Word-In-Noise Perception Test in Adults

doi:10.21203/rs.3.rs-5171623/v1

Background

The recently designed Word-in-noise Perception (WINP) test is a new technique for examining lexical-semantic scores by homotonic-monosyllabic words (HMWs) and white noise, which assesses auditory brain function in speech consonant detection. It is necessary to have a test that creates the best competitive conditions for evaluating meaning perception. Therefore, this study aimed to design a WINP test for adults using HMWs and spectrum speech noise (SSN).

Methodology:

This study was a test-development type that was conducted in a cross-sectional-comparative way, it included 110 young Persian speakers (61 men and 49 women) with mean age of 20 (0.56) years. The evaluations included checking the state of general health, sleep and mental states, basic audiological evaluations, dichotic digit test and WINP test using HMWs and SNN. HMWs consisted of 6 lists of 25 words with a vowel/consonant/vowel pattern, the words in each list have the same vowel. The content validity of HMWs was confirmed by 20 Persian language experts and calculated using content and index validity ratios. Its reliability was measured based on repeatability in test times, intraclass correlation coefficient and the comparison of HMWs scores in two repetitions. To calculate the norm values, the number of correct responses in the mean measurements was multiplied by 4 and expressed as a percentage. Mann-Whitney tests were used to compare the scores of the groups.

Results

The validity and reliability of WINP test using HMWs and SSN were determined. Mean CVR and mean CVI of the HMWs were equal to 0.99 and to 0.95 respectively. ICC in single and mean measurements were also calculated. An ICC close to 1 indicates greater validity of WPS and better consistency of HMWs across lists. The results showed that the obtained scores are stable and without measurement errors. Normal values of WINP test using HMWs and SSN were gotten, which were 54%, 69%, 82%, 90% and 94% at SNRs of -5, 0, + 5, +10, and + 15 dB, respectively. The variable of gender was not an effective factor for creating a difference in the mean scores of WINP test using HMWs and SSN (P = 0.989). There was no significant difference between WINP mean scores in SNR in different noises between the right and left ears. Also, there was significant difference between mean scores of WINP in SNRs from − 5 to + 15 for left and right ears.

Conclusion

Psychometric properties of the WINP test using HMWs and SNN have been confirmed for Persian speaking adults.

Speech Perception

Word

Noise

Validity

Reliability

Since everyday conversations and social communication take place in competitive situations, the most important human need for the sense of hearing is speech-in-noise perception [1]. The word-in-noise perception (WINP) test using homotonic monosyllabic words (HMWs) along with the use white noise was recently developed [2–5]. The WINP test consists of 6 lists, each list containing 25 HMWs, whose structural pattern is consonant/vowel/consonant, and the vowel of each list is the same. Since there are 6 vowels in Persian language, there are 6 lists of HMWs in this test (Appendix. 1). But considering that the power of the white noise signal is evenly distributed over time, white noise cannot be a suitable masker for the speech signal [2]. For this reason, the use of speech spectrum noise (SSN) can avoid additional masking of the speech signal and increase the accuracy of the WINP test [2, 4].

Speech processing is done with the participation of the auditory brain and the brainstem. The brainstem is sensitive to the fundamental frequency of sound (f0), which is the psychoacoustical equivalent of speech vowels, and it can be seen in the FFR response. This is why musicians, who are exposed to more f0 are better at speech perception in noise than normal people. In general, the brainstem is most sensitive in discovering the rhythm of music and the pitch of speech, which is done based on the tonotopic map. The auditory brain also monitors and controls the neural signals received from the brainstem as a correcting and coordinating center, and its parts even participate in pitch processing based on the tonotopic map. There is another processing in the auditory brain, which is called the phonemotopic map, and it is not sensitive to f0 in this function. Indeed, the auditory brain is sensitive to the frequency difference of the first and second formant of vowels to process the phonemes and discover the consonants, and it is the f1-f2 connection processor to understand the meaning and concepts. As a result, a sentence uttered by different speakers (woman, man and child) conveys the same meaning to the listener [1–6]. Accordingly, if HMWs are presented to the patient in order and sequence, they cause the brainstem to have the least participation in speech processing and the neural function is focused towards the detection of speech consonants by the auditory brain [3–6]. It seems that WINP using SSN and HMWs can be a valid measure in the fields of prevention, diagnosis and rehabilitation of hearing disorders. For this reason, this study was aimed to psychometric investigation of WINP test using HMWs and SSN for Persian speakers aged 18 to 22 years.

