Portuguese Version

Year:  2003  Vol. 69   Ed. 5 - (13º)

Artigo Original

Pages: 672 to 677

PDF PT  

Correlation between sentences recognition thresholds in quiet and tonal thresholds

Author(s): Fabiana Soncini1,
Maristela J. Costa2,
Tania M. Tochetto de Oliveira3,
Luis Felipe D. Lopes4

Keywords: audiology, hearing, auditory perception, speech discrimination tests, audiometry-pure tone

Abstract:
Introduction: Tests of sentences recognition in quiet still are neither very used in the clinical audiological routine and nor in researches. That's why, many times, when this evaluation is accomplished, the clinicians doesn't possess parameters to interpret or classify the results. Aim: Thus, the aim of this study was to investigate the correlation between the sentences recognition threshold in quiet (SRTs) and the average of the tonal thresholds in the frequencies of 0,5, 1 and 2 kHz. Methods: 200 ears of normal-hearing individuals, 50 men and 50 women (average 34,61 years) were examined. First, an anamnesis, meatuscopy, pure tone audiometry, speech recognition threshold and speech recognition percent index were done. After that, by using the Portuguese Sentences Lists Test, the research of sentences recognition threshold in quiet (SRTs) took place. Results: The medium SRTs obtained in the 200 examined ears was 6,15 dBHL and the average of the tonal thresholds for the frequencies 0,5, 1 and 2 kHz was 8,55 dBHL. The statistical analysis revealed significant correlation between the variables studied. Conclusions: The average of the tonal thresholds in the frequencies of 0,5, 1 and 2 kHz can be used as reference for the analysis of the results obtained in the research of SRTs. Besides, it can be expected that the values of SRTs are better than the values obtained with the average of the tonal thresholds, because the sentences supply acoustic and linguistic tracks with meaning.

INTRODUCTION

Speech understanding allows human beings to communicate in an efficient fashion, being essential for social integration. Since it is such an important skill for mankind, it is one of the most common complaints in clinical practice, that is, difficulty to understand speech.

Therefore, the skill to understand speech should be considered the most important aspect to be measured in human auditory function, because it allows the assessment of the receptive communication function, providing data about how the subject works in situations of daily listening, based on objective information that is easily quantifiable.

In audiological clinical practice, there are tests for speech recognition with different types of stimuli (syllables, lists of meaningless and meaningful words, monosyllables, disyllables, among others). Among all of them, the presentation of words in quiet has been the most frequently used test.

However, we should try to measure the speech discrimination skill in situations that are close to real life, using speech recognition tests with sentences, since they are the verbal stimuli that best simulate the reality of daily communication 1.

For this reason, tests that use sentences to assess speech recognition, both in quiet and in the presence of competitive noise 2-5, have been developed in the past decades. In Brazil, the test developed by Costa6 was the first one to use sentences in Brazilian Portuguese to that end.

Even considering so many options of instruments to assess and the importance of such findings for precise clinical diagnosis, owing to the non-existence of parameters to interpret or classify the results obtained, the sentence recognition in quiet is still not frequently used in clinical situations nor in scientific trials. This is the reason why it is relevant to define normal range parameters for this type of assessment.

Based on such consideration, the study intended to investigate whether there is a correlation between sentence recognition index in quiet and mean of pure tone thresholds in frequencies of 0.5, 1 and 2 kHz.

MATERIAL AND METHOD

The present study was conducted at the Outpatient unit of Audiology, Serviço de Atendimento Fonoaudiológico (SAF), Federal University of Santa Maria (UFSM), in the year of 2002, after having been approved by the Research Ethics Committee of the Office of projects of CCS - UFSM (protocol CEP nº 050/02 - GAP 12654).

Volunteers were only assessed after they had received the information about objectives, justification and methodology of the proposed study and signed the Free Informed Consent Term.

The measures of the present study were obtained in an acoustically treated booth using a digital two-channel audiometer brand Fonix, model FA-12, type I and auricular phones type TDH-39P, brand Telephonics. The sentences were presented using the Compact Disc Player Digital Toshiba - 4149, coupled to the audiometer referred above.

