|Year : 2014 | Volume
| Issue : 1 | Page : 32-37
Evaluation of speech perception in patients with ski slope hearing loss using Arabic consonant speech discrimination lists
Mai M El Ghazaly, Mohamed A Talaat, Mona I Mourad
Audiology Unit, ENT Department, Alexandria University, Alexandria, Egypt
|Date of Submission||24-Feb-2014|
|Date of Acceptance||07-Apr-2014|
|Date of Web Publication||28-Jul-2014|
Mai M El Ghazaly
22 Osman Galal Street, Moharam Beih Baraka, Tower 2, Alexandria
Source of Support: None, Conflict of Interest: None
Cochlear dead regions are regions of the basilar membrane where the inner hair cells and/or associated neurons function so poorly, such that they may be considered dead. Diagnosing patients suffering from ski slope hearing loss should put in consideration the possibility of cochlear dead regions. These patients often miss out high-frequency components of speech, which are consonant sounds (especially fricatives).
The current study was designed to evaluate speech perception of patients with ski slope high-frequency hearing loss on a modified Arabic consonant speech discrimination lists developed at the University of Alexandria and to evaluate the possible effect of high-frequency dead regions of the cochlear partition on such performance.
Materials and methods
Twenty patients with ski slope hearing loss were subjected to the Threshold Equalizing Noise (HL) test to bracket cochlear dead regions. The performance of each on the modified Arabic consonant speech discrimination lists was assessed and correlated with the presence or absence of cochlear dead regions and also with their extent if present across a number of spectral frequencies.
The results of this study showed that the average correct score of ears with no dead regions on the modified Arabic consonant discrimination lists was 75.32%, whereas the score of ears with dead region(s) was 61.19%. According to the extent of dead regions, the average score of ears with dead region at 4000 Hz only was 62.5%, that of ears with dead regions at 2000-4000 Hz was 61.8%, and that of ears with dead regions at 1000-4000 Hz was 56.5%. The highest probability of error in all ears was for the fricatives.
Speech tests that emphasize high-frequency speech elements are crucial in determining cochlear functional reserves in a practical manner. Psychophysical tests that investigate dead regions of the cochlea are synergistic to the high-frequency emphasis speech tests.
Keywords: Arabic consonant discrimination lists, cochlear dead regions, ski slope hearing loss, TEN (HL) test
|How to cite this article:|
El Ghazaly MM, Talaat MA, Mourad MI. Evaluation of speech perception in patients with ski slope hearing loss using Arabic consonant speech discrimination lists. Adv Arab Acad Audio-Vestibul J 2014;1:32-7
|How to cite this URL:|
El Ghazaly MM, Talaat MA, Mourad MI. Evaluation of speech perception in patients with ski slope hearing loss using Arabic consonant speech discrimination lists. Adv Arab Acad Audio-Vestibul J [serial online] 2014 [cited 2018 Jun 22];1:32-7. Available from: http://www.aaj.eg.net/text.asp?2014/1/1/32/137563
| Introduction|| |
Ski slope hearing loss is sharply falling configuration (≥30 dB increase between two successive high frequencies) where low frequencies thresholds remain about normal .
Cochlear dead regions may associate ski slope hearing loss where the inner hair cells and/or associated neurons function poorly such that mechanical vibration at a particular region of the basilar membrane will not cause signal transduction into its electrical counterpart [2,3]. A high-frequency sound may be detected by neurons that are tuned to lower frequencies. This is called off-frequency listening .
Hints in the audiogram that indicate a dead region at high frequencies may be present, such as 
(1) Threshold more than 90 dB at any high frequency.
(2) Audiometric slope is more than 50 dB/octave with increasing frequency.
Dead regions are diagnosed by the use of a masking paradigm procedure such as psychophysical tuning curves  and Threshold Equalizing Noise (TEN test). The former is used in research set up and lacks precise bracketing of the dead region . The latter is a clinical test and has the merit of accurate bracketing of cochlear dead regions [3,7].
Patients suffering from ski slope hearing loss often miss out consonant sounds (especially fricatives)  and have difficulties understanding speech in background noise.
Despite the correlation that has been demonstrated between the hearing loss for speech and the pure-tone average, it is not possible to arrive at the patient's ability to perceive speech from a pure-tone threshold audiogram. The cochlear high-frequency hearing loss sometimes score between 80 and 100% in phonetically balanced word lists, whereas they have difficulty in understanding speech conversation in certain situations .
Hence, it is necessary to use an effective consonants speech discriminative test to assess phoneme recognition difficulties for high-frequency hearing loss listeners, especially ski slope hearing loss. The effect of possible cochlear dead region on the outcome performance in such speech tests should not be overlooked .