Setting

This study was a test-making type that was conducted in a cross-sectional-comparative way. The audio recording location was the acoustic recording studio of *** University of Medical Sciences, the Fourier analysis of HMWs and psychoacoustics measurement of the recorded audio file was done at *** University of Medical Sciences, and the practical work site was the audiology department of *** University of Medical Sciences.

Inclusion criteria were Secondary education level, age less than 22 years and more than 18 years. Normal conditions in general health, night sleep, hearing thresholds, middle ear function, acoustic reflexes, word-in-noise discrimination, dichotic digit and pitch pattern scores, and without having a history of the following: Head injury, ototoxicity, metabolic and cardiovascular diseases, anemia, diabetes, hypothyroidism, hyperthyroidism, medication use psychedelic, nervous system diseases and addiction.

Exclusion criteria were unwillingness to continue cooperation in the research, inappropriate cooperation, inattention and insufficient accuracy of the participants.

Participants

110 young Persian speakers (61 men and 49 women) with mean age of 20 (0.56) years.

Ethical considerations

In this research, the privacy and security of the participants was a priority, and their evaluations were completely free.

Main outcome measures

After explaining the steps of the practical work and obtaining a written consent to cooperate in the research, 28-Item General Health Questionnaire, Petersburg Sleep Quality Index, and Mini Mental State Examination were distributed among participants (n = 131). Those who had normal results in this stage (n = 118) were evaluated by audiological tests, which consisted of acoustic immittance measurement (by clarinet middle ear immittance device), pure tone audiometry, speech reception threshold and word-in-noise discrimination (by Interacoustic AC33 audiometer), dichotic digit and pitch pattern tests (by the audio files recorded on the laptop). Out of the 118 participants, 110 of them had normal results in mentiented assessments, which were considered as our main sample size and were evaluated by the WINP test using HMWs and SSN.

In order to evaluate the validity of the content, the lists of HMWs were provided to 20 experts in Persian linguistics. After reviewing the answers, the content validity ratio (CVR) and content validity index (CVI) levels were determined for each of the items related to specificity, simplicity and clarity for the list of words. The words with CVR < 0.7 and or CVR < 0.42 were excluded. The psychoacoustics measurement of MMWs of WIN test was done using Fourier analysis. Then, in a studio equipped with the highest quality audio equipment, HMWs were recorded in silence and with the voice of a female Persian speaker without accent. HMWs with similar frequency bandwidth and energy level were selected and non-HMWs were removed. The generated SSN was white noise filtered speech noise with a flat spectrum in the frequency range of 250 to 1000 Hz and an energy decay of 6 dB per octave from 1000 to 6000 Hz [7]. Finally, SSN was added to the recorded voice at signal-to-noise ratios (SNRs) of -5, 0, + 5, +10, and + 15 dB.

