 2. Pearson correlativities between nonsubjective address parametric quantities and subjective parametric quantities Intelligibility, Articulation and Nasal Resonance ( * P & lt ; .05 ) .

Intelligibility

Articulation

Nasal resonance

EORTC H & A ; N35 Speech graduated table

R

R

R

R

/x/2

.12

.13

.33*

-.02

/k/1

.50*

.40*

.25*

-.27

/k/2

.36*

.25*

.39*

-.13

/i/F1

-.23

-.19

-.42*

.02

/i/F2

.35*

.36*

.13

-.24

/u/F1

-.11

-.11

-.37*

-.12

size I” ( HzA? )

.39*

.42*

.15

-.20

To obtain penetration into which nonsubjective parametric quantities predict subjective ( self- ) appraisals, multiple arrested development analyses were performed ( tabular arraies 3-6 ) . The consequences reveal that /k/ , F1 of /i/ , and the size of the vowel trigon predicted best subjective ( self- ) ratings. These consequences reveal equal cogency of nonsubjective address analyses. Particularly /k/ , /i/ , /x/ and the size of the vowel trigon contribute to a anticipation of subjective rating by nonsubjective address parametric quantities.

Table 3. Prediction of intelligibility by acoustic-phonetic parametric quantities. ( * P & lt ; .05 ) . RA?=45 % .

Intelligibility

B

T

/k/1

.038

3.42*

size I” ( HzA? )

7.07

4.11*

/i/F1

-.13

-2.60*

Table 4. Prediction of articulation by acoustic-phonetic parametric quantities.

( * P & lt ; .05 ) . RA?= 74 % . )

Articulation

B

Wald

/k/1

.11

8.53*

/a/ F1

-.20

6.52*

/a/ F2

.01

4.59*

/i/F1

-.06

8.31*

size I” ( HzA? )

42.87

9.23*

Table 5. Prediction of rhinal Resonance by acoustic-phonetic parametric quantities.

( * P & lt ; .05 ) . RA?= 52 % .

Nasal Resonance

B

Wald

/x/2

.20

9.36*

/k/2

.05

7.32*

/i/F1

-.03

7.67*

Table 6. subjective and nonsubjective address parametric quantities of address quality that are related to speech jobs in day-to-day life as reported by patients. ( * P & lt ; .05 ) . RA?= 45,4 % .

EORTC H & A ; N-35 Speech Scale

B

T

/x/1

1.11

2.29*

/i/F2

-.04

-2.12*

Known Group Differences

To obtain insight into the prognostic cogency of nonsubjective address analyses, Mann-Whitney trials were performed sing known group differences: patients versus controls, and within the group of patients sing tumour categorization and tumour site ( table 7 ) . Significant differences between patients and controls in acoustic-phonetic parametric quantities revealed that patients

hold a shorter force per unit area release for /k/ than controls. Patients have a higher F1 of /i/ , but a lower F2 of /i/ than controls. The size of the vowel trigon is significantly smaller for patients than for controls. Acoustic-phonetic analysis besides differentiated sing tumour phase. Patients with smaller tumor had a longer force per unit area release compared to patients with a larger tumor. Sing tumour site, /x/ distinguished between tumour location: patients with an oropharyngeal tumor had a steeper spectral incline than patients with an unwritten tumor.

Table 7. Significant differences between nonsubjective acoustic-phonetic variables measured on vowels ( formant values in Hz, size vowel trigon in Hz2 ) and consonants ( continuance of air force per unit area release ( the alleged stop consonant ) as a per centum of the entire continuance ( short silent period of force per unit area edifice + the force per unit area release ) of /k/ ; spectral incline for /x/ ) between pathological and control talkers, and sing tumour site and tumour categorization, as obtained with a Mann-Whitney trial.

Pathological vs. control talkers

Patients

Controls

/k/1

Z=-2.77

p=.006

28,5 %

sd=16

43,5 %

sd= 18

/k/2

Z=-4.15

P & lt ; .001

26,4 %

sd=19

49,8 %

sd=17

F1 /i/

Z=-2.36

p=.018

334 Hz

sd=54

296 Hz

sd=49

F2 /i/

Z=-2.42

p=.016

2105 Hz

sd=363

2325 Hz

sd=248

size I”

Z=-2.42

p=.015

.143 HzA?

sd=.12

.213 HzA?

sd=.11

Oral tumor vs. oropharyngeal tumor

Oral tumor

Oropharyngeal tumor

/x/

Z=-2.24

p=.025

-13

sd=6

-17

sd=6

T2 tumor vs. T3-4 tumor

T2

T3-4

/k/

Z=-2.09

p=.037

33 %

sd=17

23 %

sd=14

Discussion

This survey presents an stock list of address public presentation 6 months after intervention in a chiseled caput and cervix malignant neoplastic disease patient group after rehabilitative surgery and radiation therapy for advanced unwritten or oropharyngeal

malignant neoplastic disease. Speech quality was determined with nonsubjective acoustic-phonetic analyses and normally used subjective ( self- ) ratings.

The first purpose of the present survey was to look into which nonsubjective parametric quantities contribute to the anticipation of subjective ( self- ) ratings of address. Particularly acoustic-phonetic parametric quantities of /k/ , /x/ , /i/ , and the size

of the vowel trigon predicted best subjective appraisal of overall intelligibility, articulation, rhinal resonance and self-evaluation of address. The consequence sing /k/ is besides reported9, where hearers frequently judged /k/ as /x/ . Production of velar consonants such as /k/ and /x/ require a posterior move of the lingua towards the oropharyngeal part and an equal motility of the veil. Larger lingua motility corresponds with better intelligibility of consonants, including /k/.20 No old surveies report on the address sound /x/ , which may be due to the absence of /x/ in other modern Western linguistic communications except for Dutch and a few idioms like Scottish.

The size of the vowel trigon was besides found to be a forecaster of subjective address ratings. The smaller size of the vowel trigon in patients was caused by the higher F1 and lower F2 of the vowel /i/ . These consequences are in understanding with earlier research, where it was shown that a smaller size of the vowel triangle – that was besides caused by aberrant values of F1 and F2 of /i/ – was related to worse intelligibility in glossectomy patients.12 In the

present survey, the vowel /i/ itself besides proved to foretell subjective ratings: patients had a higher F1 and a lower F2. These consequences are in understanding with the consequences of earlier research on pathological speech13, 18 ( both concerned maxillectomy patients ) , but are non in understanding with consequences on research refering partial glossectomy, where it was found that merely gender and complication after surgery were of influence on altered F1

values.21

The 2nd purpose of this survey was to look into differences sing acoustic-phonetic address features between patients and controls and within the group of patients sing tumour site and tumour categorization. Between patients and controls, force per unit area release of /k/ , F1 and F2 of /i/ , and the size of the vowel trigon differentiated best. Trouble with production of /k/ originates from velar map troubles. The reduced size of the vowel trigon of patients was chiefly caused by aberrant formant values of /i/ and is in conformity with earlier studies.12, 13, 18 Inadequate motion of the lingua sing tallness ( F1 ) and anterior-posterior motion ( F2 ) may ensue in deformed vowels. Acoustic-phonetic analysis besides revealed differences between patients sing tumour phase ( /k/ ) and tumour site ( /x/ ) : patients with smaller tumor had a longer force per unit area release of /k/ compared to patients with a larger tumor. Sing tumour site, patients with an oropharyngeal tumor had a steeper spectral incline in /x/ than patients with an unwritten tumor. Due to tumour growing and intervention in the oropharyngeal country, patients with oropharyngeal malignant neoplastic disease are likely to see more trouble with the production of velar address sounds. Like

/k/ , /x/ is besides a velar consonant, which appears to be debatable for this patient population. These consequences are in understanding with earlier research9, 20, 22 The consequences refering distinction between groups can be explained by construction changes of the vocal piece of land after tumour engagement and intervention. Patients have more trouble with proper velar closing, ensuing in deformed velar address sounds. Trouble sing production of vowels is besides attributable to changes caused by tumour growing and intervention. Particularly patients who underwent intervention affecting the lingua may see more trouble with the production of vowels. In old surveies, vowels of patients treated for caput and cervix malignant neoplastic disease were considered pervert from vowels produced by healthy persons: F2 of all vowels was lowered compared to controls, and F1 of /i/ was elevated compared to controls.12, 18

In the present survey, the velar consonants /k, x/ and the vowels /a, I, u/ were selected from words that were obtained from running address. The phonological context of the selected address sounds may be of influence on the perceptual experience hereof and could besides be of influence on the consequences obtained in the present survey. Further research into speech quality of patients with caput and cervix malignant neoplastic disease could be performed on different address sounds in order to observe more features of address quality and more inside informations of specific address sounds. Besides, a different attack to objectively mensurate the address quality could be an analysis of address characteristics present in address such as nasality or voicing. Such a complex undertaking of ciphering address characteristics could be performed via automatic address acknowledgment utilizing a nervous web trained in designation of speech features.23-25 This attack might give extra penetration into the address of patients treated for caput and cervix malignant neoplastic disease. In the present survey, the consequences are based on postoperative informations merely and no efforts were made to compare these informations with preoperative address. Future research may concentrate on post- versus preoperative address quality in order to obtain more penetration into sensorimotor version capablenesss of patients to counterbalance for changes in the vocal piece of land after treatment.26, 27

Decision

Speech quality of patients after intervention of an unwritten or oropharyngeal tumor was investigated. Acoustic-phonetic analyses proved to be valid and are suited for mensurating speech quality of patients. The presented consequences

contribute to further development of a address analysis protocol to be used in clinical pattern and in clinical tests taking at bettering speech result in patients with caput and cervix malignant neoplastic disease.

Recognitions

The writers wish to thank Li Ying Chao, Pepijn Borggreven

and Milou Heiligers for their parts sing address recordings

and/or analyses.