The aim of this study was to evaluate speech perception of patients with ski slope high-frequency hearing loss on Arabic high-frequency emphasis speech materials and to evaluate the possible effect of high-frequency dead regions of the cochlear partition on such performance.
| Materials and methods|| |
The study was conducted on 20 patients with bilateral steep sloping high-frequency sensorineural hearing loss, with normal to moderate loss at 250-500 Hz, mild to moderately severe loss at 1000 Hz, and severe to total loss at 2000-8000 Hz. Categorization is borrowed from the known ranges of pure-tone averages.
Each patient was subjected to the following:
(1) Complete personal and medical history taking, full audiological assessment including pure-tone air and bone conduction thresholds and immitance measures.
(2) TEN (HL) test for diagnosis of cochlear dead regions.
(3) Speech audiometry using Arabic monosyllabic phonetically balanced words .
(4) Speech audiometry using a modified Arabic consonant speech discrimination lists .
TEN (HL) test
Threshold was measured for pure-tone detection presented in ipsilateral broadband noise using a recorded CD. The signals from the CD were fed through a two-channel audiometer, the noise on one channel and test tones on the other channel. Both channels are routed to one ear. Initially, the auditory pure-tone threshold is reached by descending manner where threshold is the lowest pure-tone level detected in 50% of trials. Then, the masked threshold is measured in the presence of a continuous background noise. Possible frequencies for the test tone are 500, 750, 1000, 1500, 2000, 3000, and 4000 Hz.
A dead region at a particular frequency is extrapolated when the masked threshold is at least 10 dB above the absolute threshold and 10 dB above the nominal noise level .
The original Arabic consonant discrimination lists
The original test developed in Alexandria University consisted of two lists containing multiple-choice items. The first list consists of 120 sense-word items and the second list consists of 156 nonsense-word items. Each set consists of four monosyllabic words; one is the stimulus word and the others are the alternatives. The syllabic structure is C V C. The words of a given item differ only in one phoneme either in initial or final position, whereas the remaining parts or the rhyme portions are spelled alike.
The modified Arabic consonant discrimination lists
The original lists had to be cut short for the sake of time in this study by the following procedure.
Construction of the new lists
The words in the two lists (sensible and nonsensible) were classified according to:
(1) The test consonants (fricatives, stops, nasals, and liquids).
(2) The position of the test consonant (initial and final).
Then, two words were chosen to represent each test consonant, one in the initial position and the other in the final one.
For sensible word list: In all, 24 words were chosen for fricatives (12 in the initial and 12 in the final position), 18 words for stops (nine in the initial and nine in the final position), four for nasals (two in the initial and two in the final position), and four for liquids (two in the initial and two in the final position).
Hence, the total for sensible word list was 50 words (25 words with the test consonant in the initial and 25 in the final positions).
For nonsensible word lists: In all, 26 words were chosen for fricatives (13 in the initial and 13 in the final position), 18 words for stops (nine in the initial and nine in the final position), four for nasals (two in the initial and two in the final position), and four for liquids (two in the initial and two in the final position).
Hence, the total for nonsensible word list was 52 words (26 words with the test consonant in the initial and 26 in the final positions).
Balancing the central vowel was considered in both lists. Then, the chosen words in each list were put into random order.
Hence, the developed lists are considered a shortened balanced version of the original lists.
The two lists were recorded separately through Jet Audio (Cowon system Inc., Korea) recording program using a microphone and connecting Dell laptop (Dell computer company, Texas, USA) to MADSEN Astera audiometer (Otometrics, Denmark). The VU meter (Madsen audiometer) was adjusted so that speech sounds are within ±1 dB HL from the audiometric zero level.
A carrier phrase (fx1) precedes the test word. The carrier phrase and the test word occupy 5 s followed by 5 s of silence so that the patient has enough time to choose the word he/she thinks is right.
Each list took about 8.5 min.
Introduction to the patient
(1) The words were presented to the patient at the level of 40 dB above the speech reception threshold.
(2) The patient had to choose the word he/she heard from a closed set of words that differ from the test word in only one consonant either initial or final, which is the test consonant.
(3) The test score was reported as a percentage of the correct responses.
Performance on phonemic errors was categorically identified and scored according to the following equation:
| Results and discussion|| |
[Figure 1] shows the average audiometric thresholds of all 20 tested ski slope patients at the test frequencies 250, 500, 1000, 2000, 4000, and 8000 Hz.
The results of the TEN (HL) test illustrated in [Figure 2] showed that, among 40 tested ears, 18 ears had no dead regions, whereas 22 ears had dead region(s). Of the ears with dead region(s), 11 had dead region at 4000 Hz, six had dead regions extending from 2000 to 4000 Hz, and five had dead regions extending from 1000 to 4000 Hz.
|Figure 2: Distribution of ears according to the presence or absence of dead regions and the extent of dead regions if present.|
Click here to view
[Figure 3] and [Figure 4] show mean audiograms of the two groups of ears (those with no dead regions and those with dead region(s), respectively). These figures show that the ears with dead region(s) had poorer thresholds at high frequencies than the ears with no dead regions. This trend has been reported by Moore et al.  in a previous study.