The practical work of the WINP test using HMWs and SSN was in this way: It was done at the comfortable listening level of the participants [3]. The recorded voice was presented through a high-quality headphone connected to a computer. Each ear was examined separately. Participants were asked to respond quickly. The interval between the two words was 2000 ms. If there was no response, the next trial was automatically initiated 2000 ms after the last word was played [4]. The inter-trial interval of each 25 HMW list was 1000 ms. In other words, after presenting the last word in each list, there was a pause for 1000 ms and the first word of the second list was then presented after 3000 ms. To calculate the norm of the WINP test, the number of HMWs repeated correctly by 4 and and each correct answer of the patient is multiplied by 4, which is the result of scoring in percentage [4, 5]. The reliability of the WINP test using HMWs and SSN was calculated in terms of repeatability in evaluation times. Repeatability in test time was evaluated by Intra-class correlation coefficient (ICC). The list of HMWs were repeated to the participants twice, and the time interval between the two assessments was two weeks, so that the words did not remain in their minds. All 6 lists of HMWs were presented to the participants. The mean of all 6 lists in each ear was considered as the mean of the selfsame ear, and the mean of the whole right and left ear was considered as the norm of the test. The values of the indices calculated in two repetitions were exactly the same and the correlation value was equal to one.

Data analysis

Quantitative variables included mean and standard deviations. Kolmogorov-Smirinov test was used to evaluate the normal distributions. Norm values were expressed as mean ± standard deviation and as percentage. The Mann-Whitney and Friedman tests were used to compare the differences between scores. Significance level was determined to be less than 0.05.

In this study, the validity and reliability of WINP test using HMWs and SSN were determined in 110 Persian speakers. Mean CVR and mean CVI of the remaining words was equal to 0.99 and to 0.95 respectively. ICC in single and mean measurements were also calculated. An ICC close to 1 indicates greater validity of WPS and better consistency of HMWs across lists. The results showed that the obtained scores are stable and without measurement errors (Table 1).

Table 1

Reliability of word-in-noise perception test using homotonic monosyllabic words and speech spectrum noise for Persian speakers aged 18 to 22 years (n = 110).
	Intraclass correlation coefficient limit (0.95)		ICC in mean of measurements	Intraclass correlation coefficient limit (0.95)		ICC in one measurement	First test score		Second test score
SNR	Upper limit	Lower limit	ICC in mean of measurements	Upper limit	Lower limit	ICC in one measurement	Standard deviation	Mean	Standard deviation	Mean
-5	0.98	0.83	0.92	1.00	0.80	0.96	1.93	53	1.03	55
0	0.99	0.78	0.90	0.99	0.78	0.92	2.22	69	2.89	68
+ 5	0.98	0.79	0.95	0.98	0.79	0.93	2.43	82	3.74	83
+ 10	0.96	0.86	0.94	0.96	0.81	0.92	1.50	90	2.63	87
+ 15	1.00	0.88	0.96	1.00	0.85	0.95	1.44	95	1.08	95

SNR; signal-to-noise ratio

Normal values of WINP test using HMWs and SSN were gotten, which were 54%, 69%, 82%, 90% and 94% at SNRs of -5, 0, + 5, +10, and + 15 dB, respectively (Table 2).

Table 2

Comparing the mean scores of the word-in-noise perception test using homotonic monosyllabic words and speech spectrum noise for Persian speakers aged 18 to 22 years in different signal to noise ratios.
Ear	SNR					P-value
	-5	0	5	10	15
	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
Right	54 ± 6.03	69 ± 4.89	82 ± 5.74	90 ± 6.63	94 ± 5.08	< 0.001
Left	53 ± 6.24	68 ± 5.61	83 ± 4.88	89 ± 7.09	96 ± 3.90	< 0.001
Power	0.99	0.99	0.99	0.99	1.00
P-value	0.421	0.108	0.229	0.733	0.022

SNR; signal-to-noise ratio

The variable of gender was not an effective factor for creating a difference in the mean scores of WINP test using HMWs and SSN (P = 0.989).

The results of the Mann-Whitney test showed that considering that the p-value of the test is more than 0.05, there is no significant difference between WINP mean in SNR in different noises between the right and left ears (Table 3).

