Citation
The Role of Clinically Applicable Temporal Resolution and Working Memory Tests in Prediction of Speech Perception and Hearing Handicap

Material Information

Title:
The Role of Clinically Applicable Temporal Resolution and Working Memory Tests in Prediction of Speech Perception and Hearing Handicap
Creator:
John, Andrew Barnabas
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (163 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Communication Sciences and Disorders
Committee Chair:
Hall, Jay
Committee Members:
Griffiths, Scott K.
Kricos, Patricia B.
West, Robin L.
Graduation Date:
8/11/2007

Subjects

Subjects / Keywords:
Auditory perception ( jstor )
Disabilities ( jstor )
Ears ( jstor )
Gin ( jstor )
Hearing loss ( jstor )
Hearing tests ( jstor )
Listening ( jstor )
Older adults ( jstor )
Temporal resolution ( jstor )
Working memory ( jstor )
Communication Sciences and Disorders -- Dissertations, Academic -- UF
hearing, memory, perception, resolution, speech, temporal, working
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Communication Sciences and Disorders thesis, Ph.D.

Notes

Abstract:
Clinical audiometric measures can only partially account for older adults? speech perception performance in difficult listening situations, such as in noise. Speech perception in noisy situations is the primary complaint of most individuals seeking hearing health care. It is not surprising, therefore, that clinical methods perform poorly in predicting self-reported hearing handicap. Two factors that have been proposed to account for this disconnect between audiometric hearing loss and speech perception difficulty are auditory temporal resolution and working memory. The role of these factors in speech perception performance in noise and in self-report of hearing handicap were examined for a group of middle-aged and older adults reporting a hearing loss. Experimental measures were chosen for their potential ease of implementation into a clinical setting. Declines in hearing sensitivity, temporal resolution, and working memory were found to be significantly correlated with decreased speech perception ability and increased hearing handicap. Loss of hearing sensitivity was the strongest predictor of speech perception difficulty. When hearing sensitivity was controlled, temporal resolution threshold was only significantly related to speech perception for the oldest participants. Self-reported hearing handicap was weakly associated with declines in hearing sensitivity, temporal resolution, and speech perception ability. The strongest predictor of hearing handicap was speech perception ability in noise. Previously unreported increases were seen in right-ear / left-ear asymmetry of temporal resolution threshold independent of decline in hearing sensitivity. In addition, more than two-thirds of participants were classified as having abnormal temporal processing according to published normative data. These findings suggest a need for re-evaluation of the temporal resolution task and establishment of age- and hearing-loss-appropriate normative values. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2007.
Local:
Adviser: Hall, Jay.
Statement of Responsibility:
by Andrew Barnabas John.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright John, Andrew Barnabas. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Classification:
LD1780 2007 ( lcc )

Downloads

This item has the following downloads:


Full Text





reported hearing handicap was weakly associated with declines in hearing sensitivity, temporal

resolution, and speech perception ability. The strongest predictor of hearing handicap was

speech perception ability in noise. Previously unreported increases were seen in right-ear / left-

ear asymmetry of temporal resolution threshold independent of decline in hearing sensitivity. In

addition, more than two-thirds of participants were classified as having abnormal temporal

processing according to published normative data. These findings suggest a need for re-

evaluation of the temporal resolution task and establishment of age- and hearing-loss-appropriate

normative values.









Little is known about the association between speech perception in noise and working

memory capacity. As detailed in the introduction, studies examining the role of working

memory in speech perception, using widely varying measurement methods, have not surprisingly

yielded widely varying results. Following the advice of Lunner (2003), sentence-length speech

stimuli were used in the present investigation in an attempt to identify working memory effects

that are not evident for simpler, word- or syllable-level stimuli (e.g., Humes et al, 1994; Humes

and Floyd, 2005). However, no working memory effects were apparent in HINT performance in

quiet or in noise. It is possible that the HINT sentences did not tax memory ability sufficiently

because they are, by design, simple in vocabulary and structure and fairly high in context (i.e.,

"The boy fell from the window;" "Strawberry jam is sweet."). Findings reported by Lunner

(2003) and Vaughan and colleagues (2006) support this possibility. These researchers reported

working memory effects for speech perception using low-context and nonsense sentences.

Furthermore, Tun and colleagues (1991) noted apparent working memory effects for recall of

portions of lengthy expository passages. Use of more complex or lower-context speech stimuli

may have revealed working memory effects not seen using the simpler HINT sentences.

A major potential sample limitation in the present investigation was the gender imbalance

seen among the participant group (17 male, 39 female). Because participants were recruited as

young as the sixth decade, it is unlikely that the gender imbalance reflects a survival effect.

Moreover, the direct recruitment method was not overtly gender-biased. Participants were

recruited from the Alachua county area principally using flyers posted on campus and at

recruitment talks given at locations around Gainesville. The best explanation for the high

proportion of female participants is that women enrolled early during the data collection period

were more likely to ask their friends and neighbors to participate. Approximately one-half of the









compared to younger listeners. The findings reported by Gordon-Salant and Fitzgibbons (1997)

support a role of memory capacity in speech perception in difficult environments. Other studies

relating a loss of some cognitive ability to speech perception difficulty can also be explained in

terms of working memory. For example, in a study by van Rooij and Plomp (1990), one-third of

the variance in speech-recognition ability by older adult listeners could be explained by cognitive

factors such as processing efficiency and memory capacity. This study used sentence-length

stimuli presented in a background of noise and required listeners to retain and repeat each

sentence in its entirety.

Kjellberg (2004) hypothesized a model for the effect of memory load on speech perception

in difficult situations, and suggested that listening in reverberant and noisy environments may

affect storage in working memory, as more cognitive resources are required for phonological

processing when distortion of speech is present. Even when speech is understood, less

information may be stored in memory, contributing to decreased understanding of speech over

time and increased cognitive fatigue. A general cognitive decline, therefore, might manifest as

difficulty in speech perception in adverse conditions, as fewer cognitive resources become

available for processing. Lyxell and colleagues (2004) suggested that when components of the

auditory signal are either altered or lost due to hearing loss or environmental distortion, more

strain is placed on cognitive strategies such as making verbal inferences and disambiguating

unclear sounds. The success of the cognitive strategies depends upon the individual's capacity to

manipulate, process, and store information quickly and over short periods of time. A large

working memory capacity enables the individual to use more contextual information in the larger

speech signal to aid in disambiguation and inference. Thus, at least in theory, cognitive

strategies may be more precise and limited to the topic of the conversation.









BIOGRAPHICAL SKETCH

Andrew Barnabas John was born in Oklahoma City, Oklahoma. Andrew received his

bachelor of arts in Communication Sciences and Disorders with a minor in Linguistics from the

University of Florida in 2002. He began his doctoral work at the University of Florida in Fall

2003 and was awarded his Ph.D. in Summer 2007. In August 2007, Andrew will join the faculty

of the Department of Communication Sciences and Disorders at the University of Oklahoma

Health Science Center in Oklahoma City as an assistant professor.









abilities to matched individuals with SNHL. However, the error patterns of the individuals with

simulated hearing loss differed from those of individuals with SNHL. While some studies find

that hearing sensitivity predicts most of the variance in speech perception in noise (i.e. Humes et

al 1987, 1994), others find that sensitivity is inadequate to explain variance in speech perception

(Plomp, 1978; Plomp & Mimpen, 1979). Additional variables, such as age and temporal

processing ability, consistently show a small but significant correlation with speech perception

results even when hearing sensitivity is controlled (Gordon-Salant & Fitzgibbons, 1993, 1995;

Halling & Humes, 2000). Overall, predictions of speech perception performance in older adults

that are based on audibility tend to be accurate for speech perception in quiet conditions but

overestimate performance in noise (Hargus & Gordon-Salant, 1995; Pichora-Fuller & Singh,

2006; Schum, Matthews, & Lee, 1991).

Deficits in hearing sensitivity are also insufficient to explain other findings related to

speech perception in noise. For example, individuals with SNHL require a greater signal-to-

noise ratio (SNR) than individuals with normal hearing for perception of speech in noise. Stated

another way, not only do individuals with SNHL require a higher-intensity speech signal, these

individuals also require that the difference between that speech and any background noise be

greater in order to optimally perceive speech. In one study illustrating this effect, Plyler and

Hedrick (2002) administered a test of stop-consonant identification to groups of individuals with

normal hearing and with hearing loss. Even when the intensity of the stimulus phonemes was

raised to a level at which all acoustic information was audible, a group of listeners with hearing

loss still did not perform as well as listeners with normal hearing.

To examine the effects of SNHL that are not explained by loss of sensitivity, researchers

often have used spectrally-shaped masking noise to simulate threshold elevation, loudness









listening in quiet to listening in noisy or reverberant environments, and from monaural

headphone presentation to listening in diffuse sound fields. Working memory effects should, in

theory, be most apparent for difficult, lengthy speech stimuli and/or speech stimuli presented in

difficult listening situations, such as in the presence of noise. It would not be surprising,

therefore, to find little or even no role for memory capacity in recall of simple stimuli, such as

nonsense syllables or monosyllabic words, or when the auditory processing system is otherwise

insufficiently taxed by the repetition task.

As a result of the substantial methodological differences among investigations, it is

difficult to be certain of the magnitude of the impact of working memory on speech perception.

While several studies have revealed compelling evidence of a role for cognitive factors,

including memory, in speech perception, laboratory findings do not consistently support a strong

association. It is likely that working memory capacity does play a role, at least in difficult

listening situations for adults with impaired memory capacity and for lengthy, complex speech

signals that place a greater strain on the individual's ability to retain, process, and contextualize a

message.

However, it is less clear whether working memory effects influence speech perception for

adults who do not have significant cognitive impairment. For example, working memory might

play a role in degraded speech perception for adults who have slight memory declines as a result

of aging or who have age-normal memory abilities but rely upon them more heavily as a result of

peripheral sensory impairment (i.e. sensorineural hearing loss). If there is a role for working

memory in adults who are generally cognitively intact, use of a simple working memory battery

(such as the digit span and digit ordering measures described above) may help to account for the









For Group 2, only BINGIN was retained as a predictor variable in the regression equation

for HHITOT (R = .39, F1, 30 = 5.28, p = .029). Inclusion of BETAI and WM increased the

correlation coefficient from .387 to .493, which was not a significant change (F2, 28 = 1.74, p =

.195).

Summary

To review, the participant group showed the expected relations among predictor and

descriptor variables. Significant declines in hearing sensitivity, temporal resolution, and working

memory were seen with increasing age. Hearing sensitivity was correlated with temporal

resolution threshold, with speech perception ability, and with self-reported hearing handicap, but

not with working memory ability.

As anticipated, declines in hearing sensitivity, temporal resolution, and working memory

were significantly correlated with decreased speech perception ability and increased hearing

handicap. Loss of hearing sensitivity was the strongest predictor of speech perception difficulty,

accounting for the majority of the variance in HINT threshold both in quiet and in noise

conditions. When hearing sensitivity was controlled, neither temporal resolution ability nor

working memory capacity was significantly correlated with speech perception threshold.

Self-reported hearing handicap was significantly but weakly associated with declines in

hearing sensitivity, temporal resolution, and speech perception ability. No association was seen

between working memory and hearing handicap score. The strongest predictor of hearing

handicap was speech perception ability in noise. Multiple regression procedures revealed that, of

the experimental variables, temporal resolution threshold was the best predictor of total HHIA/E

score, although no combination of predictors could account for more than approximately 17% of

the variance in hearing handicap.









GUIDE TO ABBREVIATIONS


AI Articulation Index

BETAI Better-ear Articulation Index score

BETGIN Better-ear Gaps-in-Noise test threshold

BINAI Mean Articulation Index score between the left and right ears

BINGIN Mean Gaps-in-Noise test threshold between the left and right ears

daPa Deca-pascals

dB A Decibels A-weighting scale

dB HL Decibels hearing level

dB SPL Decibels sound pressure level

DIGBAK Digit span backward test

DIGFWD Digit span forward test

DIGORD Digit ordering test

DPOAE Distortion-product otoacoustic emissions

GIN Gaps-in-Noise test

HHIA Hearing Handicap Inventory for Adults

HHIE Hearing Handicap Inventory for the Elderly

HHIEMOT Hearing Handicap Inventory for Adults / for the Elderly Emotional subscale
score

HHISOC Hearing Handicap Inventory for Adults / for the Elderly Social/situational
subscale score

HHITOT Hearing Handicap Inventory for Adults / for the Elderly total score

HINT Hearing in Noise Test

HINTN Hearing in Noise Test threshold in noise condition



















II'
III


125 250 500 1k
Frequency


2k 4k 8k


Figure 4-7. Mean audiogram for participants age 70-89 (N = 20).









deviation bars illustrate the substantial variability in hearing status among this rather unselected

group of adults who thought they had a hearing problem. Mean SRT and word-recognition

scores (conducted at MCL) are reported along with mean pure-tone thresholds in Table 4-2.

A repeated-measures within-subjects comparison between right- and left-ear thresholds

revealed a significant difference at 4000 Hz only (F1, 55 = 9.36, p = .003), indicating that, overall,

participants had slightly but significantly better hearing sensitivity at 4000 Hz in their right ears

than in their left ears. Right- and left-ear audiometric thresholds were not significantly different

at any other test frequency. Neither the SRT nor word-recognition scores were significantly

different between the left and right ears in this participant group.

The mean AI in the sample was 0.83 in the right ear (SD = .18) and 0.78 in the left ear (SD

=.23). The mean better-ear AI was .85 (SD = .17). These AI scores predict generally good

ability to perceive speech on average but with substantial variability within the sample.

Variability in audibility was an expected finding given the recruiting strategy for this

investigation, i.e., individuals within a wide age range and with no upper or lower limit on

degree of hearing loss. A repeated-measures within-subjects comparison of AI scores revealed

significantly better AI in the right ears compared to the left ears in this participant sample (F1, 55

= 4.02, p = .050). While significant, this difference was small, with an average difference of

0.05, equivalent to 5% speech intelligibility.

Audibility was significantly correlated with age. The Pearson's product moment

correlation (r) between right-ear AI (RAI) and age was -.75 (p < .001), and the correlation

between left-ear AI (LAI) and age was -.57 (p < .001). The principal audibility measure, better-

ear AI (BETAI), also was significantly correlated with age (r = -.76, p < .001). The correlations









multiple talkers, as in a group conversational setting, which creates fluctuating and varying noise

competition that masks not only acoustically by obscuring speech signals, but also

informationally in that the noise has verbal content. Further investigation is needed to determine

whether more complex speech perception measures, such as tasks using more difficult stimuli,

modulated noise makers, or reverberant test environments, may reveal variable associations not

seen here. In addition, speech perception tasks that purport to simulate real-world listening

environments should better account for factors described above such as multi-modal distraction

and increased load on speech processing in the presence of multiple verbal signals.

Clearly factors beyond peripheral hearing sensitivity contribute to the commonly-seen

speech perception difficulties of older adults. The failure to find all of the hypothesized variable

relations in the present investigation should by no means be considered a definitive refutation of

the associations hypothesized and demonstrated in prior research. In all likelihood, limitations

both in the sampling methodology and in the speech perception measure made the anticipated

and previously-demonstrated correlations more difficult to detect.

The current findings are instructive in that they demonstrate the difficult of simulating

everyday listening environments in a laboratory setting, or even many clinical settings. Indeed, a

primary goal of the present investigation was to attempt to replicate significant laboratory

findings of the influence of temporal resolution and working memory on speech perception using

simpler clinical measures. That is, this investigation attempted to bridge the gap between

laboratory results, frequently obtained on selected samples and with complex equipment and test

methods, and clinical testing, which must necessarily be brief, easy to implement (with minimal

cost, equipment installation, or training to clinicians), and easy to interpret. The difficulty in









efficiency of left-ear auditory pathways, which decussate at the level of the brainstem and then

must pass through the pathways of the corpus callosum to reach the left-hemisphere auditory

processing centers.

The increasing asymmetry seen on the GIN test in the present investigation may be related

to physiological changes that affect the right and left auditory pathways differently. Most of the

previous studies using gap detection paradigms have focused on a single-number, binaural

measure of temporal resolution ability. It is likely that, in these studies, only the best ear is

responding, making ear differences undetectable. Further research is needed to confirm and

explain the apparent asymmetrical aging of gap detection ability.

Working Memory Findings

Performance on all working memory tasks declined with age, consistent with much of the

previous literature on memory and aging (Carpenter et al, 1994; Hultsch et al, 1992; Salthouse,

1991; Wingfield et al, 1988). Interestingly, the rate of change for each working memory task

was different within the range of this age sample. For example, age sub-group comparison

showed significant declines in forward digit span ability only from the seventh decade on; that is,

no performance difference between participants in their 50s and participants in their 60s. Digit

ordering ability, by contrast, declined between the sixth and seventh decades, but not afterward.

Some recent working memory studies have similarly shown different age-related decline patterns

among working memory measures (Meguro, Fujii, Yamadori, Tsukiura, Suzuki, Okuda, &

Osaka, 2000; Parente, de Taussik, Ferreira, and Kristensen, 2005). Inconsistencies in age effects

may be explained by differences in the cognitive abilities accessed for different working memory

tasks. Declines in digit span ability late in the aging process are likely to be related to loss of

short-term memory capacity (Palladino & De Beni, 1999). Digit ordering, which requires more

manipulation of information than do digit span tasks, may be more sensitive to declines in









The negative r values indicated that individuals with serious hearing loss (lower AI scores)

demonstrated significantly longer gap detection thresholds (higher GIN thresholds) than

individuals with little or no hearing loss. Monaural audibility measures accounted for about 26%

of the variance in temporal resolution ability in the right ear and 41% in the left ear. When

controlling for audibility, GIN thresholds were not significantly correlated with age.

Musiek et al (2005) reported a normal average GIN threshold of 4.9 ms in the right ear (SD = 1.0

ms) and 4.8 ms (SD = 1.0 ms) in the left ear for a group of 50 normal hearing control subjects

aged 13 to 46 years. Using an upper limit for normal GIN test scores as two standard deviations

above the mean, Musiek and colleagues defined normal performance as a threshold of 7 ms or

better. Of the 56 participants in the present investigation, 40 yielded at least one GIN threshold

value greater than 7 ms. According to the normative criterion suggested by Musiek et al (2005),

therefore, 71% of the participants in the present study showed impaired auditory temporal

processing. Figure 4-4 and Table 4-4 display comparisons between the normative GIN data and

participant data collected in the present investigation.

Working Memory

Working memory was assessed using a battery of commonly-employed digit tasks: forward digit

span, backward digit span, and digit ordering. In order to create a composite working memory

score, individual test scores were standardized at the conclusion of data collection and the

resulting Z-scores were summed.

For the forward digit span test, a score of 8 approximates successful repetition of a

maximum six-digit string. The maximum possible score on forward digit span is 14. Forward

digit span (DIGFWD) scores averaged 8.1 (SD = 2.1) and ranged from 5 to 13.

Backward digit span (DIGBAK) scores averaged 6.8 (SD = 2.1). A score of 7

approximates successful repetition of a maximum four-to-five digit string, placing the digits in









in this participant sample confirm the commonly-encountered strong negative association

between hearing sensitivity and age in adults (worsening hearing with increasing age).

Of the fifty-six participants in this investigation, three demonstrated a large asymmetry in

hearing sensitivity, defined as a difference in AI of 0.25 or more between ears. One participant

presented with a congenital unilateral moderately severe hearing loss in her left ear; one

participant presented with a severe left-ear hearing loss reportedly due to Meniere's disease

diagnosed in the 1970s. Another participant presented with a moderate (20-25 dB) hearing loss

in her right ear of unknown etiology, with an apparent onset in childhood.

Temporal Resolution

Temporal resolution threshold in each ear for each participant was calculated according to

the prescribed threshold method for the GIN test. Threshold was defined as the shortest gap the

participant was able to identify in at least four out of six trials, while also identifying the next-

longest gap at least four out of six trials. The average GIN thresholds for the participant group

were 7.6 ms in the right ear (SD = 2.1 ms) and 8.4 ms in the left ear (SD = 2.8 ms). Thresholds

ranged from 5 to 15 ms in the right ear and from 5 to 20 ms in the left ear. The distribution of

GIN thresholds is graphed in Figure 4-3. A repeated-measures within-subjects comparison

between right- and left-ear GIN thresholds revealed that left-ear thresholds were significantly

longer (worse) than right ear thresholds in this group (Fi, 55 = 6.01, p = .017).

Both right- and left-ear GIN thresholds were significantly correlated with age, with

audibility (as measured by the AI in the same ear as the GIN threshold), and with each other.

Better-ear, worse-ear, and binaural GIN threshold (calculated as the average of left- and right-ear

thresholds) were also significantly correlated with age and audibility (see Table 4-3 for a

correlation matrix for these variables). Monaural GIN and AI variables were moderately

correlated (LAI x LGIN: r= -.64, p <.001; RAIx RGIN: r = -.51, p <.001).









appropriately to sound (Hall, 2000). Measurement ofDPOAE status is highly site-specific as a

measure of cochlear function. Laboratory experiments have demonstrated that OAEs can be

measured in ears which have had the auditory nerve completely severed, indicating that

connection to the auditory cortex or the brainstem is unnecessary for OAEs to be present, and

that OAEs therefore are uniquely dependent on cochlear status (Siegel & Kim, 1982).

Distortion-product OAE measurement is made using a probe microphone placed in the ear

canal. For each DPOAE stimulus, a pair of pure tones is introduced into the canal at a specified

intensity and frequency ratio. Specifically, the intensity of the lower-frequency tone (Li) is 65

dB sound pressure level (SPL) and the intensity of the higher-frequency tone (L2) is 55 dB SPL.

The frequency ratio is held constant for each stimulus pair such that the frequency of the higher-

frequency tone (f2) is equal to 1.2 times the frequency of the lower tone (fl) when rounded to the

nearest one-hundredth, which can be expressed as f2/ f = 1.2. These approximate frequency

and intensity ratios have been found to optimize the distortion response and to enhance the

sensitivity of the DP measurement to cochlear deficit. That is, using these parameters, the DP

amplitude is maximally reduced when outer hair cell pathology is present (Hall, 2000).

When the outer hair cells are intact, the introduction of the fi and f2 tones produces

distortion responses at several discrete frequencies. These responses include harmonic tones,

produced at the simple multiples of the stimulus tones (i.e. 2f2 ,2fi); summation tones, produced

at various additive combinations of the stimulus tones (i.e. fi + f2, 2fi + f2); and difference tones,

produced at various subtractive combinations of the stimulus tones (i.e. fi f2, 2fi f2). For

DPOAE testing, the difference tone at 2fi f2 is the principal measure.

Repeated amplitude measurements of the distortion responses are made by the probe

microphone. For diagnostic purposes, these measurements focus on the amplitude of the DP












their adult clients by proiding treatment and rehabilitation that reflects the tique djHticulde ,and
experiences of each idir idual. Your participation will also provide insight for hearing researther s
into some ofthe ay factors that contribute to the complex nature of real-word hearing and
communication.


6. What will be done if you take part in this research 5rud?

To determine whether you meet study criteria, the principal investigator will perform a complete
hearing evaluation. Some of the tests that make up this evaluation Vi. require you to provide
response sich as repeating words thai you hear or pressing a button when you hear a beep. Other
tests ,lll require you onld 0 I1 cquiLl' while meaElrelDmnt arc taken of the health of your eari You
will be instructed throughout the evaluation. The rill heanrg evaluation will take approximately 45
minutes to complete. You will also complete a brief cognitive screening insunnment.

WhTether you qualify forparticipaio in the study or not, the results of this evaluation will be
provided to you at the end o four appoirAuncr. and the ihcsatigaor will be available to meet with you
to discuss the meaning of those results

11' ou meem the study crilerij and choo- to parlicipate, you will be asked to complete further testing.
These tests will include repeating sentences .ou hear from a speaker while you are seated in a sound-
treatd room. linenirg iu pauses in white noise played for ou ocr beaphoces, and repeating
numbers spoken by the investigator These tests will take approximately 45 minutes to cimpletc.

In addition, you will be asked to complete a series of questionnaires related to haring,
communication, and health. You will complete these qLuesinnaires wih the investigator. These uill
taei approximately 20 minutes to complke.

All testing for this ij-Lcstigation will be completed in one appoinmneni in ihe departmnen of
Co=munication Sciences and Dsorders in Daer Hall at the Uni rsity of Florida. Your visit should
lasr approximatel% ILo hours. You will be fie to take breaks at any point during testing.


7. What are the possible dbcomforts and risks?

There ai no anticipmed disKori.fod or risk to you as a participant in this study.

Throughout the study, the investigators will notify you o tn mfonrmaton ihat may become available
and migl atTed your decision to rerrain iLrL te sfd.

If you wish to discuss the infornation above or any disomforts you may ayperiente, you may ask
questions at the dnie of our appointment or call the Prnnipal Investigaor or cart person listed on
the front page of this forr

L What ar the ponible benefits to you?

If 3ou cho&.e io parnitipatr in this study, you will be provided with a comprehensive hearing
evaluation at no charge to yow
Approved By
Uin.-ersiiay if Fiorlda
Institutiorlal Review Boarv 02 PaF 2 of 4
Protocol # 2006-U-0754
For Use Through 09/15/2007









Phillips, D. P. & Hall, S. E. (2000). Independence of frequency channels in auditory gap
detection. Journal of the Acoustical Society ofAmerica, 108(6), 2957-2963.

Phillips, D. P. & Hall, S. E. (2002). Auditory temporal gap detection for noise markers with
partially overlapping and non-overlapping spectra. Hearing Research, 174(1-2), 133-141.

Phillips, D. P., & Smith, J. C. (2004). Correlation among within-channel and between-channel
auditory gap-detection thresholds in normal listeners. Perception, 33, 371-378.

Phillips, S. L, Gordon-Salant, S., Fitzgibbons, P. J., & Yeni-Komshian, G. (2000). Frequency
and temporal resolution in elderly listeners with good and poor word recognition. Journal of
Speech Language and Hearing Research, 43, 217-228.

Pichora-Fuller, M. K., Schneider, B. A., Benson, N. J., Hamstra, S. J., & Storzer, E. (2006).
Effect of age on detection of gaps in speech and nonspeech markers varying in duration and
spectral symmetry. Journal of the Acoustical Society ofAmerica, 119(2), 1143-1155.

Pichora-Fuller, M. K., Schneider, B. A., & Daneman, M. (1995). How young and old adults
listen to and remember speech in noise. Journal of the Acoustical Society ofAmerica, 97, 593-
608.

Pichora-Fuller, M. K., Schneider, B. A., MacDonald, E., Pass, H., & Brown, S. (2007).
Temporal jitter disrupts speech intelligibility: A simulation of auditory aging. Hearing
Research, 223(1-2), 114-121.

Pichora-Fuller, M. K. & Singh, G. (2006). Effects of age on auditory and cognitive processing:
Implications for hearing aid fitting and audiologic rehabilitation. Trends in Amplification, 10(1),
29-59.

Pichora-Fuller, M. K. & Souza, P. E. (2003). Effects of aging on auditory processing of speech.
International Journal ofAudiology 42 (Suppl 2), 2S11-2S16.

Plomp, R. (1978). Auditory handicap of hearing impairment and the limited benefit of hearing
aids. Journal of the Acoustic Society ofAmerica, 63(2), 533-549.

Plomp, R. (1986). A signal-to-noise ratio model for the speech-reception threshold of the
hearing impaired. Journal of Speech and Hearing Research, 29, 146-154.

Plomp, R. & Mimpen, A. M. (1979). Speech-reception threshold for sentences as a function of
age and noise level. Journal of the Acoustic Society ofAmerica, 66(5), 1333-1342.

Plyler, P. H. & Hedrick, M. S. (2002). Effects of stimulus presentation level on stop consonant
identification in normal-hearing and hearing-impaired listeners. Journal of the American
Academy ofAudiology, 13(3), 154-159.









However, other speech-perception studies have calculated PTAs with thresholds at 500, 1000,

2000, and 4000 Hz (e.g., Divenyi & Haupt, 1997; Vaughan et al, 2006); 1000, 1500, and 2000

Hz (e.g., Plyler & Hedrick, 2002); 1000, 2000, and 4000 Hz (e.g., Gordon-Salant & Fitzgibbons,

1993; Humes, 1996); 2000 and 4000 Hz (e.g., Arlinger & Dryselius, 1990); and 2000, 3000,

4000, and 6000 Hz (e.g., Helfer & Wilber, 1990). The PTA is generally the simplest way to

quantify an individual's hearing sensitivity and is commonly used to quantify audibility. Despite

the wide use of this technique, the substantial variability in frequencies chosen for the average

illustrates the lack of agreement among researchers on a technique for quantifying audibility as a

single-number variable.