The inclusion criteria to form the studied group were presence of audibility thresholds below 25dB HL in frequencies of 250 to 8000 Hz7.

Thus, to conduct the study, we analyzed 200 ears of normal hearing adult subjects, 50 female and 50 male subjects, aged from 18 to 63 years (mean age of 34.61 years).

The procedures we conducted were: anamnesis, meatoscopy, pure tone audiometry for air conduction in frequencies of 250 to 8000 Hz and bone conduction in frequencies of 500 to 4000 Hz; speech recognition index in percentage with disyllable words, and speech recognition index in percentage for monosyllable words. In the speech recognition tests the words were read by the examiner.

Later, we conducted an investigation of Sentence Recognition Thresholds in Quiet (SRTQ) by applying the Portuguese Sentences List test, developed by Costa6. The material is recorded in a CD and comprises lists of sentences in Brazilian Portuguese and noise with speech spectrum, recorded in independent channels, allowing the presentation of sentences in noise and in quiet.

Before testing, the output of CD channel in which the sentences were recorded was calibrated using the audiometer VU-meter. The pure tone of 1000Hz presented in the CD channel in which the sentences were recorded was placed to level zero.

The sentences were presented monaurally, using auricular phones and allowing assessment of each ear separately. We can see the list of sentences used in the test in Annex I.

It is important to point out that in a study conducted with auricular phones 8, a difference of 7dB was detected between the volume of recording in the two channels presented in the CD (speech and noise). Later, we conducted a spectrographic computer analysis of the material recorded in the CD, which demonstrated that the sentences were recorded in a mean intensity of 7dB below noise intensity. For this reason, the author of the test 9 referred that in the assessment with auricular phones it is necessary to have a subtraction of 7dB from the speech values observed in the device dial. Such criterion was followed in our study.

The sentences were presented in the following order:

a) Presentation of sentences 1 to 10 of the List 1A so that the subject could get familiarized with the test;
b) Presentation of list 1B on the left ear;
c) Presentation of list 2B on the right ear.

The procedure used to collect measures of SRTQ was sequential, adaptative or ascending-descending strategy 10, which allows the determination of the necessary level for the subject to correctly identify about 50% of the presented sentences.

Thus, following this strategy, the application of the test consisted of presenting speech stimuli in a specific intensity. When the response was correct, we reduced the intensity in which the stimulus was presented. When the response was incorrect, we increased the intensity for the next stimulus. A response was considered correct when the subject repeated without mistake or omission the complete sentence.

Owing to the technical characteristics of the device available to conduct the present study, we used interval of sentence presentation of 5dB and 2.5dB, respectively. The stimuli was presented in intervals of 5dB up to the first change in response pattern, and from then on, the presentation interval was of 2.5dB up to the end of the list.

To study SRTQ, the first sentence of each list was presented 10dB above the value found in speech recognition index for disyllable words, according to the device dial. It was observed in preliminary studies with the same material 8, 11 that this intensity is necessary to have a correct response of the first sentence presented to subjects with normal hearing, ensuring motivation of the patient to carry on the test.

The levels of presentation of each sentence were marked during the test. The mean of values was calculated based on the level of presentation in which the first change took place up to the level of presentation of the last sentence of the list. Thus, to obtain the Sentence Recognition Thresholds in Quiet we subtracted 7dB from the mean values of speech presentation shown by the equipment, as described before.

Statistical method:
To investigate the level of correlation between results obtained in the study of SRTQ and mean of pure tone values for frequencies of 0.5, 1 and 2 kHz, we employed Pearson's Correlation Coefficient.
The level of rejection of the null hypothesis was fixed at a value below or equal to 0.01 (1%). The statistically significant results were marked with an asterisk (*).

To carry out the statistical study of the means obtained with SRTQ and the mean values of pure tone thresholds for frequencies of 0.5, 1 and 2 kHz, we employed the Confidence interval formula for differences between two means, using a = 1%.

Based on this formula, we obtained the value of lower limit and upper limit to differentiate the means of the studied variables.