Table 3

Comparing the mean scores of the word-in-noise perception test using homotonic monosyllabic words and speech spectrum noise for Persian speakers aged 18 to 22 years in different signal to noise ratios.
Ear	SNR					P-value
	-5	0	5	10	15
	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
Right	54 ± 6.03	69 ± 4.89	82 ± 5.74	90 ± 6.63	94 ± 5.08	< 0.001
Left	53 ± 6.24	68 ± 5.61	83 ± 4.88	89 ± 7.09	96 ± 3.90	< 0.001
P-value	0.346	0.433	0.080	0.830	0.063

SNR; signal-to-noise ratio

Also, the results of Friedman's test showed that considering that the P-value of the test is less than 0.001, there is significant difference between WINP mean in SNRs from − 5 to + 15 for left and right ears (Table 3). Since there were significant differences between SNRs, Pair-by-pair comparisons between SNRs were performed by Friedman's supplementary test (Table 4).

Table 4. Pair-by-pair comparisons between different signal-to-noise ratios of the word-in-noise perception test using homotonic monosyllabic words and speech spectrum noise for Persian speakers aged 18 to 22 years in different signal to noise ratios.

SNR				Right Ear
10-15	5-10	0-5	-5-0	Right Ear
<0.001	<0.001	<0.001	<0.001	P-value
SNR				Left Ear
10-15	5-10	0-5	-5-0	Left Ear
<0.001	<0.001	<0.001	<0.001	P-value

SNR; signal-to-noise ratio

By comparing the norm values of WINP test using SSN for Persian speakers aged 18 to 22 years (at the SNRs of -5 dB = 53 ± 1.93, 0 dB = 69 ± 2.22, + 5 dB = 82 ± 2.43, + 10 dB = 90 ± 1.50 and + 15 dB = 95 ± 1.44) with the norm values obtained from the WINP test using white noise for Persian speakers aged 18 to 22 years (at the SNRs of 0 dB = 53 ± 14.80, + 5 dB = 68.15 ± 13.20, and + 10 dB = 88 ± 12.60) in recent study [4] can be seen that the mean of the WINP test using SSN in both ears are higher and the standard deviations are lower.

Previous researchers have reported that the high standard deviation of WINP using white noise in their studies [3–5], may be due to the occurrence of central auditory processing disorder (CAPD) in young adults [4], elderly [3] and patients with renal failure [5]. In this study, it can be assumed that the lower limit of the mean WINP test scores using SSN (in SNRs of -5 dB = 48%, 0 dB = 44%, + 5 dB = 64%, + 10 dB = 68% and + 15 dB = 84% likely to occur as a hidden cognitive lesion that could not be detected.

Audiological tests that are used to diagnose CAPD do not have the ability to determine the site of lesions, which is due to the simultaneous participation of central auditory neural centers in different functions of auditory processing [1, 2, 6]. For example, word-in-noise discrimination, dichotic digit, and pitch pattern tests that can detect abnormalities of the auditory brain and brainstem [2]. For the diagnosis of CAPD, various tests are usually employed, and since the WINP test minimizes the participation of the brainstem in the detection of vowels, it can be one of the appropriate tests to assessment the function of the auditory brain in the detection of speech consonants and lexical-semantic processing [3–5].

In silent conditions, the left hemisphere of the auditory brain prefers verbal stimuli [6, 8]. The left frontal pole, left dorsolateral prefrontal cortex, and right posterior parietal lobe are activated in any function associated with the attention mechanism, especially speech perception [9, 10]. In the presence of noise, each hemisphere works with the participation of all areas of the cerebral cortexes [11, 12]. Two widespread lexical-semantic processing streams are described for speech-in-noise perception: Bilaterally and largely symmetrically flows, which organized in superior temporal gyrus. Dorsal stream or posterior portion of superior temporal gyrus for sensory-motor interaction, that is left dominant and involves structures at the parietal-temporal-frontal junction [13].