Audibility can also be quantified using the Articulation Index (e.g., Dubno & Dirks, 1989,

1990; Dubno, Horwitz, & Ahlstrom, 2005; Festen & Plomp, 1990; Lee & Humes, 1992; Schum

et al, 1991; Souza & Turner, 1999; Souza et al, 2000; Turner & Henry, 2002). The AI is often

chosen as an audibility measure instead of a PTA because the weighting built in to the index

emphasizes the contribution of frequency bands most important for speech perception.

For each of these methods, a binaural index measure for audibility is calculated using some

combination of left-ear and right-ear threshold measures, PTA, or AI. In most cases, this is

accomplished by using only the best ear measure (e.g., Duquesnoy & Plomp, 1980; George et al,

2007; Ventry & Weinstein, 1983). Less common is use of a binaural index derived by averaging

the left-ear and right-ear measures (e.g., Cokely & Humes, 1992; Humes and Floyd, 2005). As

with the PTA/AI methodology differences discussed above, no consensus exists for a preferred

method for conversion of monaural hearing sensitivity to a single-number binaural measure.

In everyday listening environments containing diffuse noise, it is likely that the better ear

predominates in detecting and processing speech signals. Thus, it is logical to presume that the










Table 4-1. Comparison of pure-tone audiogram status to DPOAE
frequency group.


status in each DPOAE


Normal Normal Abnormal Abnormal
DPOAE, DPOAE, DPOAE, DPOAE,
Normal Abnorm al Normal Abnormal.
Pure-Tone Pure-Tone Pure-Tone Pure-Tone
Right High 4 0 5 47 56
Left High 3 0 4 49 56
Right- Mid 25 0 6 25 56
Left Mid 21 4 5 26 56
Right- Low 16 0 25 15 56
Left Low 14 0 26 16 56













Rb. What are ihe posSibe beeatfi to other?

The information obtained from this study may help improve (be treaUment of eating loss in adults in
the future.


9. If you those to take part in this research study, wilt it cost you anything?

If you choose to participate in this snudy, it will cost you nothing.

10. Wil you receive compensation for taking parr la tbi, research study?

There will be no compensation for participating in this muly.


Sls. Can you withdraw from thi research study

You are fire to withdr our consent n ito stop participaring in diie reearch study at any time. Ifyou
do withdraw your sent, there will be no penalty, and you will not lose any benefits to which you are
entitled.

If you decide to withdraw yor consent to participate inls rearcb stud) Sr an% reason. you should
contact the principij rncgaws at the phone numbers above,

If 1 ou have any questions regarding your rights as a research subject you may phone the Institutional
Review Board (IRB) office at (352) 392-0433.


11L. Ifyou withdraw, can informatloa shot you still be used ander rolletred?

If you withdraw from the study, your permission will be sought to use information collected on your
prior to % our &ihidrawal


I I. Can the Principal InJoseigator withdraw you from tbb research study?

You may be withdrawn from the study without your consent for the following reasons:

You do nil quidli ro be in the study because you do not meet the study irquiremens. Ask the
ptncipal investigator if you would like more information about this.
You need medical treatinet not allowed in the study.
You are unable to keep appointments.


14a. How will %our pri acy and the confideunabty of your recmt records be protected?


Approved by
Uni rsity Of Florida
Institutional Review Bord 02 Pg 3 of 4
Protocol # 20O06.U.O75 F 3 4
For Use Througq 09/15i2'rC>7










Koehnke, J. & Besing, J. (2001). The effects of aging on binaural and spatial hearing. Seminars
in Hearing. 22(3), 241-254.

Koenig, W. (1950). Subjective effects in binaural hearing. Journal of the Acoustical Society of
America, 22, 61-62.

Koenig, A. H., Allen, J. B., Berkley, D. A., & Curtis, T. H. (1977). Determination of masking-
level differences in a reverberant environment. Journal of the Acoustical Society ofAmerica,
61, 1374-1376.

Kreisman, B. & Crandell, C. (2002). Frequency modulation (FM) systems for children with
normal hearing [Healthy Hearing]. Available at
http://www.healthyhearing.com/library/article_content.asp?articleid=160. Accessed June 1,
2007.

Larsby, B. & Arlinger, S. (1998). A method for evaluating temporal, spectral and combined
temporal-spectral resolution of hearing. Scandinavian Audiology, 27(1), 3-12.

Larsby, B., Hallgren, M., Lyxell, B., & Arlinger, S. (2005). Cognitive performance and
perceived effort in speech processing tasks: effects of different noise backgrounds in normal-
hearing and hearing-impaired subjects. International Journal ofAudiology, 44(3), 131-143.

Lee, L. W. & Humes, L. E. (1992). Factors associated with speech-recognition ability of the
hearing-impaired elderly. Journal of the American Speech Language Hearing Association,
34(10), 212.

Letowski, T. & Poch, N. (1996). Comprehension of time-compressed speech: Effects of age and
speech complexity. Journal of the American Academy ofAudiology, 7, 447-457.

Levy-Agresti, J. & Sperry, R. (1968). Differential perception capacities in major and minor
hemispheres. Proceedings of the National Academy of Sciences of the United States ofAmerica,
61, 1151.

Lindenberger, U. & Baltes, P. (1994). Sensory functioning and intelligence in old age: A strong
connection. Psychology andAging. 9, 339-355.

Lister, J., Besing, J., & Koehnke, J. (2002). Effects of age and frequency disparity on gap
discrimination. Journal of the Acoustical Society ofAmerica, 111, 2793-2800.

Lister, J., Koehnke, J., & Besing, J. (2000). Binaural gap duration discrimination in listeners
with impaired hearing and normal hearing. Ear and Hearing. 21(2), 141-150.

Lister, J., Roberts, R. A., Shackelford, J., & Rogers, C. J. (2006). An adaptive clinical test of
temporal resolution. American Journal ofAudiology, 15, 133-140.









Simon tasks and speech perception testing was "generally not seen," although certain memory

tasks were moderately and significantly correlated with speech perception performance.

In contrast, Vaughan and colleagues (2006) measured verbal working memory using a

series of "sequencing" and "non-sequencing" tasks including n-back recall and digit span.

Speech perception was evaluated using time-compressed IEEE and anomalous sentences. The

IEEE sentences are semantically and syntactically correct but have low predictability. The

anomalous sentences were designed to be similar in length and phonemic content to the IEEE

sentences but to have no sensible content (i.e. "Hang the wheel in the stupid air."). The authors

identified a significant correlation between working memory performance and speech

perception.

The methodological differences are clear between these two studies of working memory

and speech perception. These investigations were substantially different in their measures of

both working memory (visual Simon game versus verbal digit span and word recall) and speech

perception (nonsense syllables and connected speech versus time-compressed low-context and

low-semantic-content sentences). These studies also differ in age grouping criteria.

Other investigations of working memory and speech perception have similar conflicts in

methodology. In hearing research literature, working memory is identified variously using

measures such as digit span, word recall, sentence recall, subject/object pointing, digit symbol,

and visual pattern recognition and repetition. Some of these tasks do not appear to measure

working memory effectively and independently of other cognitive functions, while other tasks

are clearly tests of visual, and not verbal, working memory capacity (for example, the Simon

game). Furthermore, speech perception measures in the studies vary from identification at the

syllable level to perception and repetition of connected sentence-length speech stimuli, from









THE ROLE OF CLINICALLY APPLICABLE TEMPORAL RESOLUTION AND WORKING
MEMORY TESTS IN PREDICTION OF SPEECH PERCEPTION AND HEARING
HANDICAP




















By

ANDREW BARNABAS JOHN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2007



















GIN Summary Score Sheet .--,%

Name:____
vlb ....



U' Lrll Ir I5Ut ,f % I S
_______ a Is awe n 11 Mqw 3 MW TOW

%1 % % % %




% % % % % % % % % %
.......' ru a_ e_< as no n- a t n o






%-% % %% % *% % % r,





S, Ri tiOW C l itw Sone


% S %n 1 %%%


Falr rowdn Rkghl. Taml_______ Tr c l gItporr Fals Poirr.'
FAe. PstMis Lft x









executive function, one of the earliest cognitive processes to show changes with aging (Fisk &

Sharp, 2004; Meguro et al, 2000). The varying patterns of age-related change are notable in that

they suggest that the method of combining the three measures into a single working memory

score may have helped to balance these differing age effects across the sample.

One concern in designing the present investigation was the possible collinearity among the

predictor variables of audibility, temporal resolution, and working memory. While an

association between temporal resolution ability and hearing loss was anticipated (as described

above), no association was expected between working memory and either audibility or temporal

resolution. When controlling for age, there was no correlation between hearing loss and working

memory change. Previous research has also found independence between hearing status and

working memory when memory testing is conducted at an audible level (Gordon-Salant &

Fitzgibbons, 2001; Lyxell et al, 2004).

All temporal resolution variables were independent of working memory variables when

age was controlled. The lack of association among the predictors confirms that the GIN and

digit tasks accessed different abilities, and that testing their partial correlation with a separate

independent variable should not be confounded by any spurious collinearity, (e.g., correlation of

both variables with overall intelligence).

A review of the literature was conducted to identify reported relations among temporal

resolution and working memory variables. Three articles, all from a single research group,

described "temporal resolution of auditory perception" and verbal working memory in children

with language impairment (Fernell, Norrelgen, Bozkurt, Hellberg, & Lowing, 2002; Norrelgen,

Lacerda, & Forssberg, 2001, 2002). Like the present study design, each of these investigations

tested the separate predictive ability of a temporal resolution task and a working memory task on













20-


15 -
Right
*I H Left
0 10


0 0



2 3 4 5 6 8 10 12 15 20
GIN Threshold (ms)


Figure 4-3. Comparison of Gaps-In-Noise test thresholds between right and left ears.









loss (CHABA, 1988; Kim et al, 2005; Koehnke & Besing, 2001). Speech perception problems

are particularly common in older adults. Older adults demonstrate disproportionately more

difficulty listening in noisy environments when compared to younger adults with similar degrees

and configurations of hearing loss (Pichora-Fuller & Souza, 2003; Souza et al, 2000; Tun et al,

2002).

In the present investigation, HINT threshold in noise was more closely related to HHIA/E

score than were HINT threshold in quiet or audibility. That is, speech perception in noise was a

better predictor of hearing handicap than hearing sensitivity or speech perception in quiet. This

important finding supports the contention that speech perception in noise is a primary factor in

individuals' perception of their hearing loss. Moreover, audiologists should routinely include

speech perception in noise measures, such as the HINT, in a clinical test battery, particularly for

adults who may be candidates for hearing aids.

Hearing handicap score was weakly correlated with measures of auditory function.

Temporal resolution threshold accounted for about 14% of the variance in HHIA/E Emotional

subscale score and about 17% of the variance in total score. Audibility accounted for about 20%

of the variance in Social/Situational subscale score. Both GIN threshold and hearing sensitivity

were less closely associated with HHIA/E scores than the HINT threshold in noise.

The finding that GIN threshold was a better predictor than audibility for HHIA/E

Emotional subscale and total scores was unexpected. One possible explanation is that HHIA/E

scores may have been particularly influenced by participants' impressions of their own ability to

perceive speech in background noise, which has been shown to be related to temporal resolution

(Glasberg & Moore, 1989; Glasberg et al, 1987; Irwin & McAuley, 1987; Snell et al, 2002; Tyler

et al, 1982). Several of the questions on the HHIA/E ask either directly or indirectly about














Tist 12 /200
Location (rns) Ouraoln (rns)
I 2230.0 2
3571.3 D


3 438O.>

4 1985.9
3014.2
3745.

S 2433.6
5033.8

6 1308.9
1865.4
2681.0

7 1019.9
4179.4
5469.4

a 1275.5
2944,7
4918,3

9 872.4
1460.8
4869.5

In 3558.8

11 753.1
1290,7

12 2202.5

13 1546,5
2924.6
5014,3

14 718 7
2498.6
4546.5

Li 820-5
1675.9


Location (ms) Duration (rn
t1 1509.1 2
47595 3

19 1125.4 5

20 684 5 3
2673,1 12
3425.0 3

21 4238,4 a

22 3216-0 20

23 774.2 5
3276.4 12
4923.4 4

24 520.9 5
2799.5 5

25 1840.3 8

26 1209.1 5
5376.2 6

27 510,1 5
2549.9 20
439.3 6

2 6524.9 6
2737.8 12
4108.1 20

29 1319,7 20

30 711.7 4
43.6,1 L

31 269.9 8

32 1501.8 6


3726.3









Glasberg, B. R., Moore, B. C. J., & Bacon, S. P. (1987). Gap detection and masking in hearing-
impaired and normal-hearing subjects. Journal of the Acoustical Society of America, 81, 1546-
1556.

Goldstein, S. & Braun, L. (1974). Reversal of expected transfer as a function of increased age.
Perceptual and Motor ,k/l1, 38, 1139-1145.

Goldstein, S. & Shelly, C. (1981). Does the right hemisphere age more rapidly than the left?
Journal of Clinical Neuropsychology, 3, 67-78.

Gordon-Salant, S. (2005). Hearing loss and aging: New research findings and clinical
implications. Journal ofRehabilitation and Research Development 42 (Suppl 4), 9-24.

Gordon-Salant, S. & Fitzgibbons, P. J. (1993). Temporal factors and speech recognition
performance in young and elderly listeners. Journal of Speech and Hearing Research, 36(6),
1276-1285.

Gordon-Salant, S. & Fitzgibbons, P. J. (1995). Comparing recognition of distorted speech using
an equivalent signal-to-noise ratio index. Journal of Speech and Hearing Research, 38(3), 706-
713.

Gordon-Salant, S. & Fitzgibbons, P. J. (1997). Selected cognitive factors and speech recognition
performance among young and elderly listeners. Journal of Speech, Language, andHearing
Research, 40(2), 423-431.

Gordon-Salant, S. & Fitzgibbons, P. J. (1999). Profile of auditory temporal processing in older
listeners. Journal of Speech, Language, and Hearing Research, 42, 300-311.

Gordon-Salant, S. & Fitzgibbons, P. J. (2001). Sources of age-related recognition difficulty for
time-compressed speech. Journal of Speech, Language, and Hearing Research, 44, 709-719.

Gordon-Salant, S., Lantz, J., & Fitzgibbons, P. J. (1994). Age effects on measures of hearing
disability. Ear and Hearing. 15(3), 262-265.

Greenberg, S. (1996). Auditory processing of speech. In N.J. Lass (Ed.), Principles of
Experimental Phonetics (pp. 362-407). St. Louis: Mosby-Year Book.

Grose, J. H., Eddins, D. A., Hall, J. W. (1989). Gap detection as a function of stimulus
bandwidth with fixed high-frequency cutoff in normal-hearing and hearing-impaired listeners.
Journal of the Acoustical Society ofAmerica, 86(5), 1747-1755.

Grose, J. H. & Hall, J. W. (1992). Comodulation masking release for speech stimuli. Journal of
the Acoustical Society of America, 91(2), 1042-1050.









Pure-Tone Audiometry

Air- and bone-conduction thresholds were measured following a conventional clinical test

procedure as recommended by Carhart and Jerger (1959) and by ASHA (1978). Pure-tone

thresholds were obtained via air conduction at 250, 500, 1000, 2000, 3000, 4000, and 8000 Hz

bilaterally. Where inter-octave differences in hearing thresholds were 20 dB or greater, inter-

octave frequency thresholds were obtained at 750, 1500, and/or 6000 Hz. When conductive

hearing loss was suspected (based on tympanometry, DPOAE findings, and/or air-conduction

pure-tone thresholds), bone-conduction audiometry was conducted at the octave frequencies

from 250 to 4000 Hz. When indicated for either air- or bone-conduction testing, masking was

performed using narrow-band noise in the non-test ear.

Masking is used in audiometry when there is a possibility that the non-test ear can be

stimulated by the acoustic signal due to "crossover;" i.e., the signals presented are at intensity

levels sufficient to reach the non-test ear by either air or bone conduction. Presentation of

appropriate levels of steady-state noise corresponding to the frequency of the test stimulus

(narrow-band noise for pure-tone testing; speech-spectrum noise for speech testing) is used to

raise the threshold of the non-test ear to minimize the likelihood of a response due to a crossover

test signal.

Pure-tone hearing thresholds were obtained with the subject in a single-walled sound-

treated booth using a Grason-Stadler GSI-61 audiometer with TDH-49 supra-aural headphones

and a bone vibrator headband.

For the purpose of audiologic data analyses, hearing sensitivity was quantified using the

Articulation Index (AI), a measure derived from audiometric thresholds (ANSI, 1986; Fletcher &

Gait, 1950; French & Steinberg, 1947). The AI is used to express the amount of speech

information that is available to an individual based on hearing sensitivity level. Usually, the AI is









available to the individual for processing. Reduced working memory capacity can be overloaded

when information is more difficult to perceive, when distractors are present, or when efficiency

of data transmission is reduced (Wingfield, 1996).

A working memory overload might occur for individuals listening in difficult

environments, such as those with high levels of noise. Noise can cause speech signals to be

distorted, reducing the quality and quantity of the information received by the listener. If a

listener has reduced ability to process acoustic information, perception of the speech signal is

likely to be slow and inefficient and information may be lost (Kemper, 1992; Ryan et al, 1986).

Cohen (1987) hypothesizes that lengthy messages show the most loss in noisy situations, because

messages require storage of data and, also, the ability to relate early and late components of the

message to comprehend it in entirety. In addition to increased strain on working memory

capacity, reallocation of cognitive resources to compensate for the inability of working memory

to handle the incoming signal can cause a cascade effect in which higher processing is negatively

affected (Larsby, Hallgren, Lyxell, & Arlinger, 2005; Pichora-Fuller et al, 1995; Tun et al,

2002). Some researchers have suggested that working memory deficits appear when speed of

processing slows with aging, causing reallocation of resources to compensate and process rapidly

changing signals, (Fry & Hale, 2000; Salthouse, 1992).

Working memory may be a factor in age-related deficits in speech perception (Gordon-

Salant & Fitzgibbons, 1997; Kjellberg, 2004; Pichora-Fuller et al, 1995; Wingfield & Tun,

2001). Gordon-Salant and Fitzgibbons (1997) found that perception of sentence-length stimuli

in noise showed stronger age effects (with older listeners performing significantly more poorly)

compared to shorter (such as word-length) stimuli. Altering a recall task to increase the load on

memory during speech perception also had a greater detrimental effect on older listeners










Schum, D. J., Matthews, L. J., & Lee, F. S. (1991). Actual and predicted word-recognition
performance of elderly hearing-impaired listeners. Journal of Speech and Hearing Research, 34,
636-642.

Siegel, J. H. & Kim, D. 0. (1982). Cochlear biomechanics: Vulnerability to acoustic trauma
and other alterations as seen in neural responses and ear-canal sound pressure. In D. Hamernik,
D. Henderson, & R. Salvi (Eds.), New Perspectives on Noise-Induced Hearing Loss (pp. 137-
151). New York: Raven Press.

Smoorenburg, G. F. (1992). Speech reception in quiet and in noisy conditions by individuals
with noise-induced hearing loss in relation to their tone audiogram. Journal of the Acoustical
Society ofAmerica, 91(1), 421-437.

Snell, K. B. (1997). Age-related changes in temporal gap detection. Journal of the Acoustical
Society ofAmerica, 101, 2214-2220.

Snell, K. B. & Frisina, D. R. (2000). Relationships among age-related differences in gap
detection and word recognition. Journal of the Acoustical Society ofAmerica, 107, 1615-1626.

Snell, K. B., Mapes, F. M., Hickman, E. D., & Frisina, D. R. (2002). Word recognition in
competing babble and the effects of age, temporal processing, and absolute sensitivity. Journal
of the Acoustical Society ofAmerica, 112(2), 720-727.

Souza, P. E., Boike, K. T., Withrell, K., and Tremblay, K. (2007). Prediction of speech
recognition from audibility in older listeners with hearing loss: Effects of age, amplification, and
background noise. Journal of the American Academy ofAudiology, 18, 54-65.

Souza, P. E. & Turner, C. W. (1999). Quantifying the contribution of audibility to recognition of
compression-amplified speech. Ear and Hearing. 20(1), 12-20.

Souza, P. E., Yueh, B., Sarubbi, M., & Loovis, C. F. (2000). Fitting hearing aids with the
Articulation Index: Impact on hearing aid effectiveness. Journal of Rehabilitation Research and
Development, 37, 473-481.

Stach, B. A., Jerger, J. F., & Fleming, K. A. (1985). Central presbyacusis: A longitudinal case
study. Ear and Hearing. 6, 304-306.

Stach, B. A., Loiselle, L. H., & Jerger, J. F. (1991). Special hearing aid considerations in elderly
patients with auditory processing disorders. Ear and Hearing (Suppl), 12, 131s-137s.

Stine, E. L. & Wingfield, A. (1987). Process and strategy in memory for speech among younger
and older adults. Psychology andAging. 2(3), 272-279.

Stine, E. L., Wingfield, A., & Poon, L. W. (1986). How much and how fast: Rapid processing
of language in older adulthood. Psychology and Ain,'. 1(4), 303-311.









Distortion-Product Otoacoustic Emissions (DPOAEs) .................................. ...............64
A u d ib ility ................... ...................6...................5..........
Temporal Resolution ................................... .. .......... ............... 67
Working Memory ............................... ... .. .... ..... ................... 68
Collinearity Testing ................. ... ........... ..... .. ......... ........ 70
Speech Perception......... .......... ........ .. .. .... .... ..................70
H earin g H an dicap .............................................................................7 3
Gender.................. ...... .. .............................. ........ 76
Analysis of V ariance I: Grouping by A ge ........................................ ......... ............... 76
M multiple Regression, Grouping by Age.......................... ................... .................. 78
H IN T in Q uiet C ondition............................................. ................... ............... 78
H IN T in N oise C condition ....................................................................... ................... 79
H H IA /E Em otional Subscale................................................. .............................. 79
HHIA/E Social/Situational Subscale....................................................80
H H IA /E T total ........................ ...... ..................80
Analysis of Variance II: Grouping by Hearing Loss.................................................80
M multiple Regression, Grouping by H hearing Loss ........................................ .....................82
H IN T in Q uiet C ondition............................................. ................... ............... 82
H IN T in N oise C condition ....................................................................... ................... 83
H H IA /E Em otional Subscale................................................. .............................. 83
HHIA/E Social/Situational Subscale....................................................83
H H IA /E T total ................................................................83
S u m m ary ....................................................................................................... 8 4

5 D ISCU SS ION ......... ......... ....... .... .... ...................... ........ 110

Sam pling M ethod ...................................................... ........... ................ 111
Gaps-in-N oise (GIN) Test Findings .............................................................. ..............113
W working M em ory Findings ............................................................................. 117
Speech P perception ................................................................................ 119
H hearing H handicap ...............................................................122
L im itatio n s ................... ...................1...................2.........5
F u tu re D ire ctio n s ............................................................................................................ 12 9
S u m m ary ................... ...................1...................3.........0

APPENDIX

A INFORMED CONSENT .................................................................. ...........134

B GAPS IN N OISE TEST........................................................ 138

R E FE R E N C E L IST ...............................................................144

BIOGRAPHICAL SKETCH ................................................................... ........... 163






7









executive, which could be considered an attention-controlling system regulating the allotment of

cognitive resources to the other two "slave" components; (2) the visuospatial sketch pad, which

manipulates and processes visual images; and (3) the phonological loop, which stores and

rehearses verbal information. Working memory is an active process by which data is maintained

"online" for the purpose of problem solving, decision making, learning, and/or using language

(Baddeley, 1992; Baddeley & Hitch, 1974).

Each individual's working memory has a finite capacity or memory span that is usually

measured as the number of "chunks" of information that individual can hold in memory at once.

Miller (1956) found that young adults could hold an average of seven digits in memory at once.

Later research has clarified that the number of items in working memory is dependent upon the

nature of the stimulus items, such as whether they are numbers, letters, or words; the length and

complexity of the items (i.e., short versus long words); and whether the stimuli are familiar to the

individual tested (Hulme, Roodenrys, Brown, & Mercer, 1995).

Working memory status is commonly measured using simple tests such as digit span and

digit ordering. In a digit span paradigm, an individual is presented verbally with a series of

single-digit numbers and is asked to repeat them in the same sequence they were presented

(forward digit span) or in reverse of the sequence they were presented (backward digit span).

Digit span is available as a subtest of some intelligence scales, including all versions of the

Wechsler Adult Intelligence Scale (WAIS-III; The Psychological Corporation, 1997). The digit

ordering test is conducted similarly to digit span, but requires the tested individual not only to

retain series of numbers in memory and reproduce them, but to reproduce them in ascending

numerical order (Cooper, Sager, Jordan, Harvey, & Sullivan, 1991). For example, the tester

might say "10 1 8 2 4 7." The correct response would be "1 2 4 7 8 10."









(Glasberg & Moore, 1989; Glasberg et al, 1987; Haubert & Pichora-Fuller, 1999; Irwin &

McAuley, 1987; Phillips, Gordon-Salant, Fitzgibbons, & Yeni-Komshian, 2000; Snell et al,

2002; Souza, Boike, Withrell & Tremblay, 2007; Tyler et al, 1982), others have failed to find

such an association (Dubno & Dirks, 1990; Festen & Plomp, 1983; Snell & Frisina, 2000;

Strouse, Ashmead, Ohde, & Grantham, 1998; van Rooij & Plomp, 1991). The discrepancy in

findings among studies may be due to differences in the stimulus and procedure used for gap

detection testing. Gap thresholds are affected by characteristics of the carrier stimulus (the tone

or noise burst that contains the gap) including duration, intensity, and spectrum of the carrier

(Fitzgibbons, 1983; Fitzgibbons & Gordon-Salant, 1987; Jesteadt et al, 1982; Pichora-Fuller et

al, 2006).

Several investigations demonstrated a significant influence of carrier spectrum on gap

detection. A decrease in gap-detection threshold has been observed with increases in signal

frequency for groups of individuals with normal hearing and groups with SNHL (Fitzgibbons &

Gordon-Salant, 1987). The faster response of more broadly-tuned high-frequency auditory

filters may account for this increase in temporal acuity at higher frequency, but according to

some researchers gap-detection threshold is independent of frequency within certain ranges,

including 400 and 2000 Hz (Moore et al, 1989) and 500 and 4000 (Formby & Forrest, 1991).

The reason for frequency independence within certain ranges is not well understood, but the

existence of that independence may help to explain why gap detection studies using tonal stimuli

of varying spectral content and studies using noise stimuli often differ in findings.

There is also evidence that within-channel gap detection (detection of gaps bordered by

like carriers, such as identical tones or noise bursts) is a distinct process from between-channel

gap detection (detection of gaps bordered by different carriers, such as tones at different









CHAPTER 5
DISCUSSION

The inability of the standard audiometric test battery to account well for speech perception

ability in difficult listening situations and self-reported hearing handicap is well-known.

Audiologists commonly encounter patients whose hearing problems are not well explained by

the audiogram, or for whom certain listening situations are more difficult than would be expected

based on measurable loss of hearing. Two patients evaluated by the same audiologist, and with

similar audiometric profiles, may report and experience everyday listening difficulties that are

quite different in nature and in severity. Without appropriate test criteria to distinguish these

difficulties, the audiologist may manage both patients similarly. The outcome for each patient

may, however, vary substantially.

It is likely that both audiologic and non-audiologic factors contribute to hearing handicap

and success with rehabilitation. An audiologist may be able to identify some of the differences

among patients using good, sensitive tests of complex auditory capabilities. However, factors

that are not dependent on hearing status (such as attention and memory are generally) may be

more difficult to measure. In addition, the influence of both audiologic and non-audiologic

factors may be very different from person to person. A construction foreman whose job requires

him to communicate in very noisy situations may rely heavily on his eyesight to compensate for

his hearing loss by speechreading. A receptionist at a busy law practice may depend on attention

and working memory to listen and take messages rapidly on the phone. An elderly retiree might

have severe difficulty keeping up in evening conversations with her husband due to her inability

to focus on a single speech signal in the presence of auditory and visual distractions from the

television.









frequencies) (Divenyi & Danner, 1977; Formby, Gerber, Sherlock, & Magder, 1998; Grose,

Hall, Buss, & Hatch, 2001; Heinrich & Schneider, 2006; Phillips & Hall, 2000, 2002; Phillips &

Smith, 2004; Pichora-Fuller et al, 2006). The between-/within-channel discrepancy is most

pronounced in older adults (Fitzgibbons & Gordon-Salant, 1994, 1995; Grose, Hall, & Buss,

2001; Lister, Koehnke, & Besing, 2000; Lister & Tarver, 2004). Age effects for across-channel

gap detection may be related to difficulties experienced by older adults in resolving temporal

envelope information and integrating that information across channels (Souza et al, 2007).