ANNEX I. List of sentences used in the study

List 1A
1. Não posso perder o ônibus.
2. Vamos tomar um cafezinho.
3. Preciso ir ao médico.
4. A porta da frente está aberta.
5. A comida tinha muito sal.
6. Cheguei atrasado para a reunião.
7. Vamos conversar lá na sala.
8. Depois, liga para mim.
9. Esqueci de pagar a conta.
10. Os preços subiram ontem.
11. O jantar está na mesa.
12. As crianças estão brincando.
13. Choveu muito nesse fim-de-semana.
14. Estou morrendo de saudade.
15. Olhe bem ao atravessar a rua.
16. Preciso pensar com calma.
17. Guardei o livro na primeira gaveta.
18. Hoje é meu dia de sorte.
19. O sol está muito quente.
20. Sua mãe acabou de sair de carro.
21. Ela vai viajar nas férias.
22. Não quero perder o avião.
23. Eu não conheci sua filha.
24. Ela precisa esperar na fila.
25. O banco fechou sua conta.

List 1B
1. O avião já está atrasado.
2. O preço da roupa não subiu.
3. O jantar da sua mãe estava bom.
4. Esqueci de ir ao banco.
5. Ganhei um carro azul lindo.
6. Ela não está com muita pressa.
7. Avisei seu filho agora.
8. Tem que esperar na fila.
9. Elas foram almoçar mais tarde.
10. Não pude chegar na hora.

List 2B
1. Acabei de passar um cafezinho.
2. A bolsa está dentro do carro.
3. Hoje não é meu dia de folga.
4. Encontrei seu irmão na rua.
5. Elas viajaram de avião.
6. Seu trabalho estará pronto amanhã.
7. Ainda não está na hora.
8. Parece que agora vai chover.
9. Esqueci de comprar os pães.
10. Ouvi uma música linda.

RESULTS

Next we describe the results obtained with the assessments made in the 200 examined ears in this study.
The statistical analysis did not show statistically significant difference concerning gender. Therefore, this variable was disregarded.

In Table 1, we can see the result of the analysis of the correlation between sentence recognition index in quite and mean of pure tone thresholds for frequencies of 0.5, 1 and 2 kHz, obtained for the application of Pearson Correlation Coefficient. The analysis revealed that there was strong and significant correlation (p<0.01) between the studied variables.

Table 2 presented the results obtained based on the application of the formula of confidence interval for difference of two means, using the study of the mean value differences found in the analyzed variables.

We observed that the difference between the pure tone thresholds for frequencies 0.5, 1 and 2 kHz and the mean of SRTQ was 2.40 dB HL and the value of amplitude of difference was 1.13 dB HL. Thus, we reached a lower limit of 1.27 dB HL and an upper limit of 3.53 dB HL for the differences between the variables.

Figure 1 shows the variation of the results of sentence recognition in quiet in relation to mean of pure tone thresholds in frequencies 0.5, 1 and 2 kHz in percentages. We observed that in 77.5% (n = 155) of the examiner ears the result of SRTQ was better than the mean of pure tone thresholds, that is, the intensity necessary to have SRTQ was lower than the mean of pure tone thresholds.

Table 1. Correlation between sentence recognition thresholds in quiet and mean pure tone thresholds for frequencies 0.5, 1 and 2 kHz.

* There was statistically significant correlation (p < 0.01).



Table 2. Results of the statistical analysis referring to differences in means obtained for SRTQ and means of pure tone thresholds for frequencies 0.5, 1 and 2 kHz.




Figure 1. Variation of sentence recognition thresholds in quiet (SRTQ) concerning mean of pure tone thresholds for frequencies 0.5, 1 and 2 kHz.

DISCUSSION

Using the Portuguese Sentences List Test we noticed that the mean SRTQ obtained in 200 examined ears was 6.15dB HL and the mean of pure tone frequencies of 0.5, 1 and 2 kHz was 8.55 dB HL (Table 1). We detected strong and significant correlation between variables in the study by the statistical analysis.