[Figure 5] shows distribution of ears with dead region(s) according to the frequency of detection of a dead region at each audiometric frequency. All 22 ears with dead region(s) had a dead region at 4000 Hz, whereas 12 ears had dead regions at 2000 Hz and only five ears had dead region at 1000 Hz. No dead regions were found at 250 or 500 Hz.
|Figure 5: Distribution of ears with dead region(s) [DR(s)] according to their frequency of occurrence across frequencies.|
Click here to view
[Table 1] shows that there is a statistically significant difference in the performance of the two groups on the modified Arabic consonant discrimination lists and phonetically balanced discrimination lists; the performance on the latter list was better in both groups as denoted by the paired t-test.
[Table 2] shows a comparison between ears with no dead regions and those with dead region(s) with respect to total performance on modified Arabic consonant discrimination lists represented as percent correct. It shows that the total performance of ears with dead region(s) was significantly lower than that of the ears with no dead region. This means that Arabic discrimination lists, which are actually phoneme discrimination lists, are more sensitive in assessing performance of ski slope patients and the effect of dead regions on such performance. These results indicate that these two tests measure different aspects of auditory discrimination for speech - that is, whole word versus consonant discrimination.
|Table 1 Comparison between ears with no dead regions and those with dead region(s) according to their total score on modified Arabic consonant discrimination lists and on phonetically balanced discrimination lists|
Click here to view
|Table 2 Comparison between ears with no dead regions and those with dead region(s) with respect to percentage correct of scores on modified Arabic consonant discrimination lists|
Click here to view
This explains the variation of performance with hearing aids for those with high-frequency hearing loss reported by many previous studies. Unfortunately, in the great majority of studies, it is not clear whether or not a high-frequency dead region was present in any specific patient or group of patients [12,13].
In addition, Moore and Glasberg  studied the effect of dead regions on speech perception using low-pass filtered speech. For patients without dead regions, performance generally improved progressively with increasing cutoff frequency. This indicates that they were able to make use of high-frequency information. For patients with dead regions, two patterns of performance were observed. In the first pattern, performance initially improved with increasing cutoff frequency and then reached an asymptote. This indicates that they were not able to make use of high-frequency information. In the second pattern, performance initially improved with increasing cutoff frequency and then worsened with further increases. This indicates that amplification of high frequencies impaired performance . However, Moore's study did not provide a realistic way to assess speech performance of patients with dead regions and the effect of the extent of those dead regions on such performance, which was implemented in our study.
[Figure 6] shows that the total score of patients with dead region at 4000 Hz only was 62.5%, those with dead regions at 2000-4000 Hz was 61.8%, and those with dead regions at 1000-4000 Hz was 56.5%. This demonstrates that, as the extent of dead regions increases, the total performance of patients is degraded. This is in agreement with previous studies in which patients with contiguous dead regions in two to three frequency regions obtained less benefit from high frequencies, on average, than patients with isolated dead region(s) .
|Figure 6: Distribution of performance of ears with dead region(s) [DR(s)] on Arabic consonant discrimination lists represented as percent correct according to the extent of dead region(s).|
Click here to view
[Table 3] shows that the total probability of error was significantly higher in the dead region(s) ears than in those with no dead region. In addition, the highest probability of error was noticed for fricatives. This is explained by the known high-frequency spectrum of the fricatives especially for the voiceless ones, as the noise segment for the voiceless fricatives is wider than the voiced ones as shown in [Figure 7] [10,16]. Progressive cutoff frequency from back to front position in the vocal tract characterizes the voiceless fricatives. The noise segment begins at 2000 Hz for /ς/ consonant, below 3000 Hz for /s/ consonant then extends upward in strong and continuous manner to 10000 Hz. The friction noise for /f/ sound is very weak across the spectrum and begins at 4000 Hz. For /f/ sound, the friction noise begins at 6000-8000 Hz. The consonant /h/ has a different character than the other voiceless consonants, as it consists of formant-like bars corresponding in frequency to formants of the next vowels .
|Table 3 Comparison between ears with no dead regions and those with dead region(s) according to the probability of error of each class of consonants|
Click here to view
The energy of the voiceless fricatives /hͿ/, /x/, /δ/, /z/, /ε/, and /Я / appeared as formant-like bands, which extend with that of the following vowel. The noise source of /z/ had a cutoff frequency at 2000 Hz, that of /z/ at 4000 Hz, and that of /δ/ at 6000 Hz.