When the process of speech-in-noise perception, the intraparietal sulcus is activated and also the planum temporale acts as a computational hub and plays an important role in distinguishing two simultaneous speech signals. Left insula cortex processes fundamentals frequency sounds containing semantic information, right insula cortex processes fundamentals frequency sounds sounds containing phonetic information. Ventral premotor cortex also increases the ability to distinguish vowels, especially when the signal-to-noise ratio decreases. Mirror neurons of the medial frontal cortex, which are active in imitation functions, are also involved in speech-in-noise perception [2, 14, 15].

When signsl to noise ratio decrease, and speech-in-noise perception in becomes difficult, brain areas associated with semantic processing and speech production are employed and possibly designating the use of overt tactics by the listeners, e.g. using overt vocalization to attempt to help perception [2, 10]. Wernicke's area's principal function is speech perception, but it as well cooperates to production. Broca's area's main role is in speech production, it is also involved in perception [2, 9]. When a person read a text aloud, the imaging of the words is created in the occipital cortex. Each word is shaped in a spatial form, which is different for each person. Then, created visual signals transmit from the occipital cortex to the angular gyrus and go to Wernicke's area [16].

Regardless of whether CAPD and other auditory impairments are unilateral or bilateral, speech-in-noise perception impairments appear to be due to damages of the superior temporal lobe of the language-dominant hemisphere, suggesting that speech inputs are processed asymmetrically at early stages [2, 8]. Bilateral impairment of the superior temporal lobe produce word deafness, a condition in which pure tone hearing thresholds are within normal limits. But the ability to speech perception is effectively nil. Damage to posterior temporal lobe areas are most predictive of speech perception deficits (in aphasia), [12].

Damage to the posterior portion of superior temporal gyrus (in certain forms of aphasia) does produce deficits in speech perception. However, these conditions involve only mild phonological impairments [13]. Anterior temporal lobe regions have also been implicated both in lexical-semantic and sentence-level processing. Patients with semantic dementia have atrophy involving the anterior temporal lobe bilaterally, along with deficits on lexical tasks such as naming, semantic association and single-word perception [17]. The anterior and inferior temporal lobes seems to be more cooperated in meaning procedure, whereas the anterior temporal lobe involved in integrating certain forms of semantic knowledge across modalities [18].

Damage to posterior temporal lobes, particularly along the middle temporal gyrus is associated with speech perception deficits [14]. It is well-established that auditory inputs can produce rapid and automatic effects on speech production. For example, delayed auditory feedback of one’s own voice disrupts speech fluency. The adult onset deafness is associated with articulatory decline, indicating that auditory feedback is important in maintaining articulatory tuning [15, 18].

Damage to the left dorsal posterior superior temporal gyrus is associated with production disorders [12]. In particular, such a lesion is associated with conduction aphasia, which is a syndrome typically caused by stroke that is characterized by good speech perception, but frequent phonemic errors in production, naming difficulties and struggle with verbatim repetition [16].

Finally, since patients with CAPD and various types of peripheral hearing loss have difficulty in speech-in-noise perception [2, 10], it is necessary to perform WINP test using SSN.

Psychometric properties of the WINP test using HMWs and SNN have been confirmed for Persian-speaking adults.

Ethics approval and consent to participate: The study was approved by the research ethics committee of Hamadan University of Medical Sciences (Code: IR. IR.UMSHA.REC.1402.018). Informed written consent to participate in the study was provided by all participants.

Funding: The financial sponsor of this research was Hamadan University of Medical Sciences (registered number: 140205173928), who supported in the design of the study and data collection, analysis and interpretation and did not include the writing of the manuscript.