The findings of this study agreed with those obtained by other researchers that also conducted a comparative analysis of the mean value of SRTQ obtained with mean of pure tone thresholds of 0.5, 1 and 2 kHz and stated that these two variables of the audiological assessment were closely related. Plomp 12 referred that the mean of thresholds of 0.5, 1 and 2 kHz was close to the value obtained for sentence recognition thresholds in quiet and that the only parameters that influenced SRTQ was audibility thresholds. Smoorenburg13 emphasized the fact that the mean of 0.5, 1and 2 kHz would have a positive relation with prediction of SRTQ.

The strong relation between pure tone auditory thresholds and speech recognition thresholds has been demonstrated repetitively throughout the years 14-16. The mean of pure tone thresholds in frequencies of 0.5, 1 and 2 kHz has been used as the reference value for interpretation of the results in the study of speech recognition index, when the words are used as stimuli, being one of the tests normally conducted by audiologists.

As previously referred, the speech recognition thresholds of sentences is still little used and there are no reference values or standardization of results. However, the results obtained in this study, as well as those found in the specialized literature, show that the mean of pure tone thresholds 0.5, 1 and 2 kHz can be used as reference value for the analysis and interpretation of results obtained in sentence recognition thresholds in quiet.

Analyzing the results individually, we confirmed that SRTQ was better than the mean of pure tone thresholds in frequencies of 0.5, 1 and 2 kHz in 77.5% (n= 155) of the examined ears (Figure 1), that is, the intensity necessary to have SRTQ expressed in dB HL< was below the mean of the pure tone frequencies analyzed. The same was detected when we compared the men values found in the studied variables, as presented in Table 1. This finding was in agreement with those reported by Duquesnoy17, who stated in his study that SRTQ normally is better than that obtained as the mean of pure tone frequencies 0.5, 1 and 2 kHz.

It is believed that the best result of SRTQ when compared to the results of the mean thresholds of 0.5, 1 and 2kHz can be justified by the fact that sentences provide acoustic and linguistic clues in addition to being meaningful sound stimuli, differently from pure tone stimuli. It would facilitate the stimulus perception. Schochat1 reported that the skill in listening to speech involves intrinsic and extrinsic redundancies. In subjects with normal peripheral and central hearing, intrinsic redundancies refer to the many pathways and auditory tracts available in the central nervous system. Extrinsic redundancies refer to acoustic, syntaxes, semantic, morphological and lexical speech clues that, even though not always necessary, are extremely useful when the message is being expressed. It occurs in the case of tests that use sentences as speech stimuli. Cooper and Cutts18 and Festen and Plomp19 have also stated that the context of sentences provide clues that benefit speech intelligibility.

However, it is pointed out that these values were obtained in a studied group comprising normal hearing subjects, using the Portuguese list of sentences by Costa6 to detect SRTQ; the studied literature did not show any study that had been conducted with this kind of analysis and using the same test material.

It is suggested, therefore, that further studies are conducted so that a significant group of the population can be assessed to reach standardization of sentence recognition thresholds in quiet with the Portuguese Sentence List test. It is also highlighted the need to conduct the same analysis in groups of subjects with different audiometric configurations.

CONCLUSION

In the end of the present study, the critical analysis of results obtained with the group allowed us to conclude that in normal hearing subjects:

 there is correlation between mean of pure tone thresholds in 0.5, 1 and 2 kHz and sentence recognition thresholds in quiet (SRTQ);
 in most patients (77.5%) the values of sentence recognition thresholds in quiet were on average 2.40 dB HL better than the values of the pure tone thresholds of frequencies 0.5, 1 and 2 kHz;
 the mean of pure tone thresholds of frequencies 0.5, 1 and 2kHz can be used as reference for the analysis and interpretation of the results of the sentence recognition thresholds in quiet when we use the Portuguese Sentence List test, developed by Costa6, and the same calibration standards and procedures presented in the study.