The most commonly confused consonant in this study was /θ/ in both groups, which can be explained by weak high-frequency friction noise at 4000 Hz.
The least confused consonant in both groups was /ε/. This can be explained by that the discrimination of the voiced fricatives depends on the combination of filtering noise with vocal activity; hence, as the noise segment began earlier, the discrimination of the voiced fricatives increases .
For the ears with dead regions at 4000 Hz only, the highest probability of error was for the /f/ fricative and this can be explained by the high-frequency noise burst at 6000-8000 Hz. However, for the ears with dead regions at 2000-4000 Hz and for those with dead regions at 1000-4000 Hz, the highest probability of error was for /θ/ fricative with spectral energy at 4000 Hz.
| Conclusion|| |
(1) In this study, speech tests using consonant sounds have proved to be effective in determining the performance that reflects the spectral regions affected by cochlear pathology.
(2) Masking techniques were pioneering evaluation of cochlear reserves and dead regions but not highlighting the parallel speech perception.
(3) As rehabilitation options are increasing with technical development, that is, digital hearing aids with speech enhancement techniques, frequency transposition hearing aids, or cochlear implants, rehabilitation decision should be based on realistic measures. The most important of which is evaluating speech perception that targets specific cochlear regions, which may vary widely from very limited affection when dead regions affect very narrow cochlear partition to much degraded performance in extensive dead regions.
(4) The realistic assessment is expected to display the shortcomings of the pathology that affects speech perception. In this respect, masking techniques that uncover cochlear nonfunctioning areas and degradation of their counterpart speech perception together with high-frequency emphasis speech tests are synergistic in bracketing one or more regions of the cochlea that are not functioning and their effect on speech perception.
| Acknowledgements|| |
| References|| |
|1.||Lloyd LL, Kaplan H. Audiometric interpretation: a manual of basic audiometric. Baltimore: University Park Press; 1978. |
|2.||Moore BCJ, Glasberg BR. A model of loudness perception applied to cochlear hearing loss. Auditory Neurosci 1997; 3 :289-311. |
|3.||Moore BCJ. Dead regions in the cochlea: diagnosis, perceptual consequences, and implications for the fitting of hearing aids. Trends Amplif 2001; 5 :1-34. |
|4.||O'Loughlin BJ, Moore BCJ. Off-frequency listening: effects on psychoacoustical tuning curves obtained in simultaneous and forward masking. J Acoust Soc Am 1981; 69 :1119-1125. |
|5.||Moore BCJ, Alcantara JI. The use of psychophysical tuning curves to explore dead regions in the cochlea. Ear Hear 2001; 22 :268-278. |
|6.||Derleth RP, Dau T, Kollmeier B. In: Dau T, Hohmann V, Kollmeier B, editors. Modelling masking patterns for sinusoidal and narrowband noise maskers. Psychophysics, physiology and models of hearing. Singapore: World Scientific; 1999. 622-630. |
|7.||Moore BCJ, Huss M, Vickers DA, Glasberg BR, Alcântara JI. A test for diagnosis of dead regions in the cochlea. Br J Audiol 2000; 34 :205-224. |
|8.||Elfenbein JL, Hardin-Jones MA, Davis JM. Oral communication skills of children who are hard of hearing. J Speech Hear Res 1994; 37 :216-216. |
|9.||Moore BCJ. Cochlear hearing loss: physiological, psychological and technical issues. Chichester, UK: Wiley-Interscience; 2007. |
|10.||Khairy G. Arabic consonants discrimination test for high frequency hearing loss listeners [master's thesis]. Egypt: Faculty of Medicine, University of Alexandria; 2000 |
|11.||Soliman S. Speech discrimination audiometry using Arabic phonetically-balanced words. Ain Shams Med J 1976; 27 :27-30. |
|12.||Turner CW, Cummings KJ. Speech audibility for listeners with high-frequency hearing loss. Am J Audiol 1999; 8 :47-56. |
|13.||Amos NE, Humes LE. The Contribution of High Frequencies to Speech Recognition in Sensorineural Hearing Loss, in Physiological and Psychophysical Bases of Auditory Function (eds Breebaart DJ, Houtsma JM, Kohlrausch A, Prijs VF, Schoonhoven R), Shaker Maastricht: 2001;437-444. |
|14.||Moore BCJ, Glasberg BR. Use of a loudness model for hearing aid fitting. I. Linear hearing aids. Br J Audiol 1998; 32 :301-319. |
|15.||Cox RM, Alexander GC, Johnson J, Rivera I. Cochlear dead regions in typical hearing aid candidates: prevalence and implications for use of high-frequency speech cues. Ear Hear 2010; 32 :339-348. |
|16.||Heinz JM, Stevens KN. On the properties of voiceless fricative consonants. J Acoust Soc Am 1956; 28 :303-310. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2], [Table 3]