Emami, S.F. Is all human hearing cochlear? The Scientific World Journal. 2013. doi: 10.1155/2013/147
Emami, S.F., Shariatpanahi, E., Gohari, N. et al. Word-in-noise perception test in children. Egypt J Otolaryngol. 2024; 40: 64. doi.org/10.1186/s43163-024-00625-1
Emami SF, Shariatpanahi E, Gohari N, Mehrabifard M. 2023. Aging and speech-in-noise perception. Indian J Otolaryngol Head Neck Surg. doi:org/10.1007/s12070-023-03689-2
Emami SF., 2024. The use of homotonic monosyllabic words in the Persian language for the word-in-noise perception test. Aud Vestib Res. 33(1):28-33. doi:org/10.18502/avr.v 33i1.14271
Emami SF, Momtaz HE, Mehrabifard M., 2023. Central auditory processing impairment in renal failure. Indian J Otolaryngol Head Neck Surg. doi:org/10.1007/s12070-023-04345-5
Emami, S.F. Acoustic Sensitivity of the Saccule and Daf Music. Iran J Otorhinolaryngol. 2014;4 26(75): 105–110. PMID: 24744999
Ebrahimi, A., Mahdavi M. E., Jalilvand H. 2020. Auditory recognition of Persian digits in presence of speech-spectrum noise and multi-talker babble: a validation study. Auditory and Vestibular Research 29(1): 39-47. doi.org/10.18502/avr.v29i1.2368
Chowsilpa S, Bamiou DE, Koohi N., 2021. Effectiveness of the auditory temporal ordering and resolution tests to detect central auditory processing disorder in adults with evidence of brain pathology: a systematic review and meta-analysis. Frontier in neurology. 12: (656117). doi:10.3389/fneur.2021.656117
Moulin A., 2022. Ear Asymmetry and Contextual Influences on Speech Perception in Hearing-Impaired Patients. Frontier in neurology. 16: (801699). doi:10.3389/fnins.2022.801699
Scott SK, Sinex DG, 2010. Speech. In: Rees A, Palmer A, editors. The Oxford handbook of auditory science: the auditory brain. 1^st ed. New York: Oxford university press: 193-214.
Hickok G, Okada K, Barr W, Pa J, Rogalsky C. 2008. Bilateral capacity for speech sound processing in auditory comprehension: Evidence from Wada procedures. Brain Lang. 107(3): 179-84.doi: 10.1016/j.bandl.2008.09.006
Vaden KI Jr, Kuchinsky SE, Ahlstrom JB, Dubno JR, Eckert MA, (2015) Cortical activity predicts which older adults recognize speech in noise and when. J Neurosci. 35(9):3929-37. doi:10.1523/JNEUROSCI.2908-14.2015
Tremblay P, Perron M, Deschamps I, Kennedy-Higgins D, Houde JC, Dick AS, Descoteaux, M. 2018. The role of the arcuate and middle longitudinal fasciculi in speech perception in noise in adulthood. Hum. Brain Mapp. doi:10.1002/HBM.24367
Humphries C, Willard K, Buchsbaum B, Hickok G. 2001. Role of anterior temporal cortex in auditory sentence comprehension: An fMRI study. Neuroreport. 12:1749-52.doi: 10.1097/00001756-200106130-00046
Tremblay P, Brisson V, Deschamps I. 2021. Brain aging and speech perception: Effects of background noise and talker variability. Neuroimage. 227:117675. doi:10.1016/j.neuroimage.2020.117675
Hickok G. 2009. The functional neuroanatomy of language. Physics Life Reviews 2009; 6: 121-43. doi: 10.1016/j.plrev.06.001
Humphries C, Willard K, Buchsbaum B, Hickok G,. 2001. Role of anterior temporal cortex in auditory sentence comprehension: An fMRI study. Neuroreport. 12:1749-52. doi: 10.1097/00001756-200106130-00046
Hickok G, Poeppel D, 2007. The cortical organization of speech processing. Nat Rev Neurosci. 8(5): 393-402. doi: 10.1038/nrn211310.1016/j.bandl.2008.09.006

No competing interests reported.

Appendix.docx

Word-In-Noise Perception Test in Adults

Status:

Version 1

Abstract

Background

Methodology:

Results

Conclusion

Background

Methods

Results

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1