REFERENCES

1. Schochat E. Percepção da Fala. In: __________ Processamento Auditivo. São Paulo: Lovise; 1996. cap. 1, p.15-42.
2. Plomp R, Mimpen AM. Improving the reability of testing the speech reception threshold for sentences. Audiology 1979; 18:43-52.
3. Bronkhorst AW, Plomp RA. Clinical test for the assessment of binaural speech perception in noise. Audiology 1990; 29:275-85.
4. Nilsson MJ, Soli SD, Sullivan J. Development of the hearing in noise test for the measurement of speech reception thresholds in quiet and in noise. J Acoust Soc Am feb 1994; 95(2):1085-99.
5. Nilsson MJ, Soli SD, Sumida A. Development of norms and percent intelligibility functions for the hint. House Ear Institute feb 1995; 1-9.
6. Costa MJ. Lista de sentenças em português: apresentação & estratégias de aplicação na audiologia. Santa Maria: Pallotti; 1998. 44p.
7. Davis H, Silverman RS. Hearing and deafness. 3rd. ed. New York: Holt Rinehart & Winston; 1970. cap. 9, p. 253-79.
8. Cóser PL, Costa MJ, Cóser MJS, Fukuda Y. Reconhecimento de sentenças no silêncio e no ruído em indivíduos portadores de perda auditiva induzida pelo ruído. Rev Bras ORL jul./ago 2000; 66(4):362-70.
9. Costa MJ. Comunicação Pessoal no Serviço de Atendimento Fonoaudiológico (SAF) da Universidade Federal de Santa Maria (UFSM) abr. 2002.
10. Levitt H, Rabiner LR. Use of a sequential strategy in intelligibility testing. J Acoust Soc Am May 1967; 42(3):609-12.
11. Soldera CLC. Limiar de reconhecimento de sentenças no silêncio e no ruído: aspectos que podem influenciar estas medida. 2001. Monografia (Especialização em Fonoaudiologia) Universidade Federal de Santa Maria Santa Maria; 2001. 53p.
12. Plomp R. Auditory handicap of hearing impairment and the limited benefit of hearing aid. J Acoust Soc Am 1978; 63:533-49.
13. Smoorenburg GF. Speech reception in quiet and in noise conditions by individuals with noise - induced hearing loss in relation to their tone audiogram. J Acoust Soc Am jan 1992; 91(1):421-437.
14. Hughson W, Thompson E. Correlation of hearing acuity for speech with discrete frequency audiograms. Arch Otolaryngol1942; 36:526-40.
15. Carhart M. Monitored live-voice as a test of auditory acuity. J Acoust Soc Am 1946; 17:338-49.
16. Fletcher H. A method of calculating hearing loss for speech from the audiogram. Acta Otolaringol Suppl 1950; 90:26-37.
17. Duquesnoy AJ. The intelligibility of sentences in quiet and in noise in aged listeners. J Acoust Soc Am 1983 oct; 74(4):1136-44.
18. Cooper JC, Cutts BP. Speech discrimination in noise. J Speech Hear Res 1971 jun; 14(2):332-7.
19. Festen JM, Plomp R. Effects of fluctuating noise and interfering speech on the speech-reception threshold for impaired and normal hearing. J Acoust Soc Am 1990 oct; 88(4):1725-36.

---

1 Clinical Speech and Hearing Pathologist, Specialization in Audiology, UFSM; Master studies in Human Communication Disorders under course, UFSM.
2 Clinical Speech and Hearing Pathologist, Ph.D. in Human Communication Disorders, Area of Speech and Hearing Pathology,
UNIFESP/SP; Audiologist with Centro de Educação, UFSM.
3 Clinical Speech and Hearing Pathologist, Ph.D. in Human Communication Disorders, UNIFESP/SP; Joint Professor,
Department of Otorhinolaryngology and Speech and Hearing Pathology, UFSM.
4 Mathematician, Ph.D. in Engineering, Federal University of Santa Catarina (UFSC)/SC; Joint Professor, Department of Statistics, UFSM.
Affiliation: Federal University of Santa Maria / UFSM.
Address correspondence to: Fabiana Soncini - Av. Nossa Senhora Medianeira, 1286/103 Santa Maria/RS 97060-002.
Tel (55 55) 223.9681 / 99776748 - E-mail: fabianasoncini@terra.com.br
Study presented as free communication at 18º Encontro Internacional de Audiologia, held on April 11 - 13, 2003 in Curitiba/PR.
Article submitted on June 04, 2003. Article accepted on July 17, 2003.

Print: