Ann Otol Rhinol Laryngol

109:2000

FUNCTION OF A HEARING AID UNDER STRESSFUL CONDITIONS LOUIS W. WELSH, MD LAURIE

JOHN J. WELSH, MD

F. ROSEN, MA, CCC-A

JENKINTOWN, PENNSYLVANIA

The auditory function of individuals with normal hearing was compared with that of hearing-aided subjects of similar age to determine whether amplification remediates hearing impairment under stressful auditory situations. The specific tests of listening in a competitive noise environment and identifying moderately compressed speech were introduced to adequately aided individuals. The data indicate that noise had an impact on auditory function to a much greater degree in aided individuals than in matched counterparts with normal hearing. The data derived from acceleration of simple sentences delivered to the aided group suggested that contrary to basic tonal sensitivity, the capacity to understand the stimulus was greatly compromised. The authors discuss cochlear damage and central auditory impairment as they relate to the limitations of amplification for sensorineural hearing loss. KEY WORDS — competitive noise environment, compressed speech, hearing aid.

centrally, we recognize the components of auditory transmission through the brain stem and the thalamocortical radiations. Finally, a measure of the delayed far-field auditory-induced electrical activity may reflect the terminal cortical regions of sound recognition and speed of responsiveness or slowness of cognitive performance.

Our study examined this issue: How does the hearing-impaired ear respond to amplification under adverse conditions? Specifically, we assessed 2 areas that were of virtually universal concern to hearingaided individuals: 1) speech discrimination, that is, clarity of understanding during simultaneous exposure to a competitive noise environment, and 2) the capacity to hear and identify sentences composed of simple and common words when presented at a modestly accelerated rate.

These functional issues vary among individuals, and the reported data from this investigation confirm our experience that comparable or similar audiometric patterns do not predict identical results under the sophisticated stress of the test complexities.

Initially, we emphasize that this investigation does not include how to fit a hearing aid or the formulas for amplification and individualized frequency modulation. Neither do we examine or compare the varieties of hearing aids. Finally, we do not intend to recommend or suggest specific commercial products on the basis of this investigation.

The diagnostic analysis of hearing deficits in the peripheral or central auditory areas or a combination thereof depends on specific dedicated tests that reveal the integrity and function of each component. We used the model for the framework of word identification by Pisoni and Luce. ' This concept proposes that the sequence of events and analytic stages is initiated with peripheral auditory processing and proceeds to acoustic-phonetic and phonological analysis, thereafter to word recognition, and finally to lexical access. Normal performance is contingent on adequate quality and quantity of the received message. Defective performance at any stage before final integration into the word reservoir may cause a perceptual dysfunction.

We consider the results of this study applicable to the complaints of the hearing-aided subject. Most important, to clinicians who may be involved in the care of individuals with hearing problems, we suggest the limits of amplification, the margins of gain contingent on the nature and degree of the auditory disability, and, last, the boundaries of assistance and the need for suppression of exaggerated or unrealistic expectations. It is obviously impossible to predict in advance the exact gain achieved by the subject through amplification because of the unique and individual forms of disorders within the auditory system. This nonhomogeneity of damage includes cochlear hair cell integrity, that is, damage of the outer and inner hair cells, and the status of the primary neuronal structures, such as cellular population and function. More

The following data are based on the progression of hearing loss with age. The justification for this approach is the large database developed by longitudinal analysis over many years that defines the various functional disturbances in sensitivity, frequency analysis, speech and sentence discrimination, and hearing in competitive noise.

CORRESPONDENCE — Louis W. Welsh, MD, 179 Washington Ln, Jenkintown, PA 19046.

929

930

Welsh et al, Function of Hearing Aid Under Stressful Conditions

We recognize that the etiologic factors of hearing impairment in young individuals are different from those in older patients with presbycusis. Nonetheless, the hearing impairment still reflects dysfunction of the same functional communication skills as in elderly subjects. In addition, 90% of the subjects in this report fall within the 7th to 10th decades of life. Auditory function diminishes with age; this fact has been accepted for decades. The prevalence of age-related deafness is high, and the magnitude of the social and medical disorder increases with the progression of an aging population. Current data indicate that the incidence of hearing loss doubles per decade, beginning with 16% at 60 years of age and proceeding to 32% at 70 years of age and 64% at 80 years of age; few are spared hearing loss above 86 years of age.2 The evidence concerning hearing loss and age is briefly reviewed from a historical standpoint and in more current references to amplification. Pure tone audiometry identifies a decrease in sensitivity beginning at about 30 years of age, primarily involving the frequencies above 1 kHz and with continual progression beyond 60 years of age. The earliest reports, by Bunch,3·4 described a peripheral deficit with the characteristics of symmetric hearing loss with wide variation in auditory sensitivity (ie, degree of loss) and little correlation with age. A more recent update on this topic was compiled by the Working Group on Speech Understanding and Aging.5 Speech discrimination is depressed with aging. A broad correlation between tonal sensitivity and word identification exists; however, exceptions and disproportions may be seen.6"1 ' An analysis of speech perception in elderly patients through auditory and cognitive components discloses that in comparison with younger patients (18 to 28 years of age), the senior group (61 to 85 years of age) functions with greater individual variation, although considerable overlap is shown on tests of phonemes, spondees, and sentence perception. Also, the hearing loss of the elder group is largely related to high-frequency impairment. Sentence identification is more impaired than single-word testing. This dysfunction may begin between 30 and 39 years of age and thereafter increases rapidly with each subsequent decade.11·12 The ability to identify sentences correlates with the linguistic structure of the stimulus, the meaningfulness of the phrase, the presence of overlying syntactic constraints, and, especially, the rate of presentation.12 Consequently, as the complexity of the stimulus increases, accurate recognition falls. Conversely, simplicity of stimu-

lus, content, and structure corresponds with increased levels of identification. The recognition of speech with simultaneous sentence competition begins to diminish in the fourth decade; the failing ratio increases slowly to the seventh decade and continues more rapidly thereafter.13·14 The individual influence of distortion caused by high noise competition is so variable that subjects with similar performance in quiet may differ substantially in noise. The impact of speech babble noise on sentence identification is similar to that of a competitive sentence, namely, a function of age and hearing loss.15 Deterioration of the suppression mechanism is considered to be the foundation for this depressed function.16 The influence of age on masking level difference is discussed as a factor in decreased sensitivity17 and relates to impaired binaural auditory processing, intensity of peripheral stimulus, and reduction in the awareness of interaural difference and binaural cues.16 Temporal distortion of the speech stimulus results from time compression, interruption of speech, and reverberation of the verbal signal. These modifications of presentation produce a relatively severe performance deficit in verbal identification that is correlated with aging. Recall of sentences by older adults was found to be disproportionately depressed by increased speech rates.11·12·18"20 The diminishing performance is explained by a processing rate deficit and a superimposed defect in organization and integration of the verbal input. The "time sampling" phenomenon, a notion that implies an adequate temporal exposure within the central auditory system, is a fairly simplistic concept that may be correlated with other cognitive defects of hearing noted previously. Amplification of the stimulus level to correct intelligibility of temporally distorted speech fails to improve verbal sensitivity. More complex tests of binaural interaction and modification by frequency attenuation identify a progressive deterioration of function from the 7th to 10th decades. Presbycusis compounds the degree of performance deficit of the elderly during these examinations and may be correlated with the loss in tonal sensitivity. Our research suggests that increased loudness of the stimulus has limited value toward improvement of the dichotic input and identification of a distorted word stimulus.21 The effects of central auditory disability appear at a mean age of 60 years, with wide variation and progression thereafter. The consequences are observed in the more sophisticated aspects of hearing such as temporal modification of speech, simultaneous noise exposure, and rate of processing. Inadequate use of

Welsh et al. Function of Hearing Aid Under Stressful Conditions

temporal cues and speech perception,22 age-related behavioral slowness,23 and the diminished capacity to retain or recall information are cited as contributing factors toward inadequate verbal discrimination.10·24·25 Electrophysiological testing of auditory potentials in elderly subjects reveals a delayed wave V latency in patients 61 to 75 years of age that increases approximately 1.30 ms/y thereafter.26 An additional morphological change corresponding to a decreased amplitude of 0.18 mV/y was identified.27 Because diminished auditory processing may relate to the delay of the P-300 wave and simultaneous decreased voltage, we suggest that this observation may represent a defect in the cortical associations of hearing related to cognition. When this study began and sufficient data were accrued for analysis, we concluded that with the thenexisting instruments, and in the intervening years with new technology, amplification failed to solve hearing deficit in noise. Subsequently, Killion commented, "Hearing aids can solve the problem of loss of sensitivity..., but hearing aid circuits can not solve the SNR (signal noise ratio) problem caused by loss of inner hair cells or actual nerve deafness."28(P29) Bentler assessed the hearing aid and stated that "in high levels of background noise no previous or current hearing aid processing strategy is capable of improving SNR, ie, separating the intended speech signal from the unaided noise signal."29(P10) MATERIALS AND METHODS Each individual volunteered for the study, was not compensated, and had undergone a complete medical and otologic evaluation before entry into this project. The subsequent routine auditory studies included pure tone thresholds and speech discrimination. Thereafter, a hearing aid consistent with the demands of the patient was prescribed, and after several weeks for adaptation, evaluation of reception during noise and accelerated speech was completed. The audiological criteria for entry into the aided segment of this investigation included free-field speech reception thresholds at 25 dB sound pressure level (SPL) or better and a discrimination value greater than 70%. The performance requirements of the subjects with normal hearing used as reference to the aided individuals are contained in the subsequent section of data. These subjects are divided into 2 segments by the high-frequency function, although the speech reception threshold (SRT) and discrimination scores are quite similar. The hearing-impaired subjects were classified according to the following criteria for speech recep-

931

tion, which were coordinated with pure tone averages: 1) mild hearing loss, that is, 25 to 35 dB hearing level (HL); 2) moderate loss, that is, 35 to 55 dB HL; and 3) severe disability, greater than 60 dB HL. The focus of this report is on the second group, with a moderate loss, because in our experience this segment comprises the majority of hearing aid users, and the unaided function of auditory clarity is not markedly depressed (ie, greater than 75%). In addition to these factors, amplification usually improves the threshold of hearing to levels greater than 25 dB HL with some commensurate improvement in discrimination. Finally, the objective measurements of amplification under the adverse conditions of a testing situation are more consistent than in group 3 (severe loss), which has a greater disability in verbal sensitivity and accurate recognition. Although we do not present any data from the most impaired segment, these subjects failed to meet the criteria for entry, and when considered realistically, the functional disability is beyond complete remediation. Additional stipulations for entry into the investigation included 1) unilateral amplification; subjects with hearing aids on both ears were eliminated to provide an accurate and unbiased assessment of the isolated I -sided function; 2) the unaided ear must reveal an impairment greater than a 25-dB differential with the aided ear to eliminate crossover or binaural input; and 3) the test environment is constant, the measurement of speech is in a free field provided by paired speakers 45° azimuth from the midline, and there is negligible internal noise. INSTRUMENTATION

Testing was performed in a custom sound-treated acoustic environment. The stimuli included live voice for the measurement of SRT, and Audiotec (St Louis, Mo) cassette recordings of NU-6 lists 1 through 4 for speech discrimination in quiet. Sentence compression was presented by an Audiotec recording of the 30% rate and the speech babble (background noise) from the same source. The tapes were played on a Realistic SCT-86 cassette tape recorder-player coupled to Grason-Stadler (Milford, NH) clinical audiometer GSI-61. Sound field presentation of the verbal stimuli was through a Bose (Blythewood, SC) RM II speaker system and TDH 39P earphones (Telephonic, Farmingdale, NY). All equipment was calibrated in accordance with ANSI 1989 standards. PARAMETERS OF SPEECH STIMULUS

The intensity of speech presentation for the normal cohorts was 50 dB SPL, and for the aided subjects, sufficient amplification was added to the corrected SRT to effectively produce a 60-dB exposure level.

932

Welsh et al, Function of Hearing Aid Under Stressful Discrimination in Quiet and Noise Normal

0 20 40 60 80 100

m Mean î Range

SOU

i&oo

^Λ

10 20 §·

Age

30 6 22 29 24 34 36 3K 411 4 0 41 42 43 44 44 45 4H 50 52 58 6 0 62 64 73 T6 SI

Fig 1. Composite audiogram of normal subjects illustrates mean levels and ranges of sensitivity. Individual ages extend from 12 to 90 years. Lower graph confirms little negative impact on hearing in noise (+10-dB signal-to-noise ratio). SOUND-NOISE RELATIONSHIP

We selected a 10-dB signal-to-noise ratio (SNR) for this investigation. In general, the speech stimulus was presented at approximately 40 dB over the aided threshold response; the corrected range was thus between 55 and 65 dB HL. The SNR of 10 dB that we have used for the past 5 years appears to be adequate for verbal speech understanding. Killion28 considered the SNR required for 50% word recognition as a function of hearing loss. The linear curve for the incrementally required stimulus gain versus noise begins at a 20-dB loss. At this point, the value is approximately 5 dB, which increases to less than 10 dB at a 60-dB loss. At this level, an accelerated SNR is required for recognition beyond the approximately 70-dB threshold sensitivity. The quality and composition of the competing noise stimulus is relevant to this study. In the past, we have used broad-spectrum noise and a low-frequency background; currently, we use the simultaneous recording of 3 male voices and 1 female voice in a babble presentation (Auditec), which we have determined to be realistic and paralleling the tested words in frequency composition. The compressed speech test was used by Beasley et al30·31 in the evaluation of central auditory perceptual disorders. In addition to this application, Sticht and Gray32 examined time compression with respect to age and hearing loss. They concluded that attenuation of performance in elderly subjects was greater than that in young subjects, whereas the nature of the subjects' hearing loss did not affect the capacity for comprehension. The tape presents both 30% and 60% compression segments. We determined that the higher compression level was incomprehensible to almost all of the aided subjects and to many of the

Conditions

normal-hearing subjects, particularly elderly subjects; consequently, we deleted this rapid rate of presentation. The reported results are derived from the 30% compression sentences, which are presented in 10 sentences and are scored at 10% per unit. The individuals with normal hearing are actually measured by binaural input, and to a modest degree, the results are measurably better (less than 20%) than those for the subjects with a comparable monaural stimulus. A later report will consider the impact of competitive noise, environment, and understandability on profound unilateral deafness, but this presentation is primarily designed to evaluate individuals with 1 hearing aid and the negative influence of a "stressful" exposure to ambient noise. RESULTS We divided the data into 3 categories. Group A included individuals with normal hearing within 25 dB SPL in the frequency range of 250 to 4,000 Hz, maximal losses of 25 and 35 dB, respectively, at 4 and 8 kHz, an SRT of >20 dB, and discrimination of >90%. Group B included individuals with a loss of >45 dB at 4 and 8 kHz, an SRT between 20 and 30 dB, and discrimination of >88%. Group C was the hearing-impaired cohort with unaided sensitivity (SRT) of between 30 and 55 dB HL and a discrimination score of >70%. Early in the 5-year study, we observed the obvious impact of competitive noise on subjects with normal hearing. The common denominator in subjects who had depressed function in noise was identified as a high-frequency loss of sensitivity. Consequently, this group was separated from the totally normal subset (defined above), and comparisons of performance were made between the 2 categories (A and B) and the aided hearing-impaired subjects. A similar observation was drawn from the investigation of auditory performance under the challenge of compressed speech. The data from the 3 groups were compiled and examined by individual and group performance. INFLUENCE OF NOISE

Normal Hearing and Noise (Group A). This group included 27 individuals from 12 to 90 years of age (mean, 47.25 years). The basic composite audiogram reveals the mean pure tone sensitivity per frequency and the range of function of the group. The lower compartment of Fig 1 identifies the impact of noise on clarity. These figures are presented by age and score. Thirteen of the group had minor and probably insignificant depression of scores (less than 10%). Two subjects, 44 and 90 years of age, had some loss (28% each) as a result of the competitive noise.

933

Welsh et al. Function of Hearing Aid Under Stressful Conditions Compressed Sentences Normal Hearing jf Group A Group B

Discrimination in Quiet and Noise 0 20 40 60 80 100

High Frequency Loss t

■

ï

250

• Quiet

L

î 500

Γ"

!

t

K ^\

1060

T

X Noise

■ Mean

t

2000

ΐ Range

r~~~~-, 4000

8000

t î Î " !

! !

|

11

I

100 90 80 70 60 50

i

64 66 70 72 72 7Î 78 80 82 83 89 Age Fig 2. High-frequency loss of sensitivity is noted in audiometric pattern at 4 and 8 kHz. Impact of noise on these subjects reveals variable and individual loss of function extending from less than 10% to 50%.

Impact of Noise on High-Frequency Loss (Group B). This group comprised 11 subjects from 64 to 87 years of age (mean, 75.5 years). Figure 2 shows the configuration of the pure tone sensitivity, the mean range, and the gradual depression of the higher-frequency responses. A loss of auditory reception due to the influence of noise is confirmed by the graph, which suggests a negative impact on each individual of variable degree. The mean loss consequent to competitive noise exposure is 21%, with a range of 4% to 40%. Aided Discrimination in Quiet Versus Noise (Group C). Nineteen individuals were evaluated in this group. Their ages ranged from 24 to 96 years (mean, 72 years). Sensorineural hearing loss was the major deficit that required amplification, although two individuals had conductive losses as a result of congenital aural malformations or advanced destructive middle ear disease. The cochlear functions were generally within group B criteria, that is, high-frequency loss. Speech in Quiet and Noise: Aided P H c o JZ

3

zu

40 60 80 100

1

--4 — ( · - — at Ύ τ^

250

11 •

Quiet

» Noise

599

rr

1999

N

2999 ■

t * 1

k

f î 'I :

1

24 34

6U 61

63 71

73 74 7J

76 78

A wide range of disability is evident in Fig 3; the depression of function extends from very severe (>40%) to a relatively minor loss of 20% in noise. Nonetheless, 53% of the aided individuals were moderately to severely compromised, with >30% loss of sensitivity caused by the competitive effect of noise. IMPACT OF COMPRESSED SPEECH

Subjects With Normal Hearing and High-Frequency Loss. The individuals in groups A and B are integrated for comparison in this section, and the performance of the normal responders is compared with that of subjects with high-frequency loss (Fig 4). Although the numbers are not equal in groups A and B, an obvious difference in function is clearly noted. The preponderance of group A (85%) responded accurately to the sentence task, with a score of 90% or

79 80 81

81

Compressed Sentences Hearing Aided

100 D 80 O 3. 60 3

40 20

Age

83 87 89 9]

90 100

Since this segment (2 patients) with the mixed form of deafness, ie, sensorineural and conductive loss, was not statistically significant, comparisons were not drawn. Nonetheless, the responses were consistent with the larger segment of the sensorineural hearing loss subset.

1

T T

.

I

SO 60 70 M

% Correct Fig 4. Negative impact on understanding of compressed sentences is negligible in subjects with normal hearing (group A). High-frequency loss group (group B) is significantly impaired in this simple auditory challenge.

t Range

Be99

4999 ■

I

40

a Mean

1

r

!..lll

3

O 3

0

Fig 3. Composite mean audiogram of hearing-aided group confirms impaired sensitivity beyond normal range in each frequency. Lower graph depicts age and function and confirms loss of sensitivity with wide individual impact and variability due to simultaneous competing noise.

ε

DI

0

SO

W

70

«0

90

% Correct Fig 5. Hearing-aided subjects present wide range of functional loss; majority of group understands less than 50% of test.

934

Welsh et al, Function of Hearing Aid Under Stressful Conditions SRT

• = Air Conduction 0 10 20 30 40 50 60 70 80 90 100

DISC

·* = Bone Conduction