Gap detection can also be confounded by other factors, including contribution of off-

frequency listening regions, placement of the gap within the carrier (i.e. midpoint of the carrier

vs. near stimulus onset or offset), and presentation and response mode (He, Horwitz, Dubno, &

Mills, 1999; Pichora-Fuller et al, 2006; Schneider & Hamstra, 1999). In addition, the association

between gap detection ability and speech perception is affected by speech perception materials

and presentation method (Lister, Roberts, Shackelford, & Rogers, 2006; Ochs, 1990; Snell et al,

2002). Generally, more complex speech perception stimuli elicit stronger temporal resolution

influences.

Several gap detection studies have demonstrated that older adults do not detect a gap in the

signal until the size of the gap is about twice as large as the smallest gap detectable by younger

adults (about 6 vs. 3 ms in tonal stimuli) with gap detection threshold poorly predicted by pure-

tone hearing thresholds in listeners with good audiograms (Gordon-Salant & Fitzgibbons, 1993;

Haubert & Pichora-Fuller, 1999; Schneider et al, 1994; Snell, 1997; Strouse et al, 1998). Also,

as listeners age they experience diminished ability to resolve simple temporal cues in speech, a

finding which may partially explain age effects on speech perception not explained fully by loss

of sensitivity (Fitzgibbons & Gordon-Salant, 1994; Gordon-Salant & Fitzgibbons, 1993; Snell &










Table 4-5. Comparison of participant working memory scores to UF Language Over the
Lifespan Laboratory normative data.
Participant Group Normative Data
Age Mean SD N Mean SD N t df


DIG
FWD









DIG
BAK









DIG
ORD


50 54
55-59
60 64
65 69
70 74
75 79
80- 84
85 and up

50 54
55-59
60 64
65 69
70 74
75 79
80- 84
85 and up

50 54
55-59
60 64
65 69
70 74
75 79
80- 84
85 and up


8.63
8.33
8.20
9.78
6.78
7.50
7.00
5.00

7.38
7.56
6.70
7.33
6.00
6.00
6.00
4.00

18.63
18.44
16.60
16.33
15.44
14.83
14.25
5.00


2.26
2.06
1.81
2.17
1.92
1.22
1.63
n/a

2.72
1.67
2.26
1.41
1.87
2.61
1.83
n/a

3.81
2.70
2.95
2.24
3.84
6.79
2.22
n/a


7.92
8.75
8.02
8.02
7.63
7.44


6.08
7.63
7.13
7.00
6.35
6.88


16.50
18.37
17.23
17.34
16.27
15.41


2.70
2.70
2.50
2.20
2.40
2.50


1.80
2.70
2.50
2.30
1.80
1.70


0.2793
1.0499
1.4081
0.5672
0.5105
n/a




0.7168
0.3193
1.2862
1.0023
0.3703
n/a




0.0756
1.6914
1.5713
1.5796
1.0390
n/a


3.20
3.40
3.00
3.30
3.80
4.70


DIGFWD = forward digit span; DIGBAK


backward digit span; DIGORD


digit


ordering; SD = standard deviation of mean; N = number of participants in sample; t
= two-tailed t-statistic for difference between participant and normative data; df=
degrees of freedom











APPENDIX A
INFORMED CONSENT



Info-rm Con sent Puripmie in Research

You are being asked to take part in a research study. Ths fr provides you with infonation about the
study. The Prinipal investigatorr (the persn in charge of this research or a repress q~etiu of the Principal
Inw.i gacr rill al)s describe this study to you and awer all of your questions. Before 3ou decide
whether or not to take part, read Lhe ini.Lormat.n below and ask questions about anything you do not
understand. Your pani-,c.pann is entirely volun a


1. Name of Particpant ( "Study Subject"?




2. Title of Research Study

Audiologic nd Non-A idiologic Sources of Variance in Speech Perception and Self-Reported
Co:mrunication Status in Adults


3. Principal] lavestigator and Telephoae Number(s)


Andrew John. B.A., Ph.D. Candidate
(352) 392-2046, extension 228


James W. Hall n11, PIhD., CCC-A
(352) 273-6181


Conaci mLormaruon flo emergencies after hours or on weekends or holidays:

Andrew John, B A. Ph.D. Candidate
(352) 336-1416


4 Sourne ofFusdig or Other Marenl .Support

Univerity ofFlorida


Approved by
Urn.Versitf of Florida
irlnsidiem a Rev Ie Board 02
Protocol # 2006-U-0754
For Use Through 09 15/2007


5. What is the purpose of this resrch study?

The purpose ol this study is to e. aiuslcue me of Lhe acetoIr contribuing TO the differences among
adults in the ability to perceive speech in difficult listening situations and mn ihoe -ulls individual
helicts about their own heating and i;orcmunirjctin oatliues Both adults with and without hearing
loss report difficulty commumcaing in certain situation For most, simply measuring hearing
sensitivity is insufficient to explain thow difficulties. Ofte individuals wih the same measured
hearing ability report very different listening and commuaniation experience and problems. This
study will examine some other factors that may contribute to these differences between individuals
using some measures thia are nc-Ti picall) ulilhjd L a hearing evaluation.

You ar being asked to participate in this study to help us determine how audioogists can better serve


Pa~ I of 4


134









individuals tend to overrate disability, consistent with the findings of Gordon-Salant and

colleagues (1994), who demonstrated that increased age correlated with increased disability as

reported on the HHIE.

Speech Perception

Failure of the standard audiometric battery to fully account for the everyday listening

difficulties reported by adults seeking treatment is likely due, at least in part, to the well-

established disconnect between peripheral hearing sensitivity, the primary focus of the typical

audiologic evaluation, and speech perception in noisy, reverberant, or otherwise challenging

listening situations. Indeed, the most common complaint from individuals with hearing loss

seeking audiologic care is difficulty understanding speech in noisy environments, such as busy

restaurants, parties, or large group conversations (CHABA, 1988; Kim, Frisina, Mapes,

Hickman, & Frisina, 2005; Koehnke & Besing, 2001). These auditory problems are reported

particularly widely by older adults, who have been shown to demonstrate significantly greater

difficulty understanding and recalling spoken messages compared to younger adults, particularly

in noisy and reverberant environments, even when younger and older adults with similar degrees

and configurations of hearing loss are compared (Pichora-Fuller & Souza, 2003; Souza, Yueh,

Sarubbi, & Loovis, 2000; Tun, O'Kane, & Wingfield, 2002). For older adults, understanding

speech in difficult environments is the primary consequence of hearing loss, and is thus likely to

be a primary influence in self-report of hearing handicap and self-referral for audiologic care.

Presumably, assessment of speech understanding in adverse listening conditions should be the

primary focus of the audiologic test battery.

However, speech perception in degraded listening conditions, such as in the presence of

noise, is not well predicted by common audiometric measures. Like self-reported hearing

impairment, speech-perception-in-noise ability correlates weakly with hearing sensitivity, with









WRSGIN (r = .41, p = .002); and BINGIN (r = .44, p = .001). However, when controlling for

BETAI, no significant associations were seen between temporal resolution variables and HINT

thresholds in quiet or in noise.

Few significant associations were found between HINT scores and working memory

variables. HINTQ was weakly but significantly correlated with DIGORD (r = -.31, p = .018)

and with WM (r = .29, p = .034), while HINTN was weakly correlated with DIGORD only (r =

.31, p = .019). When controlling for BETAI, no significant associations were seen between

working memory variables and HINT thresholds in quiet or in noise.

Because there were no significant partial correlations between any temporal resolution or

working memory measure and HINTQ or HINTN, it was expected that none of those variables

would contribute significantly to a multiple regression on either HINTQ or HINTN in the

presence of BETAI. Stepwise linear regression procedures were conducted separately for

HINTQ and for HINTN to confirm this expectation. The predictor variables included were

audibility (BETAI), temporal resolution (BINGIN), and working memory (WM). The bivariate

correlations between HINT scores and the various measured (RGIN, LGIN) and derived GIN

variables (BETGIN, WRSGIN, BINGIN) were small; therefore the model was simplified to

include only the average binaural GIN threshold.

Only BETAI was retained as a predictor variable in the regression equation for HINTQ

(R= .86, F1, 54 = 150.82, p < .001). Inclusion of BINGIN and WM increased the correlation

coefficient from .858 to .863, which was not a significant change (F2, 52 = .79, p = .459).

Similarly, only BETAI was retained in the regression equation for HINTN (R = .72, F1, 54 =

58.67, p < .001). Inclusion of temporal resolution and working memory variables increased the

correlation coefficient from .722 to .731, which was not a significant change (F2, 52 = .74, p =









"young" group aged 18 to 30 years to an "old" group aged 60 to 81 years. Snell (1997)

compared two groups with age ranges of 17 to 40 and 64 to 77 years. It is possible that age

effects are only detectable when comparing such disparate age groups, and that age effects are

minimized when data are compared within a group of older adults. Treating age as a continuous

variable within a narrow age range, as in the present investigation, may obscure these effects. In

all likelihood, age influences speech perception performance, particularly in difficult listening

situations.

Hearing Handicap

Not surprisingly, none of the predictor variables predicted self-reported hearing handicap

well. Both age and audibility were significantly, albeit weakly, correlated with HHIA/E total

score and Social/Situational subscale score, but not correlated with Emotional subscale score.

The associations among age and Social/Situational measures were quite weak, explaining only

11 to 24% of the variance in HHIA/E scores. Diminished hearing sensitivity (lower AI) was

associated with increased hearing handicap (higher HHIA/E score), as would be expected.

However, as with age, the relationship was weak, with audibility accounting for 12 to 20% of the

variance in hearing handicap score. In fact, there was no difference in HHIA/E Emotional

subscale or total score when comparing audiometrically normal participants with those having a

hearing loss. The absence of an association suggests that the emotional impact of hearing loss on

the individual is so variable that it can not even be distinguished between individuals with and

without clinically-significant hearing loss.

When controlling for hearing sensitivity, the correlation between HHIA/E score and age

disappeared, indicating that the association between objective and subjective measures of hearing

loss was similar for younger and older participants. Notably, the correlations between age and

HHIA/E score and between audibility and HHIA/E score were almost identical in strength. In









(age 70 to 89 years) was analyzed separately, binaural average GIN threshold was a significant

partial predictor of HINT threshold in the quiet condition only. That is, temporal resolution

ability explained an additional 12% of the variance in speech perception above the 62%

explained by hearing sensitivity.

The overall lack of association between HINT and GIN variables suggests that the GIN

does not significantly contribute to prediction of HINT threshold, except for a modest

explanatory contribution for the oldest participants examined. Because previous research

indicated an association between temporal resolution and speech perception in noise, a partial

correlation between GIN threshold and HINTN was anticipated in the present study. However,

other investigations also have found weak or no correlations between temporal processing and

speech perception when hearing sensitivity was controlled (Dubno & Dirks, 1990; Festen &

Plomp, 1983; Snell & Frisina, 2000; Strouse et al, 1998; van Rooij & Plomp, 1991). While it

seems likely that auditory temporal resolution ability is a factor for certain populations when

listening in difficult environments, the present investigation produced little evidence of such an

association.

As noted in the review of literature, previous attempts to find an association between

working memory and difficult speech perception have produced widely varying results (Gordon-

Salant & Fitzgibbons, 1997; Humes & Floyd, 2005; Lunner, 2003; Pichora-Fuller & Singh,

2006; Pichora-Fuller et al, 1995; Tun et al, 1991; Vaughan et al, 2006). A correlation between

working memory and HINT threshold was cautiously anticipated, but the conflicting findings in

the cited studies made this association far from certain. As with temporal resolution ability, the

present investigation produced no evidence of any role for working memory in speech

perception.









CHAPTER 3
METHODS

Participants

Fifty-six adults (17 male, 39 female) served as participants for this investigation.

Participants ranged in age from 50 to 89 years of age with a mean age of 66.4. All participants

were recruited from the Alachua county area by the investigator. To be eligible to participate in

this study, participants met the following inclusion criteria:

1. Age 50 or older at the time of testing;

2. Aware of a hearing problem that interferes with communication;

3. Generally good health, with no condition preventing the participant from
completing study measures; and

4. English as primary language as reported by the participant.

All inclusion criteria were satisfied by responses to questions from the investigator during

a case history interview prior to the beginning of data collection or, if possible, during

recruitment. Three potential participants who responded to advertisements for this investigation

subsequently revealed that they did not believe they had a hearing loss. These individuals were

excluded from this investigation, but were given a brief hearing test (pure-tone air-conduction

threshold testing). Participants were not enrolled in the study if they met any of the following

exclusion criteria:

1. Prior diagnosis of dementia or mild cognitive impairment, by history;

2. Transient hearing loss, as ascertained during the pre-test interview (i.e. "my
hearing loss comes and goes");

3. Conductive hearing loss, as ascertained using aural immittance testing
and, where indicated, bone-conduction audiometry;

4. Current use of hearing aids; and

5. Inability or unwillingness to complete all study measures.









a carrier phrase ("say the word") and are recorded to compact disc. Each participant received

one half-list of 25 words per ear, presented at the participant's monaural most comfortable

listening level (MCL). If the first ten words of a list (the ten most difficult words from that list

appear first) were repeated correctly, word recognition assessment was stopped and a word-

recognition score of 100% was assigned for the test ear. Word-recognition measurement was

conducted using CD recordings of the NU-6 stimuli ordered by difficulty.

Speech reception thresholds and word-recognition scores were obtained using a Grason-

Stadler GSI-61 audiometer with TDH-49 supra-aural headphones. The tests were conducted in a

single-walled sound-treated booth.

Speech Perception in Noise Testing: Hearing In Noise Test (HINT)

Assessment of speech perception in noise was conducted using the Hearing In Noise Test,

or HINT (Nilsson, Soli, & Sullivan, 1994) in quiet and noise conditions. The HINT consists of

25 lists of 10 phonemically-balanced sentences spoken by a male voice and recorded to CD with

speech spectrum noise matched to the long-term spectrum of the sentences serving as noise

competition. A typical HINT sentence is "The boy fell from the window." Sample HINT

sentence lists can be found in the appendix of Nilsson et al (1994).

Thresholds in noise were obtained using an adaptive procedure. The noise was held steady

at 65 dB SPL and the presentation level of the sentences was varied by the investigator based

upon whether the participant was able to repeat each sentence correctly and in full. Thresholds

in quiet were obtained using a similar procedure, but without the noise competition. For

analysis, these variables are denoted as HINTQ for the threshold in quiet and HINTN for the

threshold in noise.

The HINT was administered in a double-walled sound-treated booth using a five-speaker

array surrounding the seated participant. One speaker (Tannoy System 600), located at 0 degrees









word recognition, increases in reverberation caused a much greater performance decrease for

older adults with nearly normal hearing than for young individuals with normal hearing. These

two groups performed similarly in quiet, in noise alone, and in reverberation alone; the group

difference appeared when noise and reverberation were combined. Gordon-Salant and

Fitzgibbons (1993, 1995) also found that, on tests of speech stimuli degraded by reverberation

and time compression, both age and hearing loss were significant predictors, and that age became

a greater factor as the level of distortion increased. NabElek (1988) reported similar effects of

age on perception of vowels degraded by noise and reverberation.

In addition to environmental distortion, lengthy and more everyday-speech-like stimuli

appear to elicit age effects in speech perception. Jerger (1973) examined word recognition in a

retrospective clinical study of 2,162 patients and found that, when hearing sensitivity was

controlled, word recognition declines with age were consistent in individuals with moderate to

severe hearing loss, but not for individuals with mild hearing loss. Similarly, Bergman (1980)

found that for a given threshold level, older individuals performed more poorly on speech-

perception tasks than younger individuals. In another study by Jerger (1992), perception of

sentence stimuli was characterized by more robust age effects than perception of word stimuli, a

finding reported also by Jerger and Hayes (1977). In general, studies of aging and speech

perception show that increases in the difficulty of the speech-perception task (such as longer

stimuli and more adverse listening conditions) tend to elicit more robust age effects (Gordon-

Salant & Fitzgibbons, 1993, 1995; Harris & Reitz, 1985; Jerger, Jerger, & Pirozollo, 1991;

Weinstein, 2000).

The investigation of auditory function for young versus elderly listeners presents

significant methodologic challenges. Older adults with thresholds comparable to young adults









Cameron, S., Dillon, H., & Newall, P. (2006). Development and evaluation of the listening in
spatialized noise test. Ear and Hearing. 27(1), 30-42.

Carhart, R. & Jerger, J. (1959). Preferred method for clinical determination of pure-tone
thresholds. Journal of Speech and Hearing Disorders, 16, 340-345.

Carpenter, P. A., Miyake, A., & Just, M. A. (1994). Working memory constraints in
comprehension: Evidence from individual differences, aphasia, and aging. In M.A. Gernsbacher
(Ed.), The Handbook ofPsycholinguistics. San Diego, CA: Academic Press.

Chermak, G. D. & Lee, J. (2005). Comparison of children's performance on four tests of
temporal resolution. Journal of the American Academy ofAudiology, 16, 554-563.

Clark, L. & Knowles, J. (1973). Age differences in dichotic listening performance. Journal of
Gerontology, 28, 173-178.

Cohen, G. (1987). Speech comprehension in the elderly: The effects of cognitive changes.
British Journal ofAudiology, 21, 221-226.

Committee on Hearing, Bioacoustics and Biomechanics (CHABA). (1988). Speech
understanding and aging. Journal of the Acoustical Society ofAmerica, 83, 859-920.

Cooper, J. A., Sager, H. J., Jordan, N., Harvey, N. S., & Sullivan, E. V. (1991). Cognitive
impairment in early, untreated Parkinson's disease and its relationship to motor disability. Brain,
114, 2095-2122.

Crandell, C. (1991). Individual differences in speech recognition ability: Implications for
hearing aid selection. Ear and Hearing 12 (Suppl 6), 100s-107s.

Crandell, C. & Needleman, A. (1999). Modeling hearing loss via masking: Implications for
hearing aid selection. The Hearing Journal, 52(11), 58-62.

Crandell, C. & Smaldino, J. (2002). Room acoustics and auditory rehabilitation technology. In
J. Katz (Ed.), Handbook of Clinical Audiology, Fifth Edition (pp. 607-630). Baltimore:
Lippincott Williams & Wilkins.

Danhauer, J. L. & Johnson, C. E. (1991). Perceptual features for normal listeners' phoneme
recognition in a reverberant lecture hall. Journal of the American Academy ofAudiology, 2, 91-
98.

DeDe, G., Caplan, D., Kemtes, K. & Waters, G. (2004). The relationship between age, verbal
working memory, and language comprehension. Psychology andAging'. 19, 601-616.

Divenyi, P. L. & Danner, W. F. (1977). Discrimination of time intervals marked by brief
acoustic pulses of various intensities and spectra. Perception andPsychophysics, 21(2), 125-142.










Stephens, D. & Hetu, R. (1991). Impairment, disability, and handicap in audiology: Towards a
consensus. Audiology, 30, 185-200.

Strayer, D. L., Wichkens, C. D., & Braune, R. (1987). Adult age differences in the speed and
capacity of information processing: II. An electrophysiological approach. Psychology and
Aging. 2, 99-110.

Strouse, A., Ashmead, D. H., Ohde, R. N., & Grantham, D. W. (1998). Temporal processing in
the aging auditory system. Journal of the Acoustical Society of America, 104, 2385-2399.

The Psychological Corporation (1997). Wechsler Adult Intelligence Scale Third Edition
(WAIS-III). San Antonio, TX.

Tillman, T. W. & Carhart, R. (1966). An expanded test for speech discrimination utilizing CNC
monosyllabic words. Northwestern University Auditory Test No. 6, Technical Report, SAM-
TDR-62-135. Brooks Air Force Base, TX: USAF School of Aerospace Medicine, Aerospace
Medical Division.

Trehub, S., Schneider, B., & Henderson, J. (1995). Gap detection in infants, children, and adults.
Journal of the Acoustical Society ofAmerica, 98, 2532-2541.

Tun, P., O'Kane, A. G., & Wingfield, A. (2002). Distraction by competing speech in young and
older adult listeners. Psychology andAging,. 17, 453-467.

Tun, P., Wingfield, A., & Stine, E. L. (1991). Speech-processing capacity in young and older
adults: A dual-task study. Psychology and Aging. 6(1), 3-9.

Turner, C. W. & Henry, B. A (2002). Benefits of amplification for speech recognition in
background noise. Journal of the Acoustical Society ofAmerica, 112(4), 1675-1680.

Tyler, R. S., Summerfield, Q., Wood, E. J., & Fernandes, M. A. (1982). Psychoacoustic and
phonetic temporal processing in normal and hearing-impaired listeners. Journal of the
Acoustical Society ofAmerica, 72, 740-752.

Uhlmann, R. F., Larson, E. B., Rees, T. S., Koepsell, T. D., and Duckert, L. G. (1989).
Relationship of hearing impairment to dementia and cognitive dysfunction in older adults.
Journal of the American Medical Association, 261(13), 1916-1919.

van Boxtel, M. P. J., van Beijsterveldt, T., Houx, P. J., Anteunis, L. J. C., Metsemakers, J. F. M.,
& Jolles, J. (2000). Mild hearing impairment can reduce verbal memory performance in a
healthy adult population. Journal of Clinical and Experimental Neuropsychology, 22(1), 147-
154.











. ... . . . . . . . . . . . . . . . . . ... . . i . .



........ .......................... ...... ......... ............... ........................... ............................ .................... .: ......... ........







........ . . . . . ...................................................... . . . i .


...................... ............. .............. ............. ......... ............ i .... ...



........ ............................ ........................... ........................... ........................... ............................. ............... . . .




125... 2. 5. .
. -.. . -.. . -.. . -.. . a.. . . . .


125 250 500 1k 2k
Frequency

Figure 4-9. Mean audiogram for HL Group 2 (BETAI < .98) (N


4k Bk



32).









measures of hearing including hearing sensitivity (Kannan and Lipscomb, 1974), otoacoustic

emissions (Khalfa, Morlet, Michey, Morgon, & Collet, 1997), and dichotic speech perception

(Bellis & Wilber, 2001; Jerger et al, 1994; Musiek, Wilson, & Pinheiro, 1979). One explanation

offered for auditory asymmetry is noise exposure. Examples of unbalanced noise exposures

include rifle shooting, which, for right-handed individuals, places the left ear closer to the barrel

than the right ear, or driving a car with the window down, where the left ear is closer to the

window and thus more susceptible to wind noise. However, asymmetry in auditory abilities

within and beyond the peripheral hearing system can not always be explained by noise exposure.

Numerous studies have identified a left-ear disadvantage in dichotic listening that increases with

age even when controlling for changes in hearing sensitivity (Bellis & Wilber, 2001; Clark &

Knowles, 1973; Jerger & Jordan, 1992; Jerger et al, 1990; Jerger et al, 1994; Johnson, Cole,

Bowers, Foiles, Nikaido, Patrick, & Wolliver, 1979).

Auditory system asymmetry is likely related to physiological changes with age that affect

the right and left auditory structures and pathways unequally. One hypothesis to explain

auditory asymmetry is that the right ear's connection to the left hemisphere of the brain, where

for most individuals the centers for language and verbal processing are found, provides that ear

with privileged access to speech-perception resources (Jerger et al, 1994). Moreover, neuro-

imaging studies show greater atrophy in the right cerebral hemisphere (to which the left ear has

superior access) compared to the left hemisphere (to which the right ear has access) (Goldstein &

Shelly, 1981; Levy-Agresti & Sperry, 1968).

Deterioration in function of the corpus callosum in older adults may also affect auditory

pathways by reducing inter-hemispheric communication (Duffy, McAnulty, & Albert, 1996;

Goldstein & Braun, 1974; Hellige, 1993; Jerger et al, 1994). This deterioration can reduce the









a gap detection task; nor working memory, as measured by a digit span and ordering

battery, will be correlated significantly with speech perception ability in quiet when

controlling for audibility.

HYPOTHESIS lb: It is hypothesized that audibility, as measured by peripheral hearing

sensitivity; temporal resolution, as measured by a gap detection task; and working

memory, as measured by a digit span and ordering battery, will be significantly correlated

with speech perception ability in noise.

AIM 2: The investigator will determine whether the well-established discrepancy between

objectively evaluated hearing impairment and subjective, self-reported hearing handicap

can be partially explained by the supra-threshold distortions described above and their

impact on listening in difficult environments.

HYPOTHESIS 2: It is hypothesized that audibility, as measured by peripheral hearing

sensitivity; temporal resolution, as measured by a gap detection task; and working

memory, as measured by a digit span and ordering battery, will be significantly correlated

with self-reported hearing handicap, as measured by composite score on the Hearing

Handicap Inventory for Adults (HHIA) or the Hearing Handicap Inventory for the

Elderly (HHIE).

AIM 3: In order to evaluate the efficacy of the experimental measures in a clinical setting, the

investigator will determine the degree to which temporal resolution threshold and/or

working memory capacity is predictive of speech perception difficulty in noise and/or

self-reported hearing handicap above the variance which is predicted by audibility.

HYPOTHESIS 3: It is hypothesized that, when controlling for audibility, temporal resolution









APPENDIX B
GAPS IN NOISE TEST









Auditory gap detection measures test the ability of a listener to detect the presence of a

temporal gap (silent period created by an absence of energy) within a tone or, more commonly,

noise burst. Performance on a gap detection measure is assumed to be related to speech

perception because silent gaps in speech signals can serve as phonetic cues. For example, a

silent gap is a cue for the presence of the stop consonant phoneme /p/ in the word "spoon," in

contrast to the word "soon" in which there is no gap. Because psychoacoustic gap detection

thresholds measure a listener's ability to perceive the shortest detectable gap in a sound signal

and the detection of such gaps is relevant to speech perception, it is logical to assume that a

strong relationship exists between measures of gap detection threshold and speech perception

ability (Pichora-Fuller, Schneider, Benson, Hamstra, & Storzer, 2006). There is experimental

evidence of significant correlations between measures of gap detection in noise or tone stimuli

and word identification in noise (Snell et al, 2002; Tyler et al, 1982) and reverberation (Gordon-

Salant & Fitzgibbons, 1993).

Gap detection ability is generally poorly predicted by the audiogram, as individuals with

similar degrees and configurations of hearing impairment often show widely varying ability to

perceive temporal gaps (Glasberg, Moore, & Bacon, 1987; Moore & Glasberg, 1988; Schneider,

Pichora-Fuller, Kowalchuk, & Lamb, 1994; Schneider, Speranza, & Pichora-Fuller, 1998).

Several studies have also found gap detection ability to be independent of frequency selectivity

of auditory filters (Eddins, Hall, & Grose, 1992; Grose, Eddins, & Hall, 1989; Moore, Peters, &

Glasberg, 1993).

The correlation between gap detection threshold and speech perception is not clear, even

when accounting for hearing sensitivity and spectral resolution. While several studies have

noted an association between temporal resolution ability when controlling for audibility









particularly low relations seen for individuals with more severe degrees of hearing loss

(Crandell, 1991; Plomp & Mimpen, 1979; Smoorenburg, 1992). Some studies have shown that

hearing sensitivity accounts for about half of the variance in speech perception in noise

(Middelweerd, Festen, & Plomp, 1990; Plomp, 1986; Plomp & Mimpen, 1979; Smoorenburg,

1992). The balance of this variance is not, however, explained well by tests in the classic clinical

battery. Tests of speech perception in quiet, commonly used as diagnostic tools, are poor

predictors of speech perception in noise. Furthermore, speech perception scores in quiet do not

add substantially to the predictive value of the audiogram-a finding that is explained by the

high correlation usually found between speech perception in quiet and sensitivity measured using

pure tones (Divenyi & Haupt, 1997; Festen & Plomp, 1986; Glasberg & Moore, 1989; Plomp,

1986; Plomp & Mimpen, 1979; Smoorenburg, 1992; van Rooij & Plomp, 1990).

Some researchers have suggested that speech-perception deficits in individuals with SNHL

can be explained by portions of the speech signal being rendered inaudible by the decreased

sensitivity that is a consequence of SNHL (Fabry & Van Tasell, 1986; Humes, Dirks, Bell, &

Kincaid 1987; Humes, Espinoza-Varas & Watson, 1988). The results of selected studies

suggested that increasing the audibility of the most difficult to perceive parts of speech

eliminates deficits in speech perception. Humes et al (1994) compared the performance of 50

listeners between the ages of 63 and 83 on a number of speech-recognition, auditory processing,

and cognitive tasks. In this study, hearing sensitivity accounted for 70 to 75% of the total

variance in speech-recognition performance. Auditory processing and cognitive function

accounted for little or no additional variance.

A study by Crandell and Needleman (1999) also found that individuals with normal

hearing who were given a simulated hearing loss demonstrated similar speech-recognition


















Thst 1 1I2/2001
LocaUtan ( r) Duratilon (ml)
1 1337.3 1S
3200.3 2
5277.1 5

ta 131. 15

1 262.4 6
4491,9 Io

4 1145.4 6
3449.6 20
4319.3 6

5 4466.0 4

6 139.5 12

7 2799.7 3
3421.8 4

a 1757-1 10
2875-5 t1

9 263-.4 5


11 2727-5
42105-0
5011-1

11 4014.1

13 2304.8

t4 1597,2

is 2032.1
4564.7

16 1000.B
2613.4
4190,7
17


7 1268..
1971.2

19 i193.7


Location (ms) 0-uratpon (rmi
20 726.3 2

21 4595 5

22 4024.6 8
5174.2 20

Z3 50.5 1i
4537.5 10

24 2196,3 5

25 20066. 20
3349.4 2

26 1520.3 3
5491,9 2

27 195S.9 5
3194,0 15

18 1056.3 2
3190.6 20
4358.1 8

29 1338.3 3
3802 5 4

30 884.3
2150.3 15
3386.4 20

31 4199.3 4

32 3047.4 4
5322.9 10

33 1812-0 15
Z793,5 B

34 15644 8
2255.5 6

35 11 L1.5 12
2613.0 12









at 3000 and 4000 Hz abnormal. For these four participants, DPOAE data were consistent with

audiometric data. The mismatch in this classification is an artifact of the grouping scheme. In

other words, if the mid-frequency range were further separated into low-mid and high-mid

ranges and audiometric data were compared, the low-mid range (approximately 1000 to 2000

Hz) would be normal both audiometrically and in DPOAE measurement, while the high-mid

range (approximately 2000 to 4000 Hz) would be abnormal in both respects.

Figure 4-1 provides a breakdown by ear of DPOAE status as a function of test frequency

range (low, mid, and high) for this participant group. Declines in cochlear status particularly

affecting the high-frequency area of the cochlea are evident bilaterally. As detailed in the

methods chapter, low-frequency DPOAE measurement was confounded to some degree by high

ambient noise levels in the test room for the first sixteen participants in this investigation. These

participants were tested prior to the installation of an extension cable allowing DPOAE

measurement to be conducted within the sound booth. As a result, low-frequency DPOAEs may

not be a reliable measure in this investigation. It is likely that some participants were classified

as having absent low-frequency DPOAEs when responses were present but could not be

measured due to the elevated noise floor. However, examination of higher-frequency DPOAE

data reveal declines from the mid- to high-frequency DPOAE test ranges. These declines are

consistent with the common findings of increased loss of cochlear function and high-frequency

hearing loss with age. Furthermore, comparison of DPOAE data to audiometric thresholds in

Table 4-1 suggests that more cochlear dysfunction was present within the participant group than

was apparent in pure-tone threshold testing.

Audibility

The composite audiogram for the fifty-six participants in this investigation revealed a

moderate sloping high-frequency sensorineural hearing loss bilaterally (Figure 4-2). Standard


















125 250 500 1k
Frequency


2k 4k Bk


Figure 4-8. Mean audiogram for HL Group 1 (BETAI > .98) (N = 24).


........ ........................ ............................ .................. ......... ........

. . . . . . . . . . . . . . . . . . . . . . .

........ ........................... ........................... ........................... ............................ ............................ .................. ......... ........
. . . . . . . . . . . . . . . . . . . . . . . .

........ ........................... ........................... ........................... ........................... ......... ........
............................ ...................









All exclusion criteria were satisfied by responses to questions from the investigator during

a case history interview prior to the beginning of data collection or, if possible, during the

recruitment process. Prior use of hearing aids, defined as a period of hearing aid use ending at

least one year prior to recruitment, was not considered an exclusion criterion. Four participants

reported prior use of hearing aids, but none had used them within the last five years.

Three potential participants were excluded from this investigation on the basis of presence

of conductive hearing loss. These participants were recommended to an otolaryngologist for

evaluation and possible treatment.

Prior to participating in the study, each participant was required to read carefully and sign

an Informed Consent document approved by the University of Florida Institutional Review

Board (IRB-02). This investigation was approved in September 2006 as UF protocol 2006-U-

0754. The informed consent form can be found in Appendix A.

Cognitive Screening: Mini-Mental State Examination

In order to rule out the presence of dementia, each participant completed prior to data

collection a cognitive screening instrument, the Mini-Mental State Examination, or MMSE

(Folstein, Folstein, & McHugh, 1975). The MMSE, a brief 30-item questionnaire commonly

used in health research to screen for dementia, was administered by the investigator, face-to-face.

The MMSE assesses various cognitive domains including arithmetic, memory, and orientation,

and yields a total score calculated on a 30-point scale. For the current investigation, the criterion

for exclusion from participation was a score of 26 or less. However, no participant scored below

a 27 on the MMSE.

Hearing Evaluation

Each participant underwent a comprehensive hearing evaluation, including the following

procedures: tympanometry, acoustic reflex screening, distortion-product otoacoustic emissions









Because the participant sample from the present investigation spans almost four decades

within the adult age range (50 to 89), it might be useful to examine age differences within the

sample to identify possible age differences in associations among the variables studied. For this

analysis, the participant sample was divided into three groups by age.

Group 1 (late middle-aged adults) was composed of participants age 50 to 59. This group

had 17 total members (6 male, 11 female) with a mean age of 56.0 years (SD = 3.1) (Figure 4-5).

Group 2 (young-old adults) was composed of participants age 60 to 69. This group had 19 total

members (5 male, 14 female) with a mean age of 64.7 years (SD = 2.7) (Figure 4-6). Group 3

(old-old adults) was composed of participants age 70 to 89. This group had 20 total members (6

male, 14 female) with a mean age of 76.9 years (SD = 4.8) (Figure 4-7). Mean audiometric data

for each age group is displayed in Table 4-10. A comparison of the three groups on the

experimental measures and dependent variables can be seen in Table 4-11.

An ANOVA was conducted to test for group differences on each of the experimental

measures and dependent variables (Table 4-12). Group 1 (late middle-aged adults) had

significantly higher audibility scores compared to Group 2 (young-old adults) as measured by

RAI, WORSAI, and BINAI (p < .05); significantly higher DIGORD scores (p < .05);

significantly shorter GIN thresholds as measured by LGIN and WRSGIN (p < .05); and

significantly lower HINTQ thresholds (p < .05). No significant differences were seen between

Groups 1 and 2 for LAI, BETAI, DIGFWD, DIGBAK, WM, RGIN, BETGIN, BINGIN,

HINTN, HHIEMOT, HHISOC, or HHITOT.

Group 1 had significantly higher audibility scores compared to Group 3 (old-old adults) on

all AI measures (p < .001); significantly higher working memory scores as measured by

DIGFWD (p < .05), DIGBAK (p < .05), DIGORD (p < .01), and WM (p < .01); significantly









HINT in Noise Condition

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HINTN; therefore, no regression model was computed.

For Group 2, only BETAI was retained as a predictor variable in the regression equation

for HINTN (R = .78, F1, 30 = 45.27, p < .001). Inclusion of BINGIN and WM increased the

correlation coefficient from .776 to .804, which was not a significant change (F2, 28 = 1.76, p =

.190).

HHIA/E Emotional Subscale

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HHIEMOT; therefore, no regression model was computed.

For Group 2, only BINGIN was retained as a predictor variable in the regression equation

for HHIEMOT (R = .40, F1, 30 = 5.71, df = 1, 30, p = .023). Inclusion of BETAI and WM

increased the correlation coefficient from .400 to .451, which was not a significant change (F2, 28

= .76, p = .476).

HHIA/E Social/Situational Subscale

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HHISOC; therefore, no regression model was computed.

For Group 2, only BETAI was retained as a predictor variable in the regression equation

for HHISOC (R = .36, F1, 30 = 4.43, p = .044). Inclusion of BINGIN and WM increased the

correlation coefficient from .359 to .502, which was not a significant change (F2, 28 = 2.31, p =

.118).

HHIA/E Total

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HHITOT; therefore, no regression model was computed.









reverse order. Scores ranged from 3 to 13. The maximum possible score on backward digit span

is 14.

Digit ordering (DIGORD) scores averaged 16.4 (SD = 4.0). A score of 16 approximates

successful repetition of a maximum five-digit string, placing the digits in numerical order.

Scores ranged from 3 to 24. The maximum possible score on digit ordering is 24.

Working memory test scores were compared to normative data collected by the University

of Florida Language over the Lifespan laboratory on community-dwelling older adults in the

Alachua County area. Data were available from this laboratory for age groups 60-64 (n=12), 65-

69 (n=32), 70-74 (n=48), 75-79 (n=58), 80-84 (n=41), and 85 or older (n=17). No data for

individuals between age 50 and 59 years were available. Table 4-5 provides comparisons

between the participant group in the present investigation and the normative group within these

age ranges.

Participants age 70 and up had mean scores slightly lower than normative means on all

working memory measures. However, paired-samples T-tests on all measures confirmed that the

participant group did not differ significantly from the normative group on any of the working

memory tests in any age group examined (a = .05). Only one participant in the present

investigation was over age 85. This participant scored more than two standard deviations below

the mean of the normative group on the digit ordering task. In general, comparison of data from

the present investigation to the local normative data suggests that the participant group had

working memory status within the normal range for community-dwelling adults in the

geographical area from which the group was sampled.

Moderate collinearity was calculated among all working memory measures (DIGFWD x

DIGBAK: r = .49, p < .001; DIGFWD x DIGORD: r = .37, p = .006; DIGBAK x DIGORD: r









Grose, J. H., Hall, J. W., & Buss, E. (2001). Gap duration discrimination in listeners with
cochlear hearing loss: Effects of gap and marker duration, frequency separation, and mode of
presentation. Journal of the Associationfor Research in Otolaryngology, 2(4), 388-398.

Grose, J. H., Hall, J. W., & Buss, E. (2006). Temporal processing deficits in the pre-senescent
auditory system. Journal of the Acoustical Society ofAmerica, 119(4), 2305-2315.

Grose, J. H., Hall, J. W., Buss, E., & Hatch, D. (2001). Gap detection for similar and dissimilar
gap markers. Journal of the Acoustical Society ofAmerica, 109(4), 1587-1595.

Gustafsson, H. A. & Arlinger, S. D. (1994). Masking of speech by amplitude-modulated noise.
Journal of the Acoustical Society ofAmerica, 95, 518-529.

Haggard, M. P. & Hall, J. W. (1982). Forms of binaural summation and the implications of
individual variability for binaural hearing aids. Scandanavian Audiology, 15, 47-63.

Hall, J. W. (2000). Handbook of Otoacoustic Emissions. San Diego: Singular Publishing
Group.

Halling, D. & Humes, L. (2000). Factors affecting the recognition of reverberant speech by
elderly listeners. Journal of Speech, Language, and Hearing Research, 43(2), 414-431.

Hargus, S. E. & Gordon-Salant, S. (1995). Accuracy of speech intelligibility index predictions
for noise-masked young listeners with normal hearing and for elderly listeners with hearing
impairment. Journal of Speech and Hearing Research, 38, 234-243.

Harris, R. W. & Reitz, M. L. (1985). Effects of room reverberation and noise on speech
discrimination by the elderly. Audiology, 24(5), 319-324.

Haubert, N., & Pichora-Fuller, M. K. (1999). The perception of spoken language by elderly
listeners: Contributions of auditory temporal processes. Canadian Acoustics, 27, 96-97.

Hawkins, D., Prosek, R., Walden, B., & Montgomery, A. (1987). Binaural loudness summation
in the hearing impaired. Journal of Speech and Hearing Research, 30(1), 37-43.

Hawkins, D. & Yacullo, W. (1984). Signal-to-noise advantage of binaural hearing aids and
directional microphones under different levels of reverberation. Journal of Speech andHearing
Disorders, 49, 278-286.

He, N., Horwitz, A., Dubno, J., & Mills, J. (1999). Psychometric functions for gap detection in
noise measured from young and aged subjects. Journal of the Acoustical Society ofAmerica,
106(2), 966-978.

Heinrich, A. & Schneider, B. (2006). Age-related changes in within- and between-channel gap
detection using sinusoidal stimuli. Journal of the Acoustical Society ofAmerica, 119(4), 2316-
2326.









For Group 3 (age 70 89), both BETAI and BINGIN were retained as predictor variables

in the regression equation for HINTQ (R = .86, F2, 17 = 23.99, p < .001). Inclusion ofWM

increased the correlation coefficient from .859 to .863, which was not a significant change (F1, 16

=.35, p =.565).

HINT in Noise Condition

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HINTN; therefore, no regression model was computed.

For Group 2, only BETAI was retained as a predictor variable in the regression equation

for HINTN (R = .47, F1, 17 = 4.74, p = .044). Inclusion of BINGIN and WM increased the

correlation coefficient from .467 to .469, which was not a significant change (F2, 15 = .02, p =

.982).

For Group 3, only BETAI was retained as a predictor variable in the regression equation

for HINTN (R = .90, F1, is = 77.76, p < .001). Inclusion of BINGIN and WM increased the

correlation coefficient from .901 to .911, which was not a significant change (F2, 16 = .83, p =

.452).

HHIA/E Emotional Subscale

For Group 1, no significant correlation was seen between any predictor variable (BETAI,

BINGIN, WM) and HHIEMOT; therefore, no regression model was computed.

For Group 2, only BINGIN was retained as a predictor variable in the regression equation

for HHIEMOT (R = .50, F1, 17 = 5.54, p = .031). Inclusion of BETAI and WM increased the

correlation coefficient from .496 to .532, which was not a significant change (F2, 15 = .39, p =

.681).

For Group 3, only BINGIN was retained as a predictor variable in the regression equation

for HHIEMOT (R = .45, F1, 18 = 4.62, p = .046). Inclusion of BETAI and WM increased the

































To Dr. Carl Crandell









correlation matrix in Table 4-8. Correlation coefficients for HHIA/E scales ranged from .41 to

.45 for RGIN, .30 to .32 for LGIN, .34 to .39 for BETGIN, .34 to .38 for WRSGIN, and .38 to

.41 for BINGIN. When controlling for audibility using BETAI, HHIEMOT was significantly

correlated with RGIN (r= .34, p = .010), WRSGIN (r = .29, p = .034), and BINGIN (r = .31, p =

.023). Significant correlations were also calculated between HHITOT and RGIN (r = .34, p =

.012) and between HHITOT and BINGIN (r = .28, p = .037).

There was only one significant association between HHIA/E scores and working memory

variables. Social/Situational HHI/E subscale score was weakly correlated with DIGORD (r = -

.27, p =.048).

Stepwise linear regression procedures were conducted separately for HHIEMOT,

HHISOC, and HHITOT. The predictor variables included were audibility (BETAI), temporal

resolution (BINGIN), and working memory (WM). This procedure is identical to the procedure

described above which regressed the explanatory variables on HINTQ and HINTN.

Only the temporal resolution factor BINGIN was retained as a factor in the regression

equation for HHIEMOT (R = .38, F1, 54 = 8.85, p = .004). Inclusion of BETAI and WM

increased the correlation coefficient from .375 to .380, which was not a significant change (F2, 52

=. 11, p = .895). The audibility factor BETAI was retained as a factor in the regression equation

for HHISOC (R = .45, F1, 54 = 13.56, p = .001). Inclusion of BINGIN and WM increased the

correlation coefficient from .448 to .506, which was not a significant change (F2, 52 = 1.94, p =

.155). Only the temporal resolution factor BINGIN was retained as a factor in the regression

equation for HHITOT (R = .41, F1, 54 = 11.19, p = .002). Inclusion of BETAI and WM increased

the correlation coefficient from .414 to .453, an insignificant change (F2, 52 = 1.09, p = .345).












Auh..nzed perron fmrom the Uniersiry of Flodrida the hospital or clinic (ifany) involved in this
research, and the Institutional Review Board ha. e the legal right to review your research records and
will protect the confideviality of them to the extent pennitted by law. Otherwise, yow research records
will not be released with your consent unless required by law ora cart order.

f the results of this research are published or presented at scientific meettins, you identity will not be
JscIlosad.

15. How will the researchers benefit from your being in this study?

In a enerai. presenting research results helps the career of a scientist. Therfore, the principal
investigator may tune t i' Lt r-ults ol' Lhi study are presented at sci-nTiic meeaunj s r in scientific
journals,


16 Signatures

As a reprEsentative of this study, I have explained to the panikipanr the purpoie, the procedures, the
possible hbrfit, and fth risks of this resea h study; the altcmaric to being in the jr dy. and how
pri-act ? %*ill be putected:




Sigr lure of Person C*Oininig Consent Date





You have been informed about this study'spurpose prcedum' possible benefits, and risks; the
altenives to being in the study; and how your privacy will tbe protected. You have received a copy of
this Focn. You have been given dhe opportunity to ask questions before you sign, and you have been
told that you can ask other questions at any time.

You wolurnanl agree to partcipale this study. By ; gning this form, you arc not waiving any of your
legal rights.


Signatre of Person Consenting


Date


Appo.d. bDy
Un.vrOtly Of FliOda
Irstmiuliona Review Board 02
Proj Uoco # 20 1J-07A
For Use Thrugn 9i 1512007


Page 4 of4









A significant positive correlation was seen between age and HHISOC (r = .49, p < .001)

and between age and HHITOT (r = .34, p =.01), but not between age and HHIEMOT, indicating

that perceived social and situational impact of hearing loss, but not the emotional consequences

of that hearing loss, increased slightly with age. When controlling for BETAI, no significant

correlation was seen between age and HHIA/E scores, indicating that age did not contribute

significant predictive power for HHIA/E score when controlling for hearing sensitivity.

Because listening in difficult, noisy environments is the principal complaint of most

individuals seeking help for a hearing problem, it seems logical to assume that self-report

measures of hearing handicap would correlate with speech perception ability in noise. To test

this assumption, bivariate correlations were calculated between each pair of speech perception

measures (HINTQ and HINTN) and HHIA/E scales (HHIEMOT, HHISOC, and HHITOT). In

addition, the correlations among these variables were compared to the independent correlations

between each variable and BETAI to evaluate the relative strength of associations. A correlation

matrix for these variables is displayed in Table 4-7.

Significant correlations were seen between HINTQ and HHISOC (r = .37, p = .004) and

between HINTQ and HHITOT (r = .30, p = .024). Performance on the HINT in quiet was not

significantly correlated with HHIEMOT (p = .128). Correlations were also seen between

HINTN and HHIEMOT (r = .33, p = .012), between HINTN and HHISOC (r = .45, p < .001),

and between HINTN and HHITOT (r = .42, p = .001). Overall, HHIA/E scores were better

predicted by HINTN (r = .33, .45, and .42 for HHIEMOT, HHISOC, and HHITOT respectively)

than by HINTQ (r = .21, .38, and .30) or by BETAI (r = -.23, -.45, and -.35).

There were significant correlations among all temporal resolution measures (RGIN, LGIN,

BETGIN, WRSGIN, and BINGIN) and all HHIA/E subscale and total scores as displayed in the














- I
-- --- - -- -- -- -- --- --- -- -- -


125 250 500


1k 2k 4k Ok


Frequency

Figure 4-5. Mean audiogram for participants age 50-59 (N = 17).









recruitment, and loss of spectral resolution that results from SNHL (Crandell & Needleman,

1999; Humes, 1996; Humes & Jesteadt, 1991; Humes et al, 1988). Crandell and Needleman

(1999) examined sentence recognition in a multitalker babble background for individuals with

SNHL and for matched individuals with normal hearing who were given a simulated noise-

masked hearing loss. The simulated hearing loss was created by presenting spectrally-shaped

masking noise. When noise was present, the listeners with normal hearing demonstrated hearing

thresholds similar to matched listeners with SNHL. The individuals with simulated hearing loss

performed significantly better than individuals with actual SNHL, suggesting that there was an

effect of SNHL on speech perception that could not be simulated simply by raising threshold.

However, an earlier study by Humes and Roberts (1990) found that older adults with

presbyacusis and young adults with matched simulated hearing loss performed similarly in tests

of nonsense syllable recognition in noisy and reverberant listening conditions. It is notable that,

unlike the later study by Crandell and Needleman (1999) cited above, Humes and Roberts (1990)

used closed-set testing of nonsense syllable recognition, whereas Crandell and Needleman used

an open-set protocol requiring repetition of sentences, a more complex stimulus. The

discrepancy in methodology and in findings between the two studies supports the contention that

there is a component of suprathreshold distortion in SNHL for older adults that can not be

explained by the audibility of the signal, and that the suprathreshold distortion is most evident

with stimuli similar to those found in everyday listening situations (Crandell & Needleman,

1999; Plomp & Mimpen, 1979).

Age is a factor in speech perception ability in degraded listening conditions (Frisina &

Frisina, 1997; Gordon-Salant & Fitzgibbons, 1993, 1995; Harris & Reitz, 1985; NabElek, 1988;

Pichora-Fuller, Schneider & Daneman, 1995). Harris and Reitz (1985) found that, in tests of










12.0

10.0

8.0
S6.0
Left
4.0

2.0

0.0
Participant Group Normative Group

Figure 4-4. Comparison of participant mean Gaps-In-Noise threshold values (with standard
deviation bars) to normative data collected by Musiek et al (2005) on 50 subjects (age
13 to 46 years).









noise (HINTN) and quiet (HINTQ) were calculated by averaging the presentation level of HINT

sentences five through twenty-one and adding a 1-dB correction factor for the conversion to dB

SPL. A signal-to-noise ratio (SNR) for threshold in noise was calculated by subtracting 65 (the

intensity level of the competition noise in dB SPL) from the corrected HINT threshold in noise.

The average speech perception threshold in the quiet listening condition was 38.7 dB SPL

(SD = 9.3). Thresholds ranged from 26.1 to 63.5 dB SPL, indicating large variance within the

participant group in ability to perceive spoken speech in quiet environments. As expected,

HINTQ showed a strong and significant negative correlation with BETAI (r = -.86, p <.001)

indicating that the majority (about 73%) of the variance in speech perception in quiet is

explained by loss of hearing sensitivity. With hearing sensitivity controlled, age was not a

significant predictor of speech perception performance in quiet.

The average speech perception threshold in the noise condition was 72.2 dB SPL (SD =

2.6). Thresholds ranged from 68.7 dB to 81.7 dB SPL. That is, participants perceived the

sentence stimuli correctly at an average SNR of 7.2 and in a range from 3.7 to 16.7 dB.

A strong and significant negative correlation was seen between HINTN and BETAI (r = -.72, p <

.001), indicating that hearing sensitivity accounted for about 52% of the variance in speech

perception in noise. With hearing sensitivity controlled, age was not a significant predictor of

speech perception performance in noise.

Both HINTQ and HINTN were positively correlated with temporal resolution threshold

when audibility was not controlled. Quiet-condition HINT thresholds were significantly

correlated with RGIN (r = .36, p = .006); BETGIN (r = .32, p = .018); WRSGIN (r = .28, p =

.040); and BINGIN (r = .31, p = .019). Noise-condition HINT thresholds were significantly

correlated with RGIN (r = .42, p = .001); LGIN (r = .39, p = .003); BETGIN (r = .42, p = .001);









expressed as a number between 0 and 1.0, where 0 indicates that no speech information is

received and 1.0 indicates that all speech information is received. The AI is typically calculated

by dividing the frequency spectrum into several bands, and then weighting each band between

250 and 4000 Hz based on the importance of that frequency region for understanding speech in

quiet environments. That is, unlike a simple mean of thresholds at speech frequencies, the AI

weights audiometric frequency bands unequally to emphasize the frequency ranges that are most

important for speech understanding. Using the AI, the percentage of loss can be calculated in

each band and combined into an index number representing the audibility of a typical speech

signal in a quiet background for the measured ear.

In the present study, articulation indices were calculated for the right and left ear for each

participant, and an average AI was calculated based on the right and left ear AI values. The

higher AI of these two values for each participant was considered the better-ear AI, whereas the

lower AI was considered the worse-ear AI.

A common and longstanding difficulty in regression analysis of hearing data is the method

by which audibility, the quality of a sound being perceptible, is quantified and controlled. A

review of literature was conducted to determine the most common practices for quantifying

audibility and, specifically, in studies that include analysis of predictive factors for speech

perception.

This literature review identified several methods that have been used to quantify audibility

for the purpose of regression analysis. One common method of calculating audibility is use of

the pure-tone average (PTA), or the arithmetic mean of the thresholds in dB hearing level (HL)

at a given set of test frequencies. The PTA is most commonly calculated as the average of

hearing thresholds at 500, 1000, and 2000 Hz (e.g,. George et al, 2007; Jerger et al, 1989).









The SPSS 11.5.0 software package was also used to calculate descriptive statistics for the

predictor variables, dependent variables, and other subject measures, such as age and gender,

The following variables examined and discussed in this section were not included as predictor

variables: GIN thresholds in the right ear (RGIN), left ear (LGIN), better ear (BETGIN), and

worse ear (WRSGIN); and working memory scores on forward digit span (DIGFWD), backward

digit span (DIGBAK), and digit ordering (DIGORD) tasks. All analyses were tested at alpha (a)

= .05. A glossary of abbreviations and acronyms used for the variables studied can be found in

Appendix C.

Distortion-Product Otoacoustic Emissions (DPOAEs)

Participant DPOAE data are summarized in Table 4-1 and Figure 4-1. Table 4-1 provides

a comparison table for DPOAE and audiometric status. In this table, participants demonstrating

normal audiometric data and normal DPOAE status within a frequency-range category were

classified as normal/normal; participants with abnormal DPOAEs and audiometric thresholds

were classified as abnormal/abnormal. Participants with audiometric thresholds within normal

limits but absent or abnormal DPOAEs were classified as normal/abnormal.

DPOAEs are a more sensitive index of cochlear function than is the pure tone audiogram

(Hall, 2000). The group of individuals with audiometrically-normal hearing and abnormal

DPOAEs can be interpreted as individuals who have cochlear dysfunction that is not yet detected

by a decrease in pure-tone threshold. Such individuals may demonstrate other difficulties

resulting from loss of outer hair cells, such as diminished cochlear fine-tuning and spectral

resolution ability, that are not apparent by pure-tone audiometry.

Four left ears demonstrated normal DPOAEs but abnormal audiometric thresholds for the

mid-frequency range only. Closer inspection of these four ears revealed steeply sloping hearing

loss in each of these ears, with thresholds below 2000 Hz within the normal range and thresholds









.482). As noted above, audibility accounted for the majority of variance in HINT score: 74% in

quiet and 52% in noise. The results of the regression indicate that no experimental variable

contributed a significant increase in predictive power.

Hearing Handicap

Self-reported hearing handicap was assessed using the age-appropriate version of the

Hearing Handicap Inventory: the Hearing Handicap Inventory for Adults (HHIA) for individuals

under age 65 and the Hearing Handicap Inventory for the Elderly (HHIE) for individuals age 65

and up. Scores on the HHIA/E varied widely: Emotional subscale scores (HHIEMOT) ranged

from zero to 44 with a mean of 10.8 and a standard deviation of 9.5. Social subscale scores

(HHISOC) ranged from 2 to 40 with a mean of 10.5 and a standard deviation of 6.9. Total scores

(HHITOT) ranged from 2 to 74 with a mean of 21.3 and a standard deviation of 15.1.

According to published reports, total scores of 18 or higher on the HHIE and 11 on the

HHIA indicate presence of a significant hearing handicap (Jerger et al, 1990; Weinstein, 2000;

Weinstein & Ventry, 1983). Of the 56 participants in the present investigation, 32 scored at or

above the level indicating handicap. On the HHIA, 14 of 27 participants scored above 11 (mean

= 16.4, SD = 12.1); on the HHIE, 18 of 29 participants scored above 18 (mean = 22.4, SD =

13.9).

All participants reported difficulty in at least one HHIA/E domain. The most commonly-

reported difficulties were in understanding when others speak in a whisper; when listening to

conversation while eating at a restaurant; and when listening when at a party.

Audibility was a significant predictor of both HHISOC (r = -.45, p = .001) and HHITOT (r

= -.35, p = .008). The negative valence of these correlations indicates that decreasing AI score

was correlated with increasing hearing handicap, particularly the social and situational aspect of

hearing handicap. No correlation was seen between HHIEMOT and BETAI.









CHAPTER 2
REVIEW OF LITERATURE

Hearing Impairment and Hearing Handicap

The disconnect in older adult populations between hearing impairment, determined using

audiometry and other psychoacoustic measures, and hearing handicap, as measured by personal

experiences of difficulty in real-world listening situations or by speech perception testing in

difficult listening environments, is well established. Early examinations of the Hearing

Handicap Inventory for the Elderly (HHIE), a commonly-used clinical questionnaire, found that

pure-tone sensitivity and word-recognition scores explained less than 50% of the variance in

HHIE score (Weinstein & Ventry, 1983). Other studies have found even lower correlations

between audiometric and self-report measures (Brainerd & Frankel, 1985; John & Kreisman, in

preparation; Matthews, Lee, Mills, & Schum, 1990; McKenna, 1993; Newman, Jacobson, Hug,

& Sandridge, 1997). These studies typically find that audiometric measures tend to

underestimate hearing handicap in individuals with more severe degrees of sensorineural hearing

loss (SNHL), and that the relation between sensitivity and hearing handicap becomes more

variable in older adults.

The correlation between hearing impairment and self-reported hearing handicap is

influenced by several non-auditory factors (Gatehouse, 1994, 1998; Gordon-Salant, Lantz, &

Fitzgibbons, 1994; Lutman, 1991; Lutman, Brown, & Coles, 1987; Stephens & Hetu, 1991).

Gatehouse (1998) found substantial variability in the correlation between self-reported hearing

handicap and hearing loss, with significant effects of age and personality. Lutman and

colleagues (1987) noted that age, gender, and socioeconomic status are factors in hearing self-

report, with males reporting more disability than females and older individuals reporting less

disability than younger individuals. In another study, Lutman (1991) concluded that older















II~ II


125 250 500


1k 2k 4k 8k


Frequency

Figure 4-2. Mean audiogram with standard deviation bars.









Middelweerd, M. J., Festen, J. M., & Plomp, R. (1990). Difficulties with speech intelligibility in
noise in spite of a normal pure-tone audiogram. Audiology, 29(1), 1-7.

Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our
capacity for processing information. Psychological Review, 63, 81-97.

Miller, G. A. & Licklider, J. C. R. (1950). The intelligibility of interrupted speech. Journal of
the Acoustical Society ofAmerica, 22, 167-173.

Moncur, J. & Dirks, D. (1967). Binaural and monaural speech intelligibility in reverberation.
Journal of Speech and Hearing Research, 10, 1557-1565.

Moore, B. C. J. (1997). An Introduction to the Physiology of Hearing. San Diego: Academic
Press.

Moore, B. C. J. & Glasberg, B. R. (1988). Gap detection with sinusoids and noise in normal,
impaired, and electrically-stimulated ears. Journal of the Acoustical Society ofAmerica, 83,
1093-1101.

Moore, B. C. J., Glasberg, B. R., Donaldson, E., McPherson, T., & Plack, C. J. (1989).
Detection of temporal gaps in sinusoids by normally hearing and hearing-impaired subjects.
Journal of the Acoustical Society ofAmerica, 85, 1266-1275.

Moscovitch, M. & Winocur, G. (1992). The neuropsychology of memory and aging. In F. Craik
& T. Salthouse (Eds.), The Handbook ofAging and Cognition (pp. 315-372). Hillsdale, NJ:
Erlbaum.

Mueller, H. G. & Killion, M. C. (1990). An easy method for calculating the articulation index.
The Hearing Journal, 9, 14-17.

Musiek, F. E., Shinn, J. B., Jirsa, R., Bamiou, D. E., Baran, J. A., & Zaidan, E. (2005). GIN
(Gaps-In-Noise) test performance in subjects with confirmed central auditory nervous system
involvement. Ear and Hearig. 26, 608-618.

Musiek, F. E., Wilson, D. W., & Pinheiro, M. L. (1979). Audiological manifestations in "split-
brain" patients. Ear and Hearing. 5, 25-29.

Nabelek, A. K. (1988). Identification of vowels in quiet, noise, and reverberation: Relationships
with age and hearing loss. Journal of the Acoustical Society ofAmerica, 84(2), 476-484.

Nabelek, A. K. & Robinson, P. K. (1982). Monaural and binaural speech perception in
reverberation for listeners for various ages. Journal of the Acoustical Society ofAmerica, 71(5),
1242-1248.

Neils, J., Newman, C. W., Hill, M., & Weiler, E. (1991). The effects of rate, sequencing and
memory on auditory processing in the elderly. Journal of Gerontology, 46(2), 71-75.









Working Memory

Cognitive performance is a possible explanation for suprathreshold speech-perception

deficits. However, research into this relation between cognitive function and speech perception

has produced equivocal results, with some studies demonstrating a positive correlation between

cognitive abilities and speech perception (Kjellberg, 2004; Pichora-Fuller & Singh, 2006;

Pichora-Fuller et al, 1995; Rabbitt, 1991; Smoorenburg, 1992; van Rooij & Plomp, 1990;

Wingfield, 1996) but others finding no such association (Humes & Floyd, 2005; Humes, Watson,

Christensen, Cokely, Halling, & Lee, 1994; Stach, Loiselle, & Jerger 1991; Stach et al, 1985; van

Rooij, Plomp, & Orlebeke, 1989). This inconsistency is most likely due to the broad range of

methods used by hearing researchers to assess cognition and speech perception. For example, in

two studies examining the association between speech perception and cognitive decline in older

adult subjects, van Rooij and Plomp (1990) found that one-third of the variance in an adaptive

sentence recognition task was related to cognitive status (measured by a combination of speed of

processing, memory, and intellectual ability). On the other hand, Jerger et al (1991) found that

cognitive status, as measured by a speed of processing task only, accounted for between 6

percent (for word recognition) and 14 percent (for dichotic sentence recognition) of variance.

There were differences in the two studies in the speech stimulus used to measure auditory

processing and in the battery used to assess cognitive status.

There is a bi-directional relation between peripheral processing status (such as hearing

sensitivity) and cognitive processing. Declines in peripheral sensory status, such as in loss of

hearing sensitivity, may provide a distorted and impoverished signal for later states of cognitive

processing, which may lead to declines in cognitive status (Appollonio, Carabellese, Magni,

Frattola, & Trabucci, 1995; Gennis, Garry, Haaland, Yeo, & Goodwin, 1991; Lindenberger &

Baltes, 1994; Ryan, Giles, Bartoloucci, & Henwood, 1986; Sands & Meredith, 1989; Uhlmann,









The cognitive hypotheses are supported also by Neils and colleagues (1991) who reported

that a group of older adults with mean age of 80, but not a younger group of older adults (mean

age of 70), performed significantly worse than a group of young adults (mean age of 25) on an

auditory memory task (remembering sequences of tones), particularly for long sequences and

with short intervals between tones. The performance deficit was not related to hearing

sensitivity, suggesting age-related deficits beyond loss of hearing sensitivity and involving

memory and rapid processing function. In fact, an important feature of working memory from

the perspective of the audiologist or hearing researcher is that declines in working memory

appear to be independent of changes in hearing status (Lyxell, Andersson, Borg, & Ohlsson,

2004).

To date, the few studies that have examined closely the role of working memory in

difficult speech perception have produced mixed results. While several studies provide evidence

for a role of working memory capacity in perception of difficult or degraded speech signals,

others have failed to find an association. Discrepancies in findings are likely a result of

methodological differences in testing both working memory capacity and speech perception

ability (i.e. Gordon-Salant & Fitzgibbons, 1997; Humes & Floyd, 2005; Lunner, 2003; Pichora-

Fuller & Singh, 2006; Pichora-Fuller et al, 1995; Tun, Wingfield, & Stine, 1991; Vaughan et al,

2006). For example, Humes and Floyd (2005) assessed working memory and sequence learning

using a modified version of the Simon memory game, with participants measured on their ability

to follow audiovisual cues and repeat a tapping pattern on the game panels. Speech perception

was measured using nonsense syllable repetition and the Connected Speech Test (CST) in

backgrounds of noise. The authors reported that an association between performance on the









Analysis of the participant sample separated into three age groups confirmed significant

age-related declines in all measures except HHIA/E Emotional subscale and total scores.

Separate regression procedures for these three groups revealed changing associations between

the experimental and dependent variables with increasing age.

For the two older groups, speech perception thresholds in quiet and in noise were predicted

increasingly well by hearing sensitivity and, in the case of the oldest group for speech perception

in quiet, by hearing sensitivity and temporal resolution together. For the youngest group, speech

perception was not well predicted by any measure. Hearing handicap scores were predicted

poorly or not at all by the experimental measures in each age group.

A second grouped analysis of the sample separated participants with normal hearing from

participants with hearing loss as determined by pure-tone thresholds. Restricting the analysis

only to those participants with audiometrically-measurable hearing loss did not significantly

improve the predictive ability of the experimental variables for speech perception or hearing

handicap.










Figure 3-1. Diagram of sound-field HINT test environment.



D-75A







CCD
BP-2X 1 meter BP-2X



1 meter
T-600

BP-2X BP-2X GSI-61








D-75A



BP-2X = Definitive BP-2x Bipolar Speaker (noise source)
CD = Sony CDP-CE375 5-CD Changer (stimulus and noise routing)
D-75A = Crown D-75A Two-Channel Amplifier (noise routing)
GSI-61 = GSI-61 Clinical Audiometer (stimulus routing)
T-600 = Tannoy System 600 Monitor Speaker (stimulus source)


LiU









Finally, while the GIN test appears to have good face validity as a gap detection measure,

it is a new tool with limited published data on reliability and accuracy. As noted above, the

current version of the test appears to lack gap stimuli of appropriate length (specifically 7 ms) for

precise diagnosis of impaired temporal resolution according to published norms. In addition,

GIN normative data are lacking for populations other than young adults with normal hearing.

Future versions of the GIN test, and future clinical investigations, may address these issues.

Future Directions

This investigation provides a starting place for future research into the factors influencing

speech perception in older adults. First, the associations between both of the predictor variables

(temporal resolution and working memory) should be tested with updated speech perception

measures. Specifically, the steady-state HINT competition noise should be replaced by

modulated noise in an attempt to identify temporal resolution effects. The present investigation

focused on the role of a single distortion (noise) of the speech signal. Follow-up studies might

use reverberation effects to further distort the stimulus signal and/or the noise. Previous

researchers have suggested that a combination of noise and reverberation is not only more

reflective of real-world environments, but also may affect speech perception to a greater degree

than either of the distortions alone (Danhauer and Johnson, 1991; Gelfand and Silman, 1979;

Helfer, 1992, 1994; Helfer and Huntley, 1991; Johnson, 2000).

Second, working memory effects should be examined using more complex speech stimuli,

such as low-context sentences or extended passages from which the participant must recall

information heard early in the reading. Low-context or nonsense sentences are likely to be more

difficult to recall because linguistic knowledge can not be used to compensate for sounds or

words that are not heard. Long speech passages may be a more effective test of working

memory than are the shorter HINT sentences as they would require the listener to retain










Table 4-12. ANOVA table for comparison of participant data, grouped by age, on age,
audibility, working memory, HINT, and HHIA/E variables.
Age 50 59 Age 60 69 Age 70 89 ANOVA Results
(N = 17) (N = 19) (N = 20)
Mean SD Mean SD Mean SD F Sig.


3.10 64.67


.95 .09

.92 .09

.96 .07

.92 .10

.94 .08


DIGFWD 8.47

DIGBAK 7.47

DIGORD 18.53


1.06

6.76

6.65

6.29


WRSGIN 7.12


BINGIN

HINTQ

HINTN


HHIEMOT

HHISOC


31.91

71.32

10.24


7.18


2.10

2.15

3.17

2.34

1.44

1.58

1.40

1.50

1.38


8.95

7.00

16.47


2.73 76.86

.16 .68

.24 .64

.14 .71

.24 .61

.16 .66

2.10 6.95

1.89 5.90

2.57 14.50


.55 1.83 -1.42


7.63

8.37

7.26

8.74

8.00


4.08 37.35


1.80

9.90

4.95


HHITOT 17.41 13.73


71.63

9.47

9.58

19.05


2.59 8.30

3.04 9.85

2.66 8.00

2.83 10.15

2.56 9.08

7.88 45.64

1.74 73.54

6.49 12.60

5.44 14.10

10.65 26.70


4.83 148.19

.18 16.09

.22 9.53

.16 18.19

.22 10.51

.18 15.94


5.62


2.00 3.03

4.94 5.34

2.42 6.64

1.75 2.71

3.25 6.14

1.72 3.30

3.08 6.22

2.20 5.69

9.18 15.79

3.38 4.55


11.59


8.07 5.68

18.53 2.14


* = difference is significant at the .05 level.
** = difference is significant at the .01 level.

RAI = Articulation Index in right ear; LAI = Articulation Index in left ear; BETAI = Articulation Index in
better ear; WRSAI = Articulation Index in the worse ear; BINAI = average binaural Articulation Index;
DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; WM =
composite working memory score; RGIN = Gaps-In-Noise threshold in right ear; LGIN = Gaps-In-Noise
threshold in left ear; BETGIN = Gaps-In-Noise threshold in better ear; WRSGIN = Gaps-In-Noise
threshold in worse ear; BINGIN = average Gaps-In-Noise threshold in right and left ears; HINTQ = HINT
threshold in quiet; HINTN = HINT threshold in noise; HHIEMOT = HHIA/E Emotional subscale score;
HHISOC = HHIA/E Social/Situational subscale score; HHITOT = HHIA/E total score


AGE

RAI

LAI


BETAI

WRSAI

BINAI


WM


RGIN

LGIN

BETGIN


.000**

.000**

.000**

.000**

.000**

.000**

.006**


.057


.008**

.003**

.076

.004**

.045*

.004**

.006**

.000**

.015*


.573


.006**

.127









Table 3-1. Summary of the University of Florida 65/55 Distortion Product Otoacoustic
Emissions (DPOAE) protocol.
Point f, (Hz) f2 (Hz) L2 ff, ratio DP(Hz)
(dB SPL) f(dB SPL)
1 6070 65 7277 55 1.20 4863
2 5285 65 6340 55 1.20 4230
3 4582 65 5496 55 1.20 3668
4 4020 65 4816 55 1.20 3223
5 3492 65 4184 55 1.20 2801
6 3023 65 3621 55 1.20 2426
7 2625 65 3152 55 1.20 2098
8 2297 65 2754 55 1.20 1840
9 2004 65 2402 55 1.20 1605
10 1746 65 2098 55 1.20 1395
11 1512 65 1816 55 1.20 1207
12 1324 65 1594 55 1.20 1055
13 1148 65 1371 55 1.20 926
14 984 65 1184 55 1.20 785
15 855 65 1031 55 1.20 680
16 750 65 902 55 1.20 598
17 656 65 785 55 1.20 527
18 574 65 691 55 1.20 457
19 504 65 598 55 1.20 410
fi= frequency of first stimulus tone in Hertz; L1 = intensity of first stimulus tone in
decibels sound pressure level ; f2 = frequency of second stimulus tone in Hertz; L2 =
intensity of second stimulus tone in decibels sound pressure level ; DP = distortion
product frequency in Hertz

































2007 Andrew Barnabas John









Divenyi, P. L. & Haupt, K. M. (1997). Audiological correlates of speech understanding deficits
in elderly listeners with mild-to-moderate hearing loss: Age and lateral asymmetry effects. Ear
and Hearing. 18, 42-61.

Divenyi, P. L. & Simon, H. J. (1999). Hearing in aging: Issues old and new. Current Opinions
in Otolaryngology, 7, 282-289.

Dixon Ward, W. (1983). The American Medical Association / American Academy of
Otolaryngology formula for determination of hearing handicap. Audiology, 22, 313-324.

Dobie, R. A. & Megerson, S. C. (2000). Worker's Compensation. In E.H. Berger, L.H. Royster,
J.D. Royster, D.P. Driscoll, & M. Layne, M. (Eds.), The Noise Manual, 5th Ed (pp. 698-710).
American Industrial Hygiene Association: Fairfax, VA.

Dorman, M., Marton, K., & Hannley, M. (1985). Phonetic identification by elderly normal and
hearing-impaired listeners. Journal of the Acoustical Society ofAmerica, 77, 664-670.

Dubno, J. R. & Dirks, D. D. (1989). Auditory filter characteristics and consonant recognition for
hearing-impaired listeners. Journal of the Acoustical Society ofAmerica, 85(4), 1666-1675.

Dubno, J. R. & Dirks, D. D. (1990). Associations among frequency and temporal resolution and
consonant recognition for hearing-impaired listeners. Acta Otolaryngologica (Suppl), 469, 23-
29.

Dubno, J. R., Horwitz, A. R., & Ahlstrom, J. B. (2002). Benefit of modulated makers for
speech recognition by younger and older adults with normal hearing. Journal of the Acoustical
Society ofAmerica, 111, 2897-2907.

Dubno, J. R., Horwitz, A. R., & Ahlstrom, J. B. (2005). Word recognition in noise at higher-
than-normal levels: decreases in scores and increases in masking. Journal of the Acoustical
Society ofAmerica, 118, 914-922.

Duffy, F. H., McAnulty, G. B., & Albert, M. S. (1996). Effects of age upon interhemispheric
EEG coherence in normal adults. Neurobiology ofAging. 17(4), 587-599.

Duquesnoy, A. J. & Plomp, R. (1980). Effect of reverberation and noise on the intelligibility of
sentences in cases of presbyacusis. Journal of the Acoustical Society ofAmerica, 68(2), 537-
544.

Eddins, D. A., Hall, J. W., & Grose, J. H. (1992). The detection of temporal gaps as a function
of frequency region and absolute noise bandwidth. Journal of the Acoustical Society ofAmerica,
91(2), 1069-1077.

Eisenberg, L. S., Dirks, D. D., & Bell, T. S. (1995). Speech recognition in amplitude-modulated
noise of listeners with normal and listeners with impaired hearing. Journal of Speech Language
and Hearing Research, 38, 222-233.









Other working memory tasks include reading span, which requires individuals to read

increasingly longer sets of sentences out loud while recalling the final word of each sentence in

each set, n-back recall, in which individuals are asked to track the identity or location of a series

of verbal or nonverbal stimuli and to indicate whether a stimulus is the same as the one presented

n trials previously, and pattern or sequence recall tasks, which require an individual to monitor

and replicate a pattern of visual or auditory stimuli.

Working memory capacity declines over the course of the lifespan, with steepest declines

seen in older adults after age 80 (Carpenter, Miyake, & Just, 1994; DeDe, Caplan, Kemtes, &

Walters, 2004; Foos & Wright, 1992; Gilinsky & Judd, 1994; Hultsch, Hertzog, Small,

McDonald-Miszczak, & Dixon, 1992; Norman, Kemper, & Kynette, 1992; Palladino & De Beni,

1999; Salthouse, 1991; Wingfield, Stine, Lahar, & Aberdeen, 1988). Working memory loss is a

significant factor in the overall cognitive deterioration observed in older adults (Carpenter et al,

1994; Moscovitch and Winocur, 1992; Salthouse, 1994).

Speech is a complex, rapid acoustic signal. Acoustic and linguistic information in a speech

signal is processed, on average, at a rate between 200 and 270 words per minute, with substantial

variability in speaking rate both between and within talkers (Stine, Wingfield, & Poon, 1986;

Vaughan, Storzbach, & Furukawa, 2006; Wingfield, Poon, Lombardi, & Lowe, 1985). Because

speech is received at such a high rate, listeners are required to store and process incoming

information very quickly, a function that is diminished in older adults compared to younger

adults independent of the effects of peripheral hearing loss (Gordon-Salant and Fitzgibbons,

2001; Stine & Wingfield, 1987; Vaughan & Letowski, 1997; Wingfield et al, 1985).

Still, for most older adults, perception in ideal conditions is generally unaffected, as

information is easily accessed and processed and little strain is placed on the cognitive resources









with "normal" hearing were still elevated compared to the normative averages of 4.9 and 4.8 ms,

suggesting the possibility of a small age effect. Further study is needed to separate the effects of

age and hearing loss on GIN threshold. Larger samples of age-matched groups with varying

levels of hearing loss, or hearing-loss-matched groups differing in age, might provide insight into

the role of each factor.

Age- and hearing-loss-appropriate normative criteria are needed for the GIN test. The

number of participants who would be classified abnormal under the original norms is quite

high-over 70% of a group of a non-clinical population. In fact, no participant had GIN

threshold in either ear better than 5 ms, which is marginally above the average thresholds seen in

the normative group. A second indication of a need for expanded normative data is the finding

of an apparent age-related left-ear disadvantage on the GIN. This increase is most apparent in

Table 4.12, where right- and left-ear differences in hearing sensitivity (RAI vs. LAI) and GIN

threshold (RGIN vs. LGIN) are displayed for groups of participants in their 50s (Group 1), 60s

(Group 2), and 70s and 80s (Group 3). Left-ear GIN threshold showed a more rapid decline with

age than did right-ear threshold: Inter-ear differences were 0.11 ms in Group 1, 0.64 in Group 2,

and 1.55 in Group 3. While the inter-ear difference in GIN threshold was significant only

between Group 2 and Group 3, an increase in asymmetry is evident across the age range of the

sample. This increase in asymmetry is apparently unrelated to audibility: the inter-ear

difference in AI was stable and did not change significantly between groups. If this ear

asymmetry is seen consistently in future GIN investigations, development of both clinical norms

that are both age-appropriate and ear-appropriate would be indicated.

Asymmetry favoring the right ear is common in hearing studies on older adults. Research

confirms enhanced function for the right ear, i.e., a left-ear disadvantage, in aging for selected









Two GIN lists were randomly selected from the four available lists for each participant.

One list was presented to the right ear and one to the left ear in a random order. Testing was

conducted monaurally, using one list per ear. The order of ear tested was alternated from

participant to participant. No GIN list was used twice for any participant. Prior to administering

the GIN test, the investigator explained the task according to the instructions provided with the

test materials. Each participant was then given several stimulus noise bursts from a practice list

prior to the beginning the test scoring. Threshold measures were collected at each participant's

most comfortable listening level (MCL). For all participants, MCL was between 35 and 50 dB

SL re: SRT, the recommended presentation level parameters for the GIN. Recent studies have

indicated that GIN threshold is insensitive to presentation level within this range (Musiek et al,

2005; Weihing et al, 2007).

The GIN test was conducted in a single-walled sound-treated booth under supra-aural

headphones. Gap stimuli were presented monaurally via the headphones and responses were

obtained using the audiometer's response button. To assist with scoring, the second (non-

stimulus) channel of the test CD provides beeps aligned with the gaps on the stimulus channel.

The second channel of the CD player was routed through the second channel of the audiometer

and to an absent transducer (speaker) to provide cues to the presence of the gaps to the tester,

without being audible to the participant. By monitoring the non-stimulus channel, the

investigator could determine accurately whether a gap had occurred prior to each participant

responding button press.

Monaural GIN thresholds were calculated for each ear and are denoted as RGIN (GIN

threshold in the right ear) and LGIN (GIN threshold in the left ear). In addition, the better-ear

GIN (BETGIN), worse-ear GIN (WRSGIN), and a composite binaural GIN (BINGIN) were









interhemispheric communication or deterioration of neural pathways. Asymmetric loss of

central auditory function is not a new concept in hearing research. The present findings

emphasize the need to account for this unique feature of older adult auditory processing when

attempting to explain the nature of hearing loss with age.









Self-Reported Hearing Handicap: Hearing Handicap Inventory for Adults (HHIA) or for
the Elderly (HHIE)

Participants age 65 or older completed the Hearing Handicap Inventory for the Elderly

(HHIE, Weinstein & Ventry, 1983), whereas participants under age 65 completed the Hearing

Handicap Inventory for Adults or, HHIA, Newman, Weinstein, Jacobson, & Hug, 1990). The

HHIA/E is a 25-item inventory that assesses the consequences of hearing loss on the lives of

adults. Two sub-scales comprise the HHIA/E: the HHIA/E Emotional subscale measures the

mental and emotional effects of hearing loss on the individual, while the HHIA/E

Social/Situational subscale measures the situational and interpersonal consequences of hearing

loss. Both scales are frequently used in audiologic clinics to document the need for, and benefit

from, rehabilitation.

The HHIE and HHIA are almost identical; three questions are slightly reworded in the

HHIA to make the inventory more applicable to individuals under age 65. For each item, the

participant was asked to respond "Yes", "Sometimes", or "No." After participants completed the

inventories, the investigator calculated Emotional and Social/Situational subscale scores as well

as a total score on a 100-point scale. The Emotional subscale score of the HHIA and HHIE

measures the emotional impact of hearing loss on the individual. The Social/Situational subscale

score quantifies the situational impact of hearing loss (Weinstein & Ventry, 1983).

The HHIA and HHIE scales were administered face-to-face with the answer forms marked

by the investigator. The participant was provided with a flash card containing the possible

answers. When necessary, a frequency-modulated (FM) listening system was used to ensure

audibility of the investigator's questions. For the purpose of analysis, the HHIA/E Emotional

subscale score was denoted as HHIEMOT and the Social/Situational subscale score was denoted

as HHISOC. The HHIA/E total score was denoted as HHITOT.









(DPOAEs), and assessment of pure-tone air-conduction thresholds, speech reception thresholds,

and word- recognition performance.

Aural Immittance Measurement.

Tympanometry measures the volume of the ear canal and compliance of the tympanic

membrane (eardrum) and middle ear system as air pressure in the external ear canal is

systematically varied from +200 decaPascals (daPa) to -200 daPa. Aural immittance measures

help to identify middle ear dysfunction that may be associated with conductive hearing loss.

Measurement of acoustic reflexes evaluates the integrity of the auditory pathway including the

middle ear, the cochlea, the auditory (8th cranial) nerve, and lower brainstem ponss) by the

detection of a reflex of the stapedius muscle within in the middle ear. For this investigation,

ipsilateral acoustic reflexes (measurement of stapedius muscle contraction in the same ear as the

stimulus tone was presented) were screened at 90 dB HL bilaterally for pure-tone signals of 500,

1000, 2000, and 4000 Hz. Tympanometry and acoustic reflex screening were conducted using a

Grason-Stadler GSI-16 immittance meter.

Distortion-Product Otoacoustic Emissions

Measurement of DPOAEs evaluates function of the outer hair cells in the cochlea.

DPOAEs are elicited by the simultaneous presentation of two pure tones closely spaced in

frequency. When the cochlea is healthy and the outer hair cells are motile, introduction of the

tone pair to the ear will activate the cochlea in the regions of the basilar membrane

corresponding to the frequency of those tones. When the resulting traveling waves along the

basilar membrane overlap, energy peaks are produced at several discrete frequencies related to

the frequencies of the stimulus tones. This output (a distortion product or DP) can be used

clinically as an objective test of cochlear health: if the energy peak is sufficiently high, this

suggests that the portion of the cochlea stimulated by the input tones is intact and responding









=.60, p < .001). These relations suggest an association among these measures, but substantial

variance in each unexplained by the others.

There was a significant negative correlation between age and all working memory

measures: DIGFWD (r = -.33, p = .012), DIGBAK (r = -.32, p = .017), DIGORD (r = -.48, p <

.001), and total WM (r = -.47, p < .001), suggesting an age-related decrease in working memory

status. When controlling for age, none of the working memory variables (DIGFWD, DIGBAK,

DIGORD, and WM) were significantly correlated with BETAI.

Collinearity Testing

Collinearity testing was conducted for each experimental variable (GIN thresholds and

working memory scores) to determine whether the variance each predicted in HINT or HHIA/E

score, if any, was shared with the other experimental variables. A correlation matrix for the

variables RGIN, LGIN, BETGIN, WRSGIN, BINGIN, DIGFWD, DIGBAK, DIGORD, and

WM can be seen in Table 4-6. A small but significant negative correlation was seen between

RGIN and DIGORD (r = .27, p = .046). No other correlations were significant.

Because independence between the temporal resolution and working memory measures was

anticipated, the unexpected correlation between right-ear GIN threshold and digit-ordering

ability was examined further. As described previously, all GIN and working memory variables

were significantly correlated with age. When age was controlled, the correlation between RGIN

and DIGORD disappeared (r = -.11, p = .440). It is likely that the association between RGIN

and DIGORD is a spurious correlation caused by the influence of age on each of the

experimental variables.

Speech Perception

Speech perception thresholds with and without noise competition were obtained

according to the recommended procedure for the HINT as described in Chapter 3. Thresholds in









According to scoring criteria for the GIN test, this participant's threshold of 8 ms is outside of

the normal range. However, his ability to identify correctly some gaps at 6 ms indicates that his

actual gap threshold may be closer to 7 ms and, according to the normative criterion, within

normal limits. It appears, therefore, that the current version of the GIN needs to be revised to

test gaps closer to the normative cutoff if it is to be used as a diagnostic tool.

The high percentage of participants with GIN thresholds outside of the 7 ms cutoff

suggests that the normative data published by Musiek et al may not be appropriate for adults

above age 50 or for adults with SNHL. As of this writing, no GIN test normative data have been

published for individuals in the age range recruited for this investigation or for any other sample

of older adults. The youngest participant recruited, at 50 years, 1 month of age, was still four

years older than the oldest participant in the normative group. Perhaps more importantly, no

normative data have been published for individuals with confirmed SNHL. As noted above, it

appears that the elevated GIN thresholds in this sample are a result of hearing loss and not of

age. However, no firm conclusion can be drawn from these data because, in comparison to the

normative subject sample (Musiek et al, 2005), the sample in the present study had greater

hearing loss and was older.

As expected based on the recruitment criteria (individuals who believed they had a hearing

problem), few participants presented with audiometrically-normal hearing. In all, six

participants' right ears and six left ears in the had pure-tone sensitivity at 20 dB HL or better at

all frequencies. The average GIN thresholds for these "normal" ears (mean age 55.1) were 6.2

ms on the right (mean age 55.1), and 5.8 ms on the left (mean age 58.8). These data, although

clearly limited, suggest that the average elevation in GIN threshold in the whole sample is

primarily a function of hearing loss. However, mean gap thresholds for these individual ears









rather than grouping a sample categorically, can reveal changes or trends across a continuous

range of ages or increments of hearing loss. This data analysis method also allows for

examination of the way variable associations change across age ranges, for example, determining

whether the association between speech perception and audibility is comparable for individuals

in their 50s as it is for individuals in their 70s and 80s.

Gaps-in-Noise (GIN) Test Findings

Results of GIN testing were consistent with previous studies finding that individuals with

serious hearing loss (lower AI scores) demonstrate significantly longer gap detection thresholds

(higher GIN thresholds) than individuals with little or no hearing loss (Arlinger & Dryselius,

1990; Buus & Florentine, 1985; Madden & Feth, 1992; Moore et al, 1989). Similar to findings

of most auditory temporal resolution studies, in the present study declining temporal resolution

ability was associated with decreased hearing, but much of the variance in temporal resolution

ability was unexplained by hearing sensitivity. Monaural audibility measures accounted for

about 26% of the variance in temporal resolution ability in the right ear and 41% in the left ear.

The apparent independence of GIN threshold and age, with hearing loss controlled, was also

consistent with previous research (e.g., He et al, 1999).

Importantly, analysis of the GIN test data revealed that more than 70% of the participant

group (40 of 56) demonstrated temporal thresholds outside of the normative range in at least one

ear. Of these 40 participants, 19 had worse-ear GIN thresholds of 8 ms, the gap iteration just

above the 7 ms norm (Musiek et al, 2005). Gap lengths evaluated with the GIN test are 2, 3, 4,

5, 6, 8, 10, 12, 15, and 20 ms. There is no 7 ms gap. The omission of the 7 ms gap increment

leaves open the possibility that individuals with borderline-normal temporal resolution might be

classified as abnormal. For example, in this investigation participant #2 correctly identified 6 ms

gaps two out of six times (33%) and 8 ms gaps five out of six times (83%) in his right ear.









ACKNOWLEDGMENTS

First and last and most, I acknowledge my mother, Virginia Giglio. I can never thank her

enough for her encouragement and faith in me throughout my career as a student. Ha hoo.

I am fortunate enough to have a supportive, patient, and understanding family, without

whom it would have been awfully hard to make it this far, especially my father, Andrew C. John,

and my stepfather, Neal Dunnigan. I also thank my lovely and loving fiancee, Christi Barbee,

for putting up with me throughout this year of dissertation data collection, analysis, and writing.

I know I have not always been fun to be around and I haven't been able to give the time and

attention she deserves, but I look forward to making up for it when life slows down and I can

relax a bit, some fifty or sixty years from now.

I acknowledge my doctoral committee members: Dr. Jay Hall, Dr. Scott Griffiths, Dr.

Patricia Kricos, and Dr. Robin West. I am grateful for their support, time, and energy from the

inception of this project to its completion. I am also grateful to Dr. Paul Duncan and Dr.

Christine Sapienza for their service on my academic qualifying committee.

I am indebted to my friends and colleagues at UF, particularly Drs. Brian and Nicole

Kreisman and Dr. Kristin Johnston, for moral and material support throughout my doctoral

program.

I also owe a great deal to many people whose help in completing this project was

indispensable: in addition to those named above, I acknowledge Drs. Michelle Colbum, Lisa

Edmonds, and Mini Shrivastav of the University of Florida Department of Communication

Sciences and Disorders; Dr. Mary Anne Pinner of Hampton Oaks ENT; Mrs. Shirley Bloodworth

of the Prime Time Institute; Mrs. Lynn Rousseau and Mrs. Nancy Day of the Hearing Loss

Association of Gainesville; and Mrs. Elaine Brown of the Village Retirement Community for









Table 4-11. Comparison across participant age groups for descriptive factors, experimental measures, and dependent variables.
DIG DIG DIG HINT HINT HHI HHI HHI
Age Male Female N RGIN LGIN
FWD BAK ORD Q N EMOT SOC TOT
50- 59 56.0 6 11 17 8.5 7.5 18.5 6.8 6.7 31.9 71.3 10.2 7.2 17.4
SD 3.1 2.1 2.2 3.2 1.4 1.6 4.1 1.8 9.9 5.0 13.7

60- 69 64.7 5 14 19 9.0 7.0 16.5 7.6 8.4 37.4 71.6 9.5 9.6 19.1
SD 2.7 2.1 1.9 2.6 2.6 3.0 7.9 1.8 6.5 5.4 10.7

70- 89 76.9 6 14 20 7.0 5.9 14.5 8.3 9.9 45.6 73.5 12.6 14.1 26.7
SD 4.8 1.6 2.0 4.9 1.8 3.3 9.2 3.4 11.6 8.1 18.5
N = number of participants in group; DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering;
RGIN = Gaps-In-Noise threshold in right ear; LGIN = Gaps-In-Noise threshold in left ear; HINTQ = Hearing In Noise Test
threshold in quiet condition; HINTN = Hearing In Noise Test threshold in noise condition; HHIEMOT = Hearing Handicap
Inventory for Adults / for the Elderly Emotional subscale score; HHISOC = Hearing Handicap Inventory for Adults / for the Elderly
Social/Situational subscale score; HHITOT = Hearing Handicap Inventory for Adults / for the Elderly Emotional aggregate score










Table 4-7. Correlation matrix for HINT, HHIA/E, and audibility variables (N = 56).
HHI
HINTQ HINTN EOT HHISOC HHITOT BETAI
EMOT
r 1 .660** .206 .375** .302* -.858**
Sig. ..000 .128 .004 .024 .000

HTN r .660** 1 .332* .452** .417** -.722**
HINTN
Sig. .000 .012 .000 .001 .000

HHI r .206 .332* 1 .679** .942** -.231
EMOT Sig. .128 .012 .000 .000 .087

r .375** .452** .679** 1 .886** -.448**
HHISOC
Sig. .004 .000 .000 ..000 .001

r .302* .417** .942** .886** 1 -.351**
HHITOT
Sig. .024 .001 .000 .000 ..008

r -.858** -.722** -.231 -.448** -.351** 1
BETAI
Sig. .000 .000 .087 .001 .008
= correlation is significant at the .05 level.
** = correlation is significant at the .01 level.

HINTQ = HINT threshold in quiet; HINTN = HINT threshold in noise; HHIEMOT =
HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale
score; HHITOT = HHIA/E total score; BETAI = Articulation Index score in the
better ear










Table 4-6. Correlation matrix for GIN and working memory variables (N = 56).


DIG DIG DIG
WM
FWD BAK ORD
DIG r 1 .492** .366** .764**
FWD Sig. .000 .006 .000


IN LIN BET WRS BIN
RGIN LGIN
GIN GIN GIN
.033 .095 .078 .068 .077
.807 .486 .569 .620 .572


.598** .859**
.000 .000


DIG r .366** .598** 1
ORD Sig. .006 .000

W r .764** .859** .808** 1
WM
Sig. .000 .000 .000


r .033
RGIN
Sig. .807

r .095
LGIN
Sig. .486

BET r .078
GIN Sig. .569

WRS r .068
GIN Sig. .620


.023 -.268* -.087
.865 .046 .524

.131 -.146 .033
.336 .285 .809

.072 -.222 -.030
.599 .100 .827

.103 -.185 -.006
.450 .172 .965


.023
.865


808** -.268* -.146 -.222 -.185 -.215
000 .046 .285 .100 .172 .111


-.087
.524


.033 -.030 -.006 -.017
.809 .827 .965 .900

.632** .939** .703** .862**
.000 .000 .000 .000


.632** 1
.000


.744** .971** .938**
.000 .000 .000


.939** .744** 1
.000 .000


.733** .907**
.000 .000


.703** .971** .733** 1
.000 .000 .000


.951**
.000


BIN r .077 .096 -.215 -.017
GIN Sig. .572 .481 .111 .900
* = correlation is significant at the .05 level.
** = correlation is significant at the .01 level.


.862** .938** .907** .951** 1
.000 .000 .000 .000


DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; WM
= composite working memory score; RGIN = Gaps-In-Noise threshold in right ear; LGIN =
Gaps-In-Noise threshold in left ear; BETGIN = Gaps-In-Noise threshold in better ear; WRSGIN
= Gaps-In-Noise threshold in worse ear; BINGIN = average Gaps-In-Noise threshold in right and
left ears


DIG r .492** 1
BAK Sig. .000




Full Text

PAGE 1

T HE ROLE OF CLINICALLY APPLICABLE TEMPORAL RESOLUTION AND WORKING MEMORY TESTS IN PREDICTION OF SPEECH PERCEPTION AND HEARING HANDICAP By ANDREW BARNABAS JOHN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSIT Y OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Andrew Barnabas John

PAGE 3

3 To Dr. Carl Crandell

PAGE 4

4 ACKNOWLEDGMENTS First and last and most, I a cknowledge my mother, Virginia Giglio. I can never thank her enough for her encouragement and faith in me throughout my career as a student. Ha hoo I am fortunate enough to have a supportive, patient, and understanding family, without whom it would ha ve been awfully hard to make it this far, especially my father, Andrew C. John, and my stepfather, Neal Dunnigan. I also thank my lovely and loving fiance, Christi Barbee, for putting up with me throughout this year of dissertation data collection, analy sis, and writing. attention she deserves but I look forward to making up for it when life slows down and I can relax a bit, some fifty or sixty years from now. I acknowledge my doctoral committee members: Dr. Jay Hall, Dr. Scott Griffiths, Dr. Patricia Kricos, and Dr. Robin West. I am grateful for their support, time, and energy from the inception of this project to its completion. I am also grateful to Dr. Paul Duncan and Dr. Christine Sapienza for their service on my academic qualifying committee. I am indebted to my friends and colleagues at UF, particularly D rs. Brian and Nicole Kreisman and Dr. Kristin Johnston, for moral and material support throughout my d octoral program. I also owe a great deal to many people whose help in completing this project was indispensable: in addition to those named above, I acknowledge D rs. Michelle Colburn, Lisa Edmonds, and Mini Shrivastav of the University of Florida Departme nt of Communication Sciences and Disorders; Dr. Mary Anne Pinner of Hampton Oaks ENT; Mrs. Shirley Bloodworth of the Prime Time Institute; Mrs. Lynn Rousseau and Mrs. Nancy Day of the Hearing Loss Association of Gainesville; and Mrs. Elaine Brown of the Vi llage Retirement Community for

PAGE 5

5 assistance in recruitment of participants, as well as Dr. Lori Altmann for providing test materials and n ormative data for this project and Dr. Frank Musiek for allowing the GIN test materials to be reproduced in this dissert ation. Finally, I acknowledge the staff of the UF Department of Communication Sciences and Disorders: Ms. Idella King, Mrs. Debbie Butler, Ms. Cassie Mobley, Mrs. Addie Pons, Mr. David Fleming, and Mr. Neal Musson, without whose advice and assistance I m ight still be wandering lost in the basement of Dauer Hall.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 GUIDE TO ABBREVIATIONS ................................ ................................ ................................ .... 10 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 2 REVIEW OF LITERATURE ................................ ................................ ................................ 1 5 Hearing Impairment and Hearing Handicap ................................ ................................ ........... 15 Speech Perception ................................ ................................ ................................ ................... 16 Temporal Resolution ................................ ................................ ................................ .............. 22 Working Memory ................................ ................................ ................................ ................... 28 Summary ................................ ................................ ................................ ................................ 37 Objectives for the Current Research ................................ ................................ ....................... 38 Aims and Hypotheses ................................ ................................ ................................ ............. 38 3 METHODS ................................ ................................ ................................ ............................. 41 Participants ................................ ................................ ................................ ............................. 41 Cognitive Screening: Mini Mental State E xamination ................................ .......................... 42 Hearing Evaluation ................................ ................................ ................................ ................. 42 Aural Immittance Measurement. ................................ ................................ ............................ 43 Distortion Product Otoacoustic Emissions ................................ ................................ ............. 43 Pure Tone Audiometry ................................ ................................ ................................ ........... 47 Speech Audiometry ................................ ................................ ................................ ................ 52 Speech Perception in Noise Testing: Hearing In Noise Test (HINT) ................................ ... 53 Noise Competition ................................ ................................ ................................ .................. 55 Temporal Resolution Testing: Gaps In Noise (GIN) Test ................................ .................... 56 Working Memory Testing: Digit Span and Digit Ordering Tests ................................ ......... 58 Self Reported Hearing Handicap: Hearing Handicap Inventory for Adults (HHIA) or for the Elderly (HHIE) ................................ ................................ ................................ ........ 59 4 RESULTS ................................ ................................ ................................ ............................... 63 Participant Characteri stics ................................ ................................ ................................ ...... 63 Statistical Analysis ................................ ................................ ................................ .................. 63

PAGE 7

7 Distortion Product Otoacoustic Emissions (DPOAEs) ................................ .......................... 64 Audibility ................................ ................................ ................................ ................................ 65 Temporal Resolution ................................ ................................ ................................ .............. 67 Working Memory ................................ ................................ ................................ ................... 68 Collinearity Testing ................................ ................................ ................................ ................ 70 Speech Perception ................................ ................................ ................................ ................... 70 Hearing Handicap ................................ ................................ ................................ ................... 73 Gender ................................ ................................ ................................ ................................ ..... 76 Analysis of Variance I: Grouping by Age ................................ ................................ ............. 76 Multiple Regression, Grouping by Age ................................ ................................ .................. 78 HINT in Quiet Condition ................................ ................................ ................................ 78 HINT in Noise Condition ................................ ................................ ................................ 79 HHIA/E Emotional Subscale ................................ ................................ ........................... 79 HHIA/E Social/Situational Subscale ................................ ................................ ............... 80 HHIA/E Total ................................ ................................ ................................ .................. 80 Analysis of Variance II: Grouping by Hearing Loss ................................ ............................. 80 Multiple Regression, Grouping by Hearing Loss ................................ ................................ ... 82 HINT in Quiet Condition ................................ ................................ ................................ 82 HINT in Noise Condition ................................ ................................ ................................ 83 HHIA/E Emotional Subscale ................................ ................................ ........................... 83 HHIA/E Social/Sit uational Subscale ................................ ................................ ............... 83 HHIA/E Total ................................ ................................ ................................ .................. 83 Summary ................................ ................................ ................................ ................................ 84 5 DISCUSSION ................................ ................................ ................................ ....................... 110 Sampling Method ................................ ................................ ................................ .................. 111 Gaps in Noise (GIN) Test Findings ................................ ................................ ..................... 113 Worki ng Memory Findings ................................ ................................ ................................ .. 117 Speech Perception ................................ ................................ ................................ ................. 119 Hearing Handicap ................................ ................................ ................................ ................. 122 Limitations ................................ ................................ ................................ ............................ 125 Future Directions ................................ ................................ ................................ .................. 129 Summary ................................ ................................ ................................ ............................... 130 APPENDIX A I NFORMED CONSENT ................................ ................................ ................................ ...... 134 B GAPS IN NOISE TEST ................................ ................................ ................................ ........ 138 REFERENCE LIST ................................ ................................ ................................ ..................... 144 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 163

PAGE 8

8 LIST OF TABLES Table page 3 1 Summary of the University of Florida 65/55 Distortion Product Otoacoustic Emissions (DPOAE) protocol. ................................ ................................ ........................... 60 3 2 Assignment of DPOAE test points into high mid and low frequency groups. ............. 61 4 1 Comparison of pure tone audi ogram status to DPOAE status in each DPOAE frequency group. ................................ ................................ ................................ ................ 86 4 2 Mean audiometric data (N= 56). ................................ ................................ ....................... 87 4 3 Correlation matrix for age, audibility, and GIN variables (N = 56). ................................ 88 4 4 Comparison of participant mean Gaps In Noise threshold values to normative data collected by Musiek et al (2005) on 50 subjects (age 13 to 46 years). .............................. 89 4 5 Comparison of participant working memory scores to UF Language Over the Lifespan Laboratory normative data. ................................ ................................ ................. 90 4 6 Correlation matrix for GIN and working memory variables (N = 56). ............................. 91 4 7 Correlation matrix for HINT, HHIA/E, and audibility variables (N = 56). ....................... 92 4 8 Correlation matrix for GIN and HHIA/E variables (N = 56). ................................ ............ 93 4 9 ANOVA table for comparison of male and female participant data on age, audibility, GIN, working me mory, HINT, and HHIA/E variables. ................................ .................... 94 4 10 Mean audiometric data, separated by age group. ................................ ............................... 95 4 11 Comparison across participant a ge groups for descriptive factors, experimental measures, and dependent variables. ................................ ................................ ................... 96 4 12 ANOVA table for comparison of participant data, grouped by age, on age, audibility, working memory, H INT, and HHIA/E variables. ................................ .............................. 97 4 13 Mean audiometric data, separated by hearing loss group. ................................ ................. 98 4 14 Comparison across participan ts Group 2: BETAI < .98) on descriptive factors, experimental measures, and dependent variables. ................................ ................................ ................................ ........... 99 4 15 ANOVA table for comparison of participa nt data, grouped by hearing loss (Group 1: HINT, and HHIA/E scale scores. ................................ ................................ ..................... 100

PAGE 9

9 LIST OF FIGURES Figure page 3 1 Diagram of sound field HINT test environment. ................................ ............................... 62 4 1 Distribution of distortion product otoacoustic emissions results within low, middle, and high freq uency ranges. ................................ ................................ .............................. 101 4 2 Mean audiogram with standard deviation bars. ................................ ............................... 102 4 3 Comparison of Gaps In Noise test thresholds between r ight and left ears. ..................... 103 4 4 Comparison of participant mean Gaps In Noise threshold values (with standard deviation bars) to normative data collected by Musiek et al (2005) on 50 subjects (age 13 to 46 years). ................................ ................................ ................................ ......... 104 4 5 Mean audiogram for participants age 50 59 (N = 17). ................................ .................... 105 4 6 Mean audiogram for participants age 60 69 ( N = 19). ................................ .................... 106 4 7 Mean audiogram for participants age 70 89 (N = 20). ................................ .................... 107 4 8 ). ................................ ............ 108 4 9 Mean audiogram for HL Group 2 (BETAI < .98) (N = 32). ................................ ............ 109

PAGE 10

10 GUIDE TO ABBREVIATIONS AI Articulation Index BETAI Better ear Articulation Index scor e BETGIN Better ear Gaps in Noise test threshold BINAI Mean Articulation Index score between the left and right ears BINGIN Mean Gaps in Noise test threshold between the left and right ears daPa Deca pascals dB A Decibels A weighting scale dB HL Decibels hearing level dB SPL Decibels sound pressure level DIGBAK Digit span backward test DIGFWD Digit span forward test DIGORD Digit ordering test DPOAE Distortion product otoacoustic emissions GIN Gaps in Noise test HHIA Hearing Handicap Inven tory for Adults HHIE Hearing Handicap Inventory for the Elderly HHIEMOT Hearing Handicap Inventory for Adults / for the Elderly Emotional subscale score HHISOC Hearing Handicap Inventory for Adults / for the Elderly Social/situational subscale sc ore HHITOT Hearing Handicap Inventory for Adults / for the Elderly total score HINT Hearing in Noise Test HINTN Hearing in Noise Test threshold in noise condition

PAGE 11

11 HINTQ Hearing in Noise Test threshold in quiet condition Hz Hertz LAI Left ear Art iculation Index score LGIN Left ear Gaps in Noise test threshold MCL Most comfortable loudness level MMSE Mini Mental State Examination OAE Otoacoustic emissions PTA Pure tone average RAI Right ear Articulation Index score RGIN Right ear Gaps in Noise test threshold SNHL Sensorineural hearing loss SNR Signal to noise ratio SRT Speech reception threshold WM Composite working memory score WRS Word recognition score WRSAI Worse ear Articulation Index score WRSGIN Worse ear Gaps in No ise test threshold

PAGE 12

12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ROLE OF CLINICALLY APPLICABLE TEMPORAL RESOL UTION AND WORKING MEMORY TESTS IN PREDICTION OF SPEECH PERCEPTION AND HEARING HANDICAP By Andrew Barnabas John August 2007 Chair: James W. Hall III Major: Communication Sciences and Disorders Clinical audiometric measures can only partially account for perception performance in difficult listening situations, such as in noise. Speech perception in noisy situations is the primary complaint of most individuals seeking hearing health care. It is not surprising, therefore, that clinic al methods perform poorly in predicting self reported hearing handicap. Two factors that have been proposed to account for this disconnect between audiometric hearing loss and speech perception difficulty are auditory temporal resolution and working memor y. The role of these factors in speech perception performance in noise and in self report of hearing handicap were examined for a group of middle aged and older adults reporting a hearing loss. Experimental measures were chosen for their potential ease o f implementation into a clinical setting. Declines in hearing sensitivity, temporal resolution, and working memory were found to be significantly correlated with decreased speech perception ability and increased hearing handicap. Loss of hearing sensitiv ity was the strongest predictor of speech perception difficulty. When hearing sensitivity was controlled, temporal resolution threshold was only significantly related to speech perception for the oldest participants. Self

PAGE 13

13 reported hearing handicap was we akly associated with declines in hearing sensitivity, temporal resolution, and speech perception ability. The strongest predictor of hearing handicap was speech perception ability in noise. Previously unreported increases were seen in right ear / left ea r asymmetry of temporal resolution threshold independent of decline in hearing sensitivity. In addition, more than two thirds of participants were classified as having abnormal temporal processing according to published normative data. These findings sug gest a need for re evaluation of the temporal resolution task and establishment of age and hearing loss appropriate normative values.

PAGE 14

14 CHAPTER 1 INTRODUCTION The impact of hearing loss on the everyday listening situations of each individual is of primary interest to professionals in audiology and otolaryngology. While a diverse array of procedures exist to measure the function of the auditory system pure tone audiometry, otoacoustic emissions, and aural immittance measures, for example these tests often n eglect the complex and often highly individual factors that mediate the effect of hearing loss on daily function. The standard clinical audiologic test battery, including simple speech perception tasks and a handful of other primarily physiologic measures fails to address these factors. As a result, the same rehabilitation methods and treatments, including amplification selection, may be offered to two individuals who present with the same audiometric profile but have dissimilar cognitive and auditory pr ocessing abilities, listening experiences, and needs. The present study examined potential sources of the variance among individuals in the relation among hearing sensitivity, self report of hearing handicap, and speech perception performance in difficul t listening situations. Data from the study may help to define the individual factors that affect the experience of hearing loss for older adults. Findings may also lead to improved clinical strategies taking into account important factors that lay outsid e the current audiologic clinical protocol.

PAGE 15

15 CHAPTER 2 REVIEW OF LITERATURE Hearing Impairment and Hearing Handicap The disconnect in older adult populations between hearing impairment, determined using audiometry and other psychoacoustic measures, an d hearing handicap, as measured by personal experiences of difficulty in real world listening situations or by speech perception testing in difficult listening environments, is well established. Early examinations of the Hearing Handicap Inventory for the Elderly (HHIE), a commonly used clinical questionnaire, found that pure tone sensitivity and word recognition scores explained less than 50% of the variance in HHIE score (Weinstein & Ventry, 1983). Other studies have found even lower correlations betwee n audiometric and self report measures (Brainerd & Frankel, 1985; John & Kreisman, in preparation; Matthews, Lee, Mills, & Schum, 1990; McKenna, 1993; Newman, Jacobson, Hug, & Sandridge, 1997). These studies typically find that audiometric measures tend t o underestimate hearing handicap in individuals with more severe degrees of sensorineural hearing loss (SNHL), and that the relation between sensitivity and hearing handicap becomes more variable in older adults. The correlation between hearing impairmen t and self reported hearing handicap is influenced by several non auditory factors (Gatehouse, 1994, 1998; Gordon Salant, Lantz, & Fitzgibbons, 1994; Lutman, 1991; Lutman, Brown, & Coles, 1987; Stephens & Htu, 1991). Gatehouse (1998) found substantial va riability in the correlation between self reported hearing handicap and hearing loss, with significant effects of age and personality. Lutman and colleagues (1987) noted that age, gender, and socioeconomic status are factors in hearing self report, with m ales reporting more disability than females and older individuals reporting less disability than younger individuals. In another study, Lutman (1991) concluded that older

PAGE 16

16 individuals tend to overrate disability, consistent with the findings of Gordon Sala nt and colleagues (1994), who demonstrated that increased age correlated with increased disability as reported on the HHIE. Speech Perception Failure of the standard audiometric battery to fully account for the everyday listening difficulties reported by adults seeking treatment is likely due, at least in part, to the well established disconnect between peripheral hearing sensitivity, the primary focus of the typical audiologic evaluation, and speech perception in noisy, reverberant, or otherwise challengi ng listening situations. Indeed, the most common complaint from individuals with hearing loss seeking audiologic care is difficulty understanding speech in noisy environments, such as busy restaurants, parties, or large group conversations (CHABA, 1988; K im, Frisina, Mapes, Hickman, & Frisina, 2005; Koehnke & Besing, 2001). These auditory problems are reported particularly widely by older adults, who have been shown to demonstrate significantly greater difficulty understanding and recalling spoken message s compared to younger adults, particularly in noisy and reverberant environments, even when younger and older adults with similar degrees and configurations of hearing loss are compared (Pichora Fuller & Souza, 2003; Souza, Yueh, Sarubbi, & Loovis, 2000; T speech in difficult environments is the primary consequence of hearing loss, and is thus likely to be a primary influence in self report of hearing handicap and self referral for audiologic c are. Presumably, assessment of speech understanding in adverse listening conditions should be the primary focus of the audiologic test battery. However, speech perception in degraded listening conditions, such as in the presence of noise, is not well pre dicted by common audiometric measures. Like self reported hearing impairment, speech perception in noise ability correlates weakly with hearing sensitivity, with

PAGE 17

17 particularly low relations seen for individuals with more severe degrees of hearing loss (Cra ndell, 1991; Plomp & Mimpen, 1979; Smoorenburg, 1992). Some studies have shown that hearing sensitivity accounts for about half of the variance in speech perception in noise (Middelweerd, Festen, & Plomp, 1990; Plomp, 1986; Plomp & Mimpen, 1979; Smoorenbu rg, 1992). The balance of this variance is not, however, explained well by tests in the classic clinical battery. Tests of speech perception in quiet, commonly used as diagnostic tools, are poor predictors of speech perception in noise. Furthermore, spee ch perception scores in quiet do not add substantially to the predictive value of the audiogram a finding that is explained by the high correlation usually found between speech perception in quiet and sensitivity measured using pure tones (Divenyi & Haupt, 1997; Festen & Plomp, 1986; Glasberg & Moore, 1989; Plomp, 1986; Plomp & Mimpen, 1979; Smoorenburg, 1992; van Rooij & Plomp, 1990). Some researchers have suggested that speech perception deficits in individuals with SNHL can be explained by portions of th e speech signal being rendered inaudible by the decreased sensitivity that is a consequence of SNHL (Fabry & Van Tasell, 1986; Humes, Dirks, Bell, & Kincaid 1987; Humes, Espinoza Varas & Watson, 1988). The results of selected studies suggested that increa sing the audibility of the most difficult to perceive parts of speech eliminates deficits in speech perception. Humes et al (1994) compared the performance of 50 listeners between the ages of 63 and 83 on a number of speech recognition, auditory processin g, and cognitive tasks. In this study, hearing sensitivity accounted for 70 to 75% of the total variance in speech recognition performance. Auditory processing and cognitive function accounted for little or no additional variance. A study by Crandell and Needleman (1999) also found that individuals with normal hearing who were given a simulated hearing loss demonstrated similar speech recognition

PAGE 18

18 abilities to matched individuals with SNHL. However, the error patterns of the individuals with simulated hea ring loss differed from those of individuals with SNHL. While some studies find that hearing sensitivity predicts most of the variance in speech perception in noise (i.e. Humes et al 1987, 1994), others find that sensitivity is inadequate to explain varia nce in speech perception (Plomp, 1978; Plomp & Mimpen, 1979). Additional variables, such as age and temporal processing ability, consistently show a small but significant correlation with speech perception results even when hearing sensitivity is controll ed (Gordon Salant & Fitzgibbons, 1993, 1995; Halling & Humes, 2000). Overall, predictions of speech perception performance in older adults that are based on audibility tend to be accurate for speech perception in quiet conditions but overestimate performa nce in noise (Hargus & Gordon Salant, 1995; Pichora Fuller & Singh, 2006; Schum, Matthews, & Lee, 1991). Deficits in hearing sensitivity are also insufficient to explain other findings related to speech perception in noise. For example, individuals with S NHL require a greater signal to noise ratio (SNR) than individuals with normal hearing for perception of speech in noise. Stated another way, not only do individuals with SNHL require a higher intensity speech signal, these individuals also require that t he difference between that speech and any background noise be greater in order to optimally perceive speech. In one study illustrating this effect, Plyler and Hedrick (2002) administered a test of stop consonant identification to groups of individuals wit h normal hearing and with hearing loss. Even when the intensity of the stimulus phonemes was raised to a level at which all acoustic information was audible, a group of listeners with hearing loss still did not perform as well as listeners with normal hea ring. To examine the effects of SNHL that are not explained by loss of sensitivity, researchers often have used spectrally shaped masking noise to simulate threshold elevation, loudness

PAGE 19

19 recruitment, and loss of spectral resolution that results from SNHL (Crandell & Needleman, 1999; Humes, 1996; Humes & Jesteadt, 1991; Humes et al, 1988). Crandell and Needleman (1999) examined sentence recognition in a multitalker babble background for individuals with SNHL and for matched individuals with normal hearing who were given a simulated noise masked hearing loss. The simulated hearing loss was created by presenting spectrally shaped masking noise. When noise was present, the listeners with normal hearing demonstrated hearing thresholds similar to matched liste ners with SNHL. The individuals with simulated hearing loss performed significantly better than individuals with actual SNHL, suggesting that there was an effect of SNHL on speech perception that could not be simulated simply by raising threshold. However an earlier study by Humes and Roberts (1990) found that older adults with presbyacusis and young adults with matched simulated hearing loss performed similarly in tests of nonsense syllable recognition in noisy and reverberant listening conditions. It i s notable that, unlike the later study by Crandell and Needleman (1999) cited above, Humes and Roberts (1990) used closed set testing of nonsense syllable recognition, whereas Crandell and Needleman used an open set protocol requiring repetition of sentenc es, a more complex stimulus. The discrepancy in methodology and in findings between the two studies supports the contention that there is a component of suprathreshold distortion in SNHL for older adults that can not be explained by the audibility of the signal, and that the suprathreshold distortion is most evident with stimuli similar to those found in everyday listening situations (Crandell & Needleman, 1999; Plomp & Mimpen, 1979). Age is a factor in speech perception ability in degraded listening condi tions (Frisina & Frisina, 1997; Gordon Pichora Fuller, Schneider & Daneman, 1995). Harris and Reitz (1985) found that, in tests of

PAGE 20

20 word recognition, increases in reverberation caused a much greater performance decrease for older adults with nearly normal hearing than for young individuals with normal hearing. These two groups performed similarly in quiet, in noise alone, and in reverberation alone; the group difference appeared when no ise and reverberation were combined. Gordon Salant and Fitzgibbons (1993, 1995) also found that, on tests of speech stimuli degraded by reverberation and time compression, both age and hearing loss were significant predictors, and that age became a greate age on perception of vowels degraded by noise and reverberation. In addition to environmental distortion, lengthy and more everyday speech like stimuli appear to eli cit age effects in speech perception. Jerger (1973) examined word recognition in a retrospective clinical study of 2,162 patients and found that, when hearing sensitivity was controlled, word recognition declines with age were consistent in individuals wi th moderate to severe hearing loss, but not for individuals with mild hearing loss. Similarly, Bergman (1980) found that for a given threshold level, older individuals performed more poorly on speech perception tasks than younger individuals. In another study by Jerger (1992), perception of sentence stimuli was characterized by more robust age effects than perception of word stimuli, a finding reported also by Jerger and Hayes (1977). In general, studies of aging and speech perception show that increase s in the difficulty of the speech perception task (such as longer stimuli and more adverse listening conditions) tend to elicit more robust age effects (Gordon Salant & Fitzgibbons, 1993, 1995; Harris & Reitz, 1985; Jerger, Jerger, & Pirozollo, 1991; Weins tein, 2000). The investigation of auditory function for young versus elderly listeners presents significant methodologic challenges. Older adults with thresholds comparable to young adults

PAGE 21

21 are often difficult to locate and, even if found and enrolled in a study, older adults with good confound some research designs Not all age effects can be attributed to small threshold differences. A study by Helfer and Wilber (1990) found a strong negative correlation between age and nonsense syllable perception in reverb In this study, in a comparison of older and younger individuals with hearing loss, older individuals performed slightly worse despite actually having better thresholds than younger counterparts. As the degree of distortion increased, elderly listeners with mild hearing loss performed less and less like young listeners with normal hearing, a finding confirmed by The reasons for age related deficits in speech perception are not well understood. Researchers have hypothesized that age related deficits in speech perception are related to peripheral pathology (Fabry & Van Tasell, 1986; Humes et al, 1987 ; Humes et al, 1988), to central auditory deficits (Jerger, Jerger, Oliver, & Pirozzolo, 1989; Jerger, Oliver, & Pirozzolo, 1990; Ochs, 1990; Stach, Jerger, & Fleming, 1985; Wingfield, 1996), and/or to cognitive deficits (Kjellberg, 2004; Letowski & Poch, 1996; Pichora Fuller et al, 1995; Rabbitt, 1991; Smoorenburg, 1992). More recently, Pichora Fuller and Singh (2006) integrated these hypotheses into a unified view of auditory processing that takes into account multiple etiologies and pathologies, and foc uses on the interaction among peripheral hearing sensitivity, central auditory processing, and cognitive processing.

PAGE 22

22 Within both the etiological and unified models of auditory processing, several potential sources of suprathreshold distortion and diminis hed speech perception have been identified. The present investigation will examine two factors that have received recent attention in the hearing science literature: auditory temporal resolution and working memory capacity. Temporal Resolution Auditory t emporal resolution refers to the ability of an individual to detect differences in the duration of auditory stimuli and in the time intervals between auditory stimuli over time. Silent intervals and rapid modulations of intensity are important for speech understanding, providing cues at the phoneme, word, and sentence level, Examples of silent intervals include transitions between phonemes, voicing cues, cues signaling the end of a word, phrase, or sentence, and prosodic cues (Dorman, Marton, & Hannley, 19 85; Greenberg, 1996; Pichora Fuller & Singh, 2006; Rosen, 1992; Schneider & Pichora Fuller, 2001). Several studies have noted an association between temporal resolution ability and perception of degraded or distorted speech (Bergman, 1980; Divenyi & Simon 1999; Gordon Salant & Fitzgibbons, 1999; Irwin & MacAuley, 1987; Pichora Fuller, Schneider, MacDonald, Pass, & Brown, 2007; Pichora Fuller & Souza, 2003; Schneider & Pichora Fuller, 2000, 2001; Snell, Mapes, Hickman, & Frisina, 2002; Tyler, Summerfield, Wood, & Fernandes, 1982). Although temporal resolution ability is highly variable among individuals with SNHL, listeners with SNHL have generally poorer temporal resolution ability than individuals with normal hearing (Arlinger & Dryselius, 1990; Buus & Florentine, 1985; Madden & Feth, 1992; Moore, Glasberg, Donaldson, McPherson, & Plack, 1989). Temporal resolution ability can be measured using several methods, including modulation detection, perception of compressed speech, and perception of interrupte d speech. However, temporal resolution is most commonly tested using gap detection measures.

PAGE 23

23 Auditory gap detection measures test the ability of a listener to detect the presence of a temporal gap (silent period created by an absence of energy) within a tone or, more commonly, noise burst. Performance on a gap detection measure is assumed to be related to speech perception because silent gaps in speech signals can serve as phonetic cues. For example, a silent gap is a cue for the presence of the stop co nsonant phoneme /p/ in the word "spoon," in contrast to the word "soon" in which there is no gap. Because psychoacoustic gap detection thresholds measure a listener's ability to perceive the shortest detectable gap in a sound signal and the detection of s uch gaps is relevant to speech perception, it is logical to assume that a strong relationship exists between measures of gap detection threshold and speech perception ability (Pichora Fuller, Schneider, Benson, Hamstra, & Storzer, 2006). There is experime ntal evidence of significant correlations between measures of gap detection in noise or tone stimuli and word identification in noise (Snell et al, 2002; Tyler et al, 1982) and reverberation (Gordon Salant & Fitzgibbons, 1993). Gap detection ability is gen erally poorly predicted by the audiogram, as individuals with similar degrees and configurations of hearing impairment often show widely varying ability to perceive temporal gaps (Glasberg, Moore, & Bacon, 1987; Moore & Glasberg, 1988; Schneider, Pichora F uller, Kowalchuk, & Lamb, 1994; Schneider, Speranza, & Pichora Fuller, 1998). Several studies have also found gap detection ability to be independent of frequency selectivity of auditory filters (Eddins, Hall, & Grose, 1992; Grose, Eddins, & Hall, 1989; M oore, Peters, & Glasberg, 1993). The correlation between gap detection threshold and speech perception is not clear, even when accounting for hearing sensitivity and spectral resolution. While several studies have noted an association between temporal res olution ability when controlling for audibility

PAGE 24

24 (Glasberg & Moore, 1989; Glasberg et al, 1987; Haubert & Pichora Fuller, 1999; Irwin & McAuley, 1987; Phillips, Gordon Salant, Fitzgibbons, & Yeni Komshian, 2000; Snell et al, 2002; Souza, Boike, Withrell & T remblay, 2007; Tyler et al, 1982) others have failed to find such an association (Dubno & Dirks, 1990; Festen & Plomp, 1983; Snell & Frisina, 2000; Strouse, Ashmead, Ohde, & Grantham, 1998; van Rooij & Plomp, 1991). The discrepancy in findings among stu dies may be due to differences in the stimulus and procedure used for gap detection testing. Gap thresholds are affected by characteristics of the carrier stimulus (the tone or noise burst that contains the gap) including duration, intensity, and spectrum of the carrier (Fitzgibbons, 1983; Fitzgibbons & Gordon Salant, 1987; Jesteadt et al, 1982; Pichora Fuller et al, 2006). Several investigations demonstrated a significant influence of carrier spectrum on gap detection. A decrease in gap detection thres hold has been observed with increases in signal frequency for groups of individuals with normal hearing and groups with SNHL (Fitzgibbons & Gordon Salant, 1987). The faster response of more broadly tuned high frequency auditory filters may account for thi s increase in temporal acuity at higher frequency, but according to some researchers gap detection threshold is independent of frequency within certain ranges, including 400 and 2000 Hz (Moore et al, 1989) and 500 and 4000 (Formby & Forrest, 1991). The re ason for frequency independence within certain ranges is not well understood, but the existence of that independence may help to explain why gap detection studies using tonal stimuli of varying spectral content and studies using noise stimuli often differ in findings. There is also evidence that within channel gap detection (detection of gaps bordered by like carriers, such as identical tones or noise bursts) is a distinct process from between channel gap detection (detection of gaps bordered by different c arriers, such as tones at different

PAGE 25

25 frequencies) (Divenyi & Danner, 1977; Formby, Gerber, Sherlock, & Magder, 1998; Grose, Hall, Buss, & Hatch, 2001; Heinrich & Schneider, 2006; Phillips & Hall, 2000, 2002; Phillips & Smith, 2004; Pichora Fuller et al, 200 6). The between /within channel discrepancy is most pronounced in older adults (Fitzgibbons & Gordon Salant, 1994, 1995; Grose, Hall, & Buss, 2001; Lister, Koehnke, & Besing, 2000; Lister & Tarver, 2004). Age effects for across channel gap detection may be related to difficulties experienced by older adults in resolving temporal envelope information and integrating that information across channels (Souza et al, 2007). Gap detection can also be confounded by other factors, including contribution of off f requency listening regions, placement of the gap within the carrier (i.e. midpoint of the carrier vs. near stimulus onset or offset), and presentation and response mode (He, Horwitz, Dubno, & Mills, 1999; Pichora Fuller et al, 2006; Schneider & Hamstra, 19 99). In addition, the association between gap detection ability and speech perception is affected by speech perception materials and presentation method (Lister, Roberts, Shackelford, & Rogers, 2006; Ochs, 1990; Snell et al, 2002). Generally, more comple x speech perception stimuli elicit stronger temporal resolution influences. Several gap detection studies have demonstrated that older adults do not detect a gap in the signal until the size of the gap is about twice as large as the smallest gap detectable by younger adults (about 6 vs. 3 ms in tonal stimuli) with gap detection threshold poorly predicted by pure tone hearing thresholds in listeners with good audiograms (Gordon Salant & Fitzgibbons, 1993; Haubert & Pichora Fuller, 1999; Schneider et al, 1994 ; Snell, 1997; Strouse et al, 1998). Also, as listeners age they experience diminished ability to resolve simple temporal cues in speech, a finding which may partially explain age effects on speech perception not explained fully by loss of sensitivity (Fi tzgibbons & Gordon Salant, 1994; Gordon Salant & Fitzgibbons, 1993; Snell &

PAGE 26

26 Frisina, 2000; Strouse et al, 1998). Age effects on temporal processing appear more consistent and strong when measures are made with complex stimuli in difficult listening condit ions (Fitzgibbons & Gordon Salant, 1996; Fuller et al, 2006). Deficits in temporal processing may be evident as early as middle age (Grose, Hall, & Buss, 2001, 2006). In order to minimize the confounds of stimulus frequency, task complexity, and age in temporal resolution testing, researchers at the University of Connecticut recently developed a new temporal resolution test called the Gaps In Noise (GIN) test. The GIN consists of series of 6 second segments of broadband (white) noise containing between zero and three silen t gaps per segment. These gaps range in length from two to 20 ms. For each set of noise segments, a total of 60 gaps are presented with six gaps each of 2, 3, 4, 5, 6, 8, 10, 12, 15, and 20 ms. Listeners are instructed to press a response button wheneve r a gap is detected. When administration of the GIN procedure is complete, a gap threshold is calculated (Musiek, Shinn, Jirsa, Bamiou, Baran, & Zaidan, 2005). The GIN procedure has several potential advantages over other temporal resolution tests. Firs t, it can be administered with conventional clinical instrumentation. That is, the GIN test is recorded to a compact disc and requires only a CD player with standard audio outputs, a two channel audiometer, and a response button, thus eliminating an obsta cle to clinical assessment of temporal resolution ability and the reason why gap detection is infrequently performed by audiologists (Emanuel, 2002). Second, as noted above, assessment of gap detection can be frequency and age dependent. The GIN attempt s to minimize these effects by using broadband noise, which is presumed by the authors to be unaffected by hearing for any single frequency because the noise contains energy over a range of frequencies. Broadband noise stimuli have the additional advantag e of masking the spectral splatter heard when a tonal or narrow band signal is

PAGE 27

27 rapidly switched off. The unwanted spectral energy could provide a cue for a gap that otherwise would not be detected (Trehub, Schneider, & Henderson, 1995). Finally, the simpl e response method (pressing a button when a gap is heard, rather than verbally responding or counting gaps) is designed to place minimal cognitive and linguistic/verbal demands on the listener, a feature of the GIN test which may reduce age effects (Jerger & Musiek, 2000). Third, because it requires the listener to respond to the presence of a silent interval, the GIN is a true test of gap detection. The two most common purported gap detection measures, the Random Gap Detection Test (RGDT; Keith, 2000) and the Auditory Fusion Test Revised (AFT R; McCroskey & Keith, 1996) are actually auditory fusion tests. Auditory fusion tasks require the listener to determine the point at which two stimuli (tones, noise bursts) fuse into one stimulus. The fusion thresho ld, therefore, is the smallest gap between two stimuli that allows the listener to perceive the stimuli as distin ct from one another (Lister, Be sing, & Koehnke, 2002). Auditory fusion could be considered the inverse of gap detection. The two terms are of ten used interchangeably to describe the same function (Keith, 2000); however, there is neurophysiologic evidence that the two abilities may have different underlying processes (Chermak & Lee, 2005). Finally, the GIN test is designed as a monaural auditor y measure and, thus, provides laterality information about auditory function. Preliminary studies of the GIN test have shown good reliability and sensitivity when administered to individuals with normal hearing and individuals with CANS lesions (Chermak & Lee, 2005; Musiek et al, 2005). Temporal resolution ability is usually enhanced at higher presentation levels for listeners with normal hearing (e.g. Fitzgibbons & Gordon Salant, 1987; Fitzgibbons, 1983; Jesteadt et al, 1982). In contrast, recent researc h suggests that the GIN test is insensitive to presentation level within a range of approximately 35 50 dB SL re: threshold for GIN test noise (Weihing, Musiek, & Shinn, 2007).

PAGE 28

28 Working Memory Cognitive performance is a possible explanation for suprathresho ld speech perception deficits. However, research into this relation between cognitive function and speech perception has produced equivocal results, with some studies demonstrating a positive correlation between cognitive abilities and speech perception ( Kjellberg, 2004; Pichora Fuller & Singh, 2006; Pichora Fuller et al, 1995; Rabbitt, 1991; Smoorenburg, 1992; van Rooij & Plomp, 1990; Wingfield, 1996) but others finding no such association (Humes & Floyd, 2005; Humes, Watson, Christensen, Cokely, Halling, & Lee, 1994; Stach, Loiselle, & Jerger 1991; Stach et al, 1985; van Rooij, Plomp, & Orlebeke, 1989). This inconsistency is most likely due to the broad range of methods used by hearing researchers to assess cognition and speech perception. For example, in two studies examining the association between speech perception and cognitive decline in older adult subjects, van Rooij and Plomp (1990) found that one third of the variance in an adaptive sentence recognition task was related to cognitive status (meas ured by a combination of speed of processing, memory, and intellectual ability). On the other hand, Jerger et al (1991) found that cognitive status, as measured by a speed of processing task only, accounted for between 6 percent (for word recognition) and 14 percent (for dichotic sentence recognition) of variance. There were differences in the two studies in the speech stimulus used to measure auditory processing and in the battery used to assess cognitive status. There is a bi directional relation betwe en peripheral processing status (such as hearing sensiti vity) and cognitive processing. Declines in peripheral sensory status such as in loss of hearing sensitivity, may provide a distorted and impoverished signal for later states of cognitive processing which may lead to declines in cognitive status (Appollonio, Carabellese, Magni, Frattola, & Trabucci, 1995; Gennis, Garry, Haaland, Yeo, & Goodwin, 1991; Lindenberger & Baltes, 1994; Ryan, Giles, Bartoloucci, & Henwood, 1986; Sands & Meredith, 1989; Uhlm ann,

PAGE 29

29 Larson, Rees, Koepsell, & Duckert, 1989; van Boxtel, van Beijsterveldt, Houx, Anteunis, Metsemakers, & Jolles, 2000). In addition, age related declines in cognitive processes may affect higher order hearing processes such as speech perception, even when peripheral hearing status is unimpaired (Gordon Salant, 2005; Gordon Salant & Fitzgibbons, 1997; Helfer & Wilber, 1990; Lunner, 2003; Pichora Fuller et al, 1995; Rabbitt, 1991). However, the particular aspects of cognitive status and the mechanisms b y which cognitive declines affect speech perception are still a topic of debate. Some studies have isolated individual components of cognitive function that appear to play a role in speech perception, including attention (Alain & Woods, 1999; Cameron, Dill on, & Newall, 2006; Noble & Gatehouse, 2006; Rakerd, Hartmann, & Hsu, 2000; Snell et al, 2002), speed of processing (Brebion, 2001; Fitzgibbons & Gordon Salant, 1996; Gordon Salant & Fitzgibbons, 1993, 2001; Strayer, Wichkens, & Braune, 1987; Wingfield, 19 96), and memory (Gordon Salant & Fitzgibbons, 1997; Kemper, 1992; Lunner, 2003; McCoy, Tun, Cox, Colangelo, Stewart, & Wingfield, 2005; Pichora Fuller & Singh, 2006; Verhaegen, 1999; Wingfield & Tun, 2001). A full analysis of all potential cognitive facto rs in speech perception is beyond the scope of the present investigation. However, working memory is a component that has received particular attention in recent research, and appears to hold significant potential for further study within the field of hear ing science. Working memory, therefore, will be the primary cognitive focus of the present study. Working memory refers to the processes involved in temporary storage and processing of information used for complex cognitive tasks. The term first appeared in the psychology literature in the 1960s. Baddeley and Hitch (1974).proposed the most prominent accepted definition, which described working memory as a system with three components: (1) the central

PAGE 30

30 executive, which could be considered an attention cont rolling system regulating the allotment of manipulates and processes visual images; and (3) the phonological loop, which stores and rehearses verbal information Working memory is an active process by which data is maintained (Baddeley, 1992; Baddeley & Hitch, 1974). y or memory span that is usually Miller (1956) found that young adults could hold an average of seven digits in memory at once. Later research has clarified tha t the number of items in working memory is dependent upon the nature of the stimulus items, such as whether they are numbers, letters, or words; the length and complexity of the items (i.e., short versus long words); and whether the stimuli are familiar to the individual tested (Hulme, Roodenrys, Brown, & Mercer, 1995). Working memory status is commonly measured using simple tests such as digit span and digit ordering. In a digit span paradigm, an individual is presented verbally with a series of single di git numbers and is asked to repeat them in the same sequence they were presented (forward digit span) or in reverse of the sequence they were presented (backward digit span). Digit span is available as a subtest of some intelligence scales, including all versions of the Wechsler Adult Intelligence Scale (WAIS III; The Psychological Corporation, 1997). The digit ordering test is conducted similarly to digit span, but requires the tested individual not only to retain series of numbers in memory and reproduc e them, but to reproduce them in ascending numerical order (Cooper, Sager, Jordan, Harvey, & Sullivan, 1991). For example, the tester 1 8 2 4 2 4 7 8

PAGE 31

31 Other working memory tasks include reading span, which requires individuals to read increasingly longer sets of sentences out loud while recalling the final word of each sentence in each set, n back recall, in which individuals are asked to track the identity or location of a series of verbal or nonverbal stimuli and to indicate whether a stimulus is the same as the one presented n trials previously, and pattern or sequence recall tasks, which require an individual to monitor and replicate a pattern of visual or auditory stimuli. Wor king memory capacity declines over the course of the lifespan, with steepest declines seen in older adults after age 80 (Carpenter, Miyake, & Just, 1994; DeDe, Caplan, Kemtes, & Walters, 2004; Foos & Wright, 1992; Gilinsky & Judd, 1994; Hultsch, Hertzog, S mall, McDonald Miszczak, & Dixon, 1992; Norman, Kemper, & Kynette, 1992; Palladino & De Beni, 1999; Salthouse, 1991; Wingfield, Stine, Lahar, & Aberdeen, 1988). Working memory loss is a significant factor in the overall cognitive deterioration observed in older adults (Carpenter et al, 1994; Moscovitch and Winocur, 1992; Salthouse, 1994). Speech is a complex, rapid acoustic signal. Acoustic and linguistic information in a speech signal is processed, on average, at a rate between 200 and 270 words per minu te, with substantial variability in speaking rate both between and within talkers (Stine, Wingfield, & Poon, 1986; Vaughan, Storzbach, & Furukawa, 2006; Wingfield, Poon, Lombardi, & Lowe, 1985). Because speech is received at such a high rate, listeners ar e required to store and process incoming information very quickly, a function that is diminished in older adults compared to younger adults independent of the effects of peripheral hearing loss (Gordon Salant and Fitzgibbons, 2001; Stine & Wingfield, 1987; Vaughan & Letowski, 1997; Wingfield et al, 1985). Still, for most older adults, perception in ideal conditions is generally unaffected, as information is easily accessed and processed and little strain is placed on the cognitive resources

PAGE 32

32 available to t he individual for processing. Reduced working memory capacity can be overloaded when information is more difficult to perceive, when distractors are present, or when efficiency of data transmission is reduced (Wingfield, 1996). A working memory overload might occur for individuals listening in difficult environments, such as those with high levels of noise. Noise can cause speech signals to be distorted, reducing the quality and quantity of the information received by the listener. If a listener has re duced ability to process acoustic information, perception of the speech signal is likely to be slow and inefficient and information may be lost (Kemper, 1992; Ryan et al, 1986). Cohen (1987) hypothesizes that lengthy messages show the most loss in noisy s ituations, because messages require storage of data and, also, the ability to relate early and late components of the message to comprehend it in entirety. In addition to increased strain on working memory capacity, reallocation of cognitive resources to compensate for the inability of working memory to handle the incoming signal can cause a cascade effect in which higher processing is negatively affected (Larsby, Hallgren, Lyxell, & Arlinger, 2005; Pichora Fuller et al, 1995; Tun et al, 2002). Some resea rchers have suggested that working memory deficits appear when speed of processing slows with aging, causing reallocation of resources to compensate and process rapidly changing signals, (Fry & Hale, 2000; Salthouse, 1992). Working memory may be a factor i n age related deficits in speech perception (Gordon Salant & Fitzgibbons, 1997; Kjellberg, 2004; Pichora Fuller et al, 1995; Wingfield & Tun, 2001). Gordon Salant and Fitzgibbons (1997) found that perception of sentence length stimuli in noise showed stro nger age effects (with older listeners performing significantly more poorly) compared to shorter (such as word length) stimuli. Altering a recall task to increase the load on memory during speech perception also had a greater detrimental effect on older l isteners

PAGE 33

33 compared to younger listeners. The findings reported by Gordon Salant and Fitzgibbons (1997) support a role of memory capacity in speech perception in difficult environments. Other studies relating a loss of some cognitive ability to speech perc eption difficulty can also be explained in terms of working memory. For example, in a study by van Rooij and Plomp (1990), one third of the variance in speech recognition ability by older adult listeners could be explained by cognitive factors such as pro cessing efficiency and memory capacity. This study used sentence length stimuli presented in a background of noise and required listeners to retain and repeat each sentence in its entirety. Kjellberg (2004) hypothesized a model for the effect of memory lo ad on speech perception in difficult situations, and suggested that listening in reverberant and noisy environments may affect storage in working memory, as more cognitive resources are required for phonological processing when distortion of speech is pres ent. Even when speech is understood, less information may be stored in memory, contributing to decreased understanding of speech over time and increased cognitive fatigue. A general cognitive decline, therefore, might manifest as difficulty in speech per ception in adverse conditions, as fewer cognitive resources become available for processing. Lyxell and colleagues (2004) suggested that when components of the auditory signal are either altered or lost due to hearing loss or environmental distortion, mor e strain is placed on cognitive strategies such as making verbal inferences and disambiguating manipulate, process, and store information quickly and over sh ort periods of time. A large working memory capacity enables the individual to use more contextual information in the larger speech signal to aid in disambiguation and inference. Thus, at least in theory, cognitive strategies may be more precise and limi ted to the topic of the conversation.

PAGE 34

34 The cognitive hypotheses are supported also by Neils and colleagues (1991) who reported that a group of older adults with mean age of 80, but not a younger group of older adults (mean age of 70), performed significantl y worse than a group of young adults (mean age of 25) on an auditory memory task (remembering sequences of tones), particularly for long sequences and with short intervals between tones. The performance deficit was not related to hearing sensitivity, sugge sting age related deficits beyond loss of hearing sensitivity and involving memory and rapid processing function. In fact, an important feature of working memory from the perspective of the audiologist or hearing researcher is that declines in working mem ory appear to be independent of changes in hearing status (Lyxell, Andersson, Borg, & Ohlsson, 2004). To date, the few studies that have examined closely the role of working memory in difficult speech perception have produced mixed results. While several studies provide evidence for a role of working memory capacity in perception of difficult or degraded speech signals, others have failed to find an association. Discrepancies in findings are likely a result of methodological differences in testing both wo rking memory capacity and speech perception ability (i.e. Gordon Salant & Fitzgibbons, 1997; Humes & Floyd, 2005; Lunner, 2003; Pichora Fuller & Singh, 2006; Pichora Fuller et al, 1995; Tun, Wingfield, & Stine, 1991; Vaughan et al, 2006). For example, Hum es and Floyd (2005) assessed working memory and sequence learning using a modified version of the Simon memory game, with participants measured on their ability to follow audiovisual cues and repeat a tapping pattern on the game panels. Speech perception was measured using nonsense syllable repetition and the Connected Speech Test (CST) in backgrounds of noise. The authors reported that an association between performance on the

PAGE 35

35 ertain memory tasks were moderately and significantly correlated with speech perception performance. In contrast, Vaughan and colleagues (2006) measured verbal working memory using a back recall and digit span. Speech perception was evaluated using time compressed IEEE and anomalous sentences. The IEEE sentences are semantically and syntactically correct but have low predictability. The anomalous sentences were designed to be similar in length and phonemic content to the IEEE identified a significant correlation between working memory performance and speech perception. The methodological differenc es are clear between these two studies of working memory and speech perception. These investigations were substantially different in their measures of both working memory (visual Simon game versus verbal digit span and word recall) and speech perception ( nonsense syllables and connected speech versus time compressed low context and low semantic content sentences). These studies also differ in age grouping criteria. Other inv estigations of working memory and speech perception have similar conflicts in me thodology. In hearing research literature, working memory is identified variously using measures such as digit span, word recall, sentence recall, subject/object pointing, digit symbol, and visual pattern recognition and repetition. Some of these tasks d o not appear to measure working memory effectively and independently of other cognitive functions, while other tasks are clearly tests of visual, and not verbal, working memory capacity (for example, the Simon game). Furthermore, speech perception measure s in the studies vary from identification at the syllable level to perception and repetition of connected sentence length speech stimuli, from

PAGE 36

36 listening in quiet to listening in noisy or reverberant environments, and from monaural headphone presentation to listening in diffuse sound fields. Working memory effects should, in theory, be most apparent for difficult, lengthy speech stimuli and/or speech stimuli presented in difficult listening situations, such as in the presence of noise. It would not be surp rising, therefore, to find little or even no role for memory capacity in recall of simple stimuli, such as nonsense syllables or monosyllabic words, or when the auditory processing system is otherwise insufficiently taxed by the repetition task. As a resul t of the substantial methodological differences among investigations, it is difficult to be certain of the magnitude of the impact of working memory on speech perception. While several studies have revealed compelling evidence of a role for cognitive fact ors, including memory, in speech perception, laboratory findings do not consistently support a strong association. It is likely that working memory capacity does play a role, at least in difficult listening situations for adults wit h impaired memory capac ity and for lengthy, complex speech signals message. However, it is less clear whether working memory effects influence speech perception for adults who do not have significant cognitive impairment. For example, working memory might play a role in degraded speech perception for adults who have slight memory declines as a result of aging or who have age normal memory abilities but rely upon them more heavily as a result of peripheral sensory impairment (i.e. sensorineural hearing loss). If there is a role for working memory in adults who are generally cognitively intact, use of a simple working memory battery (such as the digit span and digit ordering measures de scribed above) may help to account for the

PAGE 37

37 commonly seen speech perception difficulties which are not fully explained by loss of hearing sensitivity. Summary In summary, while audibility appears to be a primary explanatory factor for everyday listening dif ficulty in adults with hearing loss other factors clearly play an important role in speech perception In fact, speech perception predictions based on audibility perform poorly for complex speech perception tasks and simulations of real world listening, such as inclusion of competing noise signals and other environmental distortions and distractions. Peripheral central auditory, and cognitive supra threshold distortion factors account for a substantial portion of the variance in speech perception in dif ficult listening situations. The peripheral/central factor of temporal resolution and the cognitive factor of working memory capacity are two such supra threshold determinants. Each factor has been demonstrated to play some role in degraded speech percep tion ability and, significantly, each can be assessed using measures that (1) are simple to administer with minimal training; (2) do not require specialized personnel or equipment beyond that found in a standard audiometric test suite; and (3) are widely a vailable and accepted as valid measurement tools. With these considerations in mind, the purpose of the present investigation is to evaluate the predictive ability and clinical efficacy of a temporal resolution measure, the Gaps In Noise test, and a work ing memory measure, the digit span and digit ordering battery. The population to be investigated predominates among those persons self re ferring for hearing health care: adults ranging from late middle age to older adulthood who are aware of a hearing lo ss that interferes with daily communication activity. A group of fifty six adults in this population served as participants for the investigation.

PAGE 38

38 Objectives for the Current Research The impetus for the present research is the well established disconnect between measures of hearing impairment (i.e., diagnostic audiometry) and measures of hearing handicap (i.e., speech perception in noise) for adults experiencing hearing loss. With that disconnect in mind, the goal of this study is to examine potential fac tors in hearing handicap, defined as the consequences and impact of hearing loss on daily living, that are not currently measured in a typical audiologic clinical protocol. While no combination of tests and questionnaires is likely to account fully for ea over current clinical methods, which insufficiently address the problems experienced by individuals in daily communication. Two factors that may play a role in speech perception a bility, and for which relations to communication function have been empirically demonstrated, are auditory temporal resolution ability and working memory capacity. The present study will incorporate clinically feasible tests for each factor into an audiolo gic test battery to determine what influence, if any, the factors have on hearing handicap above the influence of hearing sensitivity, and thus what benefit assessment of temporal resolution and working memory capacity might provide in clinical diagnosis a nd rehabilitation planning. Aims and Hypotheses AIM 1: The investigator will evaluate the contributions of audibility, temporal resolution threshold, and working memory capacity in prediction of speech perception loss in quiet and in noisy environments. HYPOTHESIS 1a: It is hypothesized that audibility, as measured by peripheral hearing sensitivity, is significantly correlated with speech perception ability in quiet. Consistent with previous studies, it is hypothesized that neither temporal resolution, as measured by

PAGE 39

39 a gap detection task; nor working memory, as measured by a digit span and ordering battery, will be correlated significantly with speech perception ability in quiet when controlling for audibility. HYPOTHESIS 1b: It is hypothesized that au dibility, as measured by peripheral hearing sensitivity; temporal resolution, as measured by a gap detection task; and working memory, as measured by a digit span and ordering battery, will be significantly correlated with speech perception ability in noi se. AIM 2: The investigator will determine whether the well established discrepancy between objectively evaluated hearing impairment and subjective, self reported hearing handicap can be partially explained by the supra threshold distortions described ab ove and their impact on listening in difficult environments. HYPOTHESIS 2: It is hypothesized that audibility, as measured by peripheral hearing sensitivity; temporal resolution, as measured by a gap detection task; and working memory, as measured by a d igit span and ordering battery, will be significantly correlated with self reported hearing handicap, as measured by composite score on the Hearing Handicap Inventory for Adults (HHIA) or the Hearing Handicap Inventory for the Elderly (HHIE). AIM 3: In or der to evaluate the efficacy of the experimental measures in a clinical setting, the investigator will determine the degree to which temporal resolution threshold and/or working memory capacity is predictive of speech perception difficulty in noise and/or self reported hearing handicap above the variance which is predicted by audibility. HYPOTHESIS 3: It is hypothesized that, when controlling for audibility, temporal resolution

PAGE 40

40 threshold and/or working memory capacity will remain significantly correlated with speech perception ability in noise.

PAGE 41

41 CHAPTER 3 METHODS Participants Fifty six adults (17 male, 39 female) served as participants for this investigation. Participants ranged in age from 50 to 89 years of age with a mean age of 66.4. All participants were recruited from the Alachua county area by the investigator. To be eligible to participate in this study, participants met the following inclusion criteria: 1. Age 50 or older at the time of testing; 2. Aware of a hearing problem that interferes with comm unication; 3. Generally good health, with no condition preventing the participant from completing study measures; and 4. English as primary language as reported by the participant. All inclusion criteria were satisfied by responses to questions from the inve stigator during a case history interview prior to the beginning of data collection or, if possible, during recruitment. Three potential participants who responded to advertisements for this investigation subsequently revealed that they did not believe the y had a hearing loss. These individuals were excluded from this investigation, but were given a brief hearing test (pure tone air conduction threshold testing). Participants were not enrolled in the study if they met any of the following exclusion criter ia: 1. Prior diagnosis of dementia or mild cognitive impairment, by history; 2. Transient hearing loss, as ascertained during the pre 3. Conductive hearing loss, as ascertained using aural immittance testin g and, where indicated, bone conduction audiometry; 4. Current use of hearing aids; and 5. Inability or unwillingness to complete all study measures.

PAGE 42

42 All exclusion criteria were satisfied by responses to questions from the investigator during a case history interview prior to the beginning of data collection or, if possible, during the recruitment process. Prior use of hearing aids, defined as a period of hearing aid use ending at least one year prior to recruitment, was not considered an exclusion criterion Four participants reported prior use of hearing aids, but none had used them within the last five years. Three potential participants were excluded from this investigation on the basis of presence of conductive hearing loss. These participants were r ecommended to an otolaryngologist for evaluation and possible treatment. Prior to participating in the study, each participant was required to read carefully and sign an Informed Consent document approved by the University of Florida Institutional Review B oard (IRB 02). This investigation was approved in September 2006 as UF protocol 2006 U 0754. The informed consent form can be found in Appendix A. Cognitive Screening: Mini Mental State Examination In order to rule out the presence of dementia, each p articipant completed prior to data collection a cognitive screening instrument, the Mini Mental State Examination, or MMSE (Folstein, Folstein, & McHugh, 1975). The MMSE, a brief 30 item questionnaire commonly used in health research to screen for dementi a, was administered by the investigator, face to face. The MMSE assesses various cognitive domains including arithmetic, memory, and orientation, and yields a total score calculated on a 30 point scale. For the current investigation, the criterion for ex clusion from participation was a score of 26 or less. However, no participant scored below a 27 on the MMSE. Hearing Evaluation Each participant underwent a comprehensive hearing evaluation, including the following procedures: tympanometry, acoustic refl ex screening, distortion product otoacoustic emissions

PAGE 43

43 (DPOAEs), and assessment of pure tone air conduction thresholds, speech reception thresholds, and word recognition performance. Aural Immittance Measurement. Tympanometry measures the volume of the e ar canal and compliance of the tympanic membrane (eardrum) and middle ear system as air pressure in the external ear canal is systematically varied from +200 decaPascals (daPa) to 200 daPa. Aural immittance measures help to identify middle ear dysfunctio n that may be associated with conductive hearing loss. Measurement of acoustic reflexes evaluates the integrity of the auditory pathway including the middle ear, the cochlea, the auditory (8 th cranial) nerve, and lower brainstem (pons) by the detection of a reflex of the stapedius muscle within in the middle ear. For this investigation, ipsilateral acoustic reflexes (measurement of stapedius muscle contraction in the same ear as the stimulus tone was presented) were screened a t 90 dB HL bilaterally for pu re tone signals of 500, 1000, 2000, and 4000 Hz. Tympanometry and acoustic reflex screening were conducted using a Grason Stadler GSI 16 immittance meter. Distortion Product Otoacoustic Emissions Measurement of DPOAEs evaluates function of the outer hai r cells in the cochlea. DPOAEs are elicited by the simultaneous presentation of two pure tones closely spaced in frequency. When the cochlea is healthy and the outer hair cells are motile, introduction of the tone pair to the ear will activate the cochlea in the regions of the basilar membrane corresponding to the frequency of those tones. When the resulting traveling waves along the basilar membrane overlap, energy peaks are produced at several discrete frequencies related to the frequencies of the stimu lus tones. This output (a distortion product or DP) can be used clinically as an objective test of cochlear health: if the energy peak is sufficiently high, this suggests that the portion of the cochlea stimulated by the input tones is intact and respond ing

PAGE 44

44 appropriately to sound (Hall, 2000). Measurement of DPOAE status is highly site specific as a measure of cochlear function. Laboratory experiments have demonstrated that OAEs can be measured in ears which have had the auditory nerve completely severe d, indicating that connection to the auditory cortex or the brainstem is unnecessary for OAEs to be present, and that OAEs therefore are uniquely dependent on cochlear status (Siegel & Kim, 1982). Distortion product OAE measurement is made using a probe mi crophone placed in the ear canal. For each DPOAE stimulus, a pair of pure tones is introduced into the canal at a specified intensity and frequency ratio. Specifically, the intensity of the lower frequency tone (L 1 ) is 65 dB sound pressure level (SPL) an d the intensity of the higher frequency tone (L 2 ) is 55 dB SPL. The frequency ratio is held constant for each stimulus pair such that the frequency of the higher frequency tone (f 2 ) is equal to 1.2 times the frequency of the lower tone (f 1 ) when rounded t o the nearest one hundredth, which can be expressed as f 2 / f 1 = 1.2. These approximate frequency and intensity ratios have been found to optimize the distortion response and to enhance the sensitivity of the DP measurement to cochlear deficit. That is, using these parameters, the DP amplitude is maximally reduced when outer hair cell pathology is present (Hall, 2000). When the outer hair cells are intact, the introduction of the f 1 and f 2 tones produces distortion responses at several discrete frequenci es. These responses include harmonic tones, produced at the simple multiples of the stimulus tones (i.e. 2f 2 ,2f 1 ); summation tones, produced at various additive combinations of the stimulus tones (i.e. f 1 + f 2 2f 1 + f 2 ); and difference tones, produced a t various subtractive combinations of the stimulus tones (i.e. f 1 f 2 2f 1 f 2 ). For DPOAE testing, the difference tone at 2f 1 f 2 is the principal measure. Repeated amplitude measurements of the distortion responses are made by the probe microphone. For diagnostic purposes, these measurements focus on the amplitude of the DP

PAGE 45

45 frequency (2f 1 f 2 ) and the amplitude of the noise floor, or the average amplitude within a narrow frequency range surrounding the DP frequency. The DP and noise floor amplitude are averaged and plotted on a graph often called a DPgram. The DPgram displays the amplitude of the DP and the noise floor as a function of frequency. The DP amplitude can then be compared to normative data to determine whether the DP meets criteria to be labeled as normal. A full diagnostic (University of Florida 65/55 protocol) DPOAE test was conducted in each ear. The UF 65/55 protocol evaluates distortion products at nineteen test frequencies between 750 and 8000 Hz. A numerical summary of this pro tocol including frequency and intensity levels for each stimulus tone pair can be seen in Table 3 1. Scoring of DPOAE data was conducted by examining the DPgram and labeling each DP ve data; noise floor amplitude within 6 dB of the noise floor). A normal r esponse suggests good outer hair cell health in the area of the cochlea where the DP was produced. An abnormal response suggests some outer hair cell pathology but enough structure remaining to produce a diminished response. In the absence of confounding factors (such as middle ear pathology or malfunctioning equipment), an absent response suggests a sufficient loss of outer hair cells in the measured portion of the cochlea to diminish the DP response to below measurable levels. Following labeling, findin gs for DPOAE data were summarized in three categories based

PAGE 46

46 test frequ encies in each category for each ear was analyzed and labeled as normal, abnormal, or absent overall. See Table 3 2 for a summary of the DPOAE category grouping. All DPOAE recordings were replicated to assess reliability of results. Responses were obtai ned using a Grason Stadler Audera evoked emissions device. Calibration of the Audera calibration program. The probe was inserted into a test cavity and was activa ted using the calibration program, which measured the tone outputs from the probe in the frequency range used for DPOAE testing. Output intensities were within normal tolerances for the duration of data collection. During the early stage of data collectio n (late September to November 2006), DPOAE measurement was conducted in the main area of the test room (Dauer Hall 23). Due to high levels of background noise resulting from hallway traffic and a large adjacent construction project, low frequency OAEs eit her could not be obtained or were confounded by high noise floor measurements for participants seen early in the investigation (a total of 16 participants). In late November, an extension cable was installed allowing the Audera module to be moved adjacent to the single walled sound treated booth in the test room. From December 2006 until the end of data collection in May 2007, DPOAE measurement was conducted inside the sound booth with the door closed as much as possible with the cord from the probe assem bly within the door jam (approximately inch opening). It should be noted that noise from the construction project did not affect any other testing. All other tasks were completed inside of the sound booth with the door closed (i.e., pure tone audiometr y, Gaps in Noise test) or in a room from which the construction noise was inaudible (i.e. Hearing in Noise Test, Hearing Handicap Inventories).

PAGE 47

47 Pure Tone Audiometry Air and bone conduction thresholds were measured following a conventional clinical test pr ocedure as recommended by Carhart and Jerger (1959) and by ASHA (1978). Pure tone thresholds were obtained via air conduction at 250, 500, 1000, 2000, 3000, 4000, and 8000 Hz bilaterally. Where inter octave differences in hearing thresholds were 20 dB o r greater, inter octave frequency thresholds were obtained at 750, 1500, and/or 6000 Hz. When conductive hearing loss was suspected (based on tympanometry, DPOAE findings, and/or air conduction pure tone thresholds), bone conduction audiometry was conduct ed at the octave frequencies from 250 to 4000 Hz. When indicated for either air or bone conduction testing, masking was performed using narrow band noise in the non test ear. Masking is used in audiometry when there is a possibility that the non test ear can be stimulated by the ac levels sufficient to reach the non test ear by either air or bone conduction. Presentation of appropriate levels of steady state noise corresponding to the frequency of the test stimulus (narrow band noise for pure tone testing; speech spectrum noise for speech testing) is used to raise the threshold of the non test ear to minimize the likelihood of a response due to a crossover test signal. Pure tone hearing thresholds were obtained with the subject in a single walled sound treated booth using a Grason Stadler GSI 61 audiometer with TDH 49 supra aural headphones and a bone vibrator headband. For the purpose of audiologic data analyses, hearing sensiti vity was quantified using the Articulation Index (AI), a measure derived from audiometric thresholds (ANSI, 1986; Fletcher & Galt, 1950; French & Steinberg, 1947). The AI is used to express the amount of speech information that is available to an individua l based on hearing sensitivity level. Usually, the AI is

PAGE 48

48 expressed as a number between 0 and 1.0, where 0 indicates that no speech information is received and 1.0 indicates that all speech information is received. The AI is typically calculated by dividing the frequency spectrum into several bands, and then weighting each band between 250 and 4000 Hz based on the importance of that frequency region for understanding speech in quiet environments. That is, unlike a simple mean of thresholds at speech frequen cies, the AI weights audiometric frequency bands unequally to emphasize the frequency ranges that are most important for speech understanding. Using the AI, the percentage of loss can be calculated in each band and combined into an index number representi ng the audibility of a typical speech signal in a quiet background for the measured ear. In the present study, articulation indices were calculated for the right and left ear for each participant, and an average AI was calculated based on the right and l eft ear AI values. The higher AI of these two values for each participant was considered the better ear AI, whereas the lower AI was considered the worse ear AI. A common and longstanding difficulty in regression analysis of hearing data is the method b y which audibility, the quality of a sound being perceptible, is quantified and controlled. A review of literature was conducted to determine the most common practices for quantifying audibility and, specifically, in studies that include analysis of predi ctive factors for speech perception. This literature review identified several methods that have been used to quantify audibility for the purpose of regression analysis. One common method of calculating audibility is use of the pure tone average (PTA), or the arithmetic mean of the thresholds in dB hearing level (HL) at a given set of test frequencies. The PTA is most commonly calculated as the average of hearing thresholds at 500, 1000, and 2000 Hz (e.g,. George et al, 2007; Jerger et al, 1989).

PAGE 49

49 Howeve r, other speech perception studies have calculated PTAs with thresholds at 500, 1000, 2000, and 4000 Hz (e.g., Divenyi & Haupt, 1997; Vaughan et al, 2006); 1000, 1500, and 2000 Hz (e.g., Plyler & Hedrick, 2002); 1000, 2000, and 4000 Hz (e.g., Gordon Salant & Fitzgibbons, 1993; Humes, 1996); 2000 and 4000 Hz (e.g., Arlinger & Dryselius, 1990); and 2000, 3000, 4000, and 6000 Hz (e.g., Helfer & Wilber, 1990). The PTA is generally the simplest way to used to quantify audibility. Despite the wide use of this technique, the substantial variability in frequencies chosen for the average illustrates the lack of agreement among researchers on a technique for quantifying audibility as a single number variab le. Audibility can also be quantified using the Articulation Index (e.g., Dubno & Dirks, 1989, 1990; Dubno, Horwitz, & Ahlstrom, 2005; Festen & Plomp, 1990; Lee & Humes, 1992; Schum et al, 1991; Souza & Turner, 1999; Souza et al, 2000; Turner & Henry, 200 2). The AI is often chosen as an audibility measure instead of a PTA because the weighting built in to the index emphasizes the contribution of frequency bands most important for speech perception. For each of these methods, a binaural index measure for audibility is calculated using some combination of left ear and right ear threshold measures, PTA, or AI. In most cases, this is accomplished by using only the best ear measure (e.g., Duquesnoy & Plomp, 1980; George et al, 2007; Ventry & Weinstein, 1983) Less common is use of a binaural index derived by averaging the left ear and right ear measures (e.g., Cokely & Humes, 1992; Humes and Floyd, 2005). As with the PTA/AI methodology differences discussed above, no consensus exists for a preferred method for conversion of monaural hearing sensitivity to a single number binaural measure. In everyday listening environments containing diffuse noise, it is likely that the better ear predominates in detecting and processing speech signals. Thus, it is logical to presume that the

PAGE 50

50 sensitivity should be emphasized when estimating total audibility. However, it is equally certain that the better ear does not contribut e independently to listening in these environments. Except for cases of severe asymmetry (i.e., profound unilateral deafness), the worse ear does contribute to binaural listening. Specifically, the worse hearing ear is important for psychoacoustic functio ns, among them localization, i.e., the ability to locate a sound source in three dimensional space (Hirsh, 1950; Kock, 1950; Litovzky & MacMillan, 1994; MacKeith & Coles, 1971); binaural summation, i.e, the increase in perceived loudness when a signal is h eard with both ears (Festen & Plomp, 1986; Haggard & Hall, 1982; Hawkins, Prosek, Walden, & Montgomery, 1987; Hawkins & Yacullo, 1984; Moncur & Dirks, 1967; Ricketts, 2000); and binaural squelch, i.e., the suppression of interfering sounds when listening t o a single signal in a noisy environment (K aplan & Pickett, 1981; Koenig, 1950; Koenig, Allen, Berkley, & Curtis, 1977; Ideally, any index for audibility would account for these factors in a weighted better ear / worse ear average. However, the appropriate weighting is unknown and may well be unknowable. It is likely that the contribution of the worse ear is different depending on the degree of asymmetry, as well as the central auditory status of the individual and his or her ability to maximize dichotic cues. The con tribution of the worse ear to listening is important in the calculation of hearing impairment for medico Compensation. As of 2006, the four calculations in active use in the United States at the state level for Worke in calculating binaural impairment (Dobie & Megerson, 2000; John & Kreisman, in preparation). As noted by Dixon Ward (1983), however, no empirical rationale exists for the choice of these

PAGE 51

51 weighting schemes. For example, the widely used AMA 1947 formula, which uses a 7:1 better ear to worse rationale (American Medical Association, 1947). In fact, recent research suggests that weighting schemes heavily emphasizing the better ear may be inappropriate and may insufficiently account for the contribution of the worse ear to total hearing handicap (John & Kreisman, in preparation). For studies including participant groups that are restricted to individuals with essentially symmetrical hearing loss, the question of better ear/worse ear contribution is not a pressi ng one (e.g. George et al, 2007; Humes and Floyd, 2005). That is, the better ear AI or PTA and the binaural average AI or PTA are very similar in cases of symmetrical hearing loss. The present investigation, however, was designed to evaluate the predicti ve value and clinical efficacy of the experimental measures for the predominant adult population that are likely to seek clinical services: middle aged and older adults who are aware that they have a hearing loss. It is not uncommon for individuals in th is population to present with asymmetrical hearing thresholds. In summary, the AI is a widely used and validated method for accounting for the ability of the ear to detect sounds in the environment. Furthermore, the AI offers an improvement over the vario us forms of the PTA, described above, each of which weights its component frequencies equally despite the fact that certain frequency regions certainly contribute more to speech understanding than others. One potential advantage to using the PTA as the au dibility measure in this analysis was that it can be easily and quickly calculated from the audiogram by clinical audiologists. Determining the AI, by contrast, requires either an arithmetic formula or a visual employed by Mueller and Killion (1990). While not difficult, this method is slower than calculating the PTA, and is rarely employed in the

PAGE 52

52 clinical setting. The distinction between the PTA and the AI can be summarized based on a choice between maximizin g convenience or accuracy. The AI is considered a gold standard measure for audibility in the hearing research community and is certainly a more accurate reflection of the relation between hearing sensitivity and speech perception in quiet environments. For this reason, the AI was employed in this investigation as the audibility variable. Articulation indices were denoted for analysis as follows: right ear AI (RAI), left ear AI (LAI), better ear AI (BETAI), worse ear AI (WRSAI), and binaural average AI ( BINAI). The contributions of the better and worse ears to total auditory perception are difficult to quantify. Because previous research has not yielded a well validated ear weighting mechanism and because, as stated previously, the predominant method f or controlling audibility is use of a better ear single number index, in the present investigation the better ear AI (BETAI) was used as the principal measure of audibility. Speech Audiometry Speech reception thresholds were used to confirm pure tone thres hold results. Participants were asked to repeat a series of spondaic (spondee) words at varying intensity levels until threshold for repetition (the lowest intensity level at which two out of three words were correctly repeated) was found. Spondaic words are two syllable words with equal stress on each syllable when spoken aloud. Some sample spondaic words are baseball, airplane, cowboy, and greyhound. Speech reception thresholds were obtained using a monitored live voice presentation with a descending method. Word recognition testing was conducted using the Northwestern University Auditory Test number 6 (NU 6) ordered by difficulty word lists (Tillman & Carhart, 1966). These lists consist of phonemically balanced consonant nucleus consonant words commo nly used in spoken English. Some sample NU 6 words are mouse, knock, chat, and thin. Words are presented using

PAGE 53

53 one half list of 25 words per ear, presented at listening level (MCL). If the first ten words of a list (the ten most difficult words from that list appear first) were repeated correctly, word recognition assessment was stopped and a word recognition score of 100% was assigned for the test ear. Word recognition measurement was conducted using CD recordings of the NU 6 stimuli ordered by difficulty. Speech reception thresholds and word recognition scores were obtained using a Grason Stadler GSI 61 audiometer with TDH 49 supra aural headphones. The tests were conducted in a single walled sound treated booth. Speech Perception in Noise Testing: Hearing In Noise Test (HINT) Assessment of speech perception in noise was conducted using the Hearing In Noise Test or HINT (Nilsson, Soli, & Sullivan, 1994) in quiet and noise conditions. The HINT consists of 25 lists of 10 phonemically balanced sentences spoken by a male voice and recorded to CD with speech spectrum noise matched to the long term spectrum of the se ntences serving as noise sentence lists can be found in the appendix of Nilsson et al (1994). Thresholds in noise were obtained using an adaptive procedure. The noise wa s held steady at 65 dB SPL and the presentation level of the sentences was varied by the investigator based upon whether the participant was able to repeat each sentence correctly and in full. Thresholds in quiet were obtained using a similar procedure, b ut without the noise competition. For analysis, these variables are denoted as HINTQ for the threshold in quiet and HINTN for the threshold in noise. The HINT was administered in a double walled sound treated booth using a five speaker array surrounding the seated participant. One speaker (Tannoy System 600), located at 0 degrees

PAGE 54

54 azimuth, 1 meter from the participant, was used to deliver the sentences. Four speakers (Definitive BP 2X bipolar) located at 45, 135, 225, and 315 degrees azimuth (in the cor ners of the booth), 1 meter from the participant, were used to deliver the noise competition. This speaker array simulated listening to a single talker within a diffuse noise field typically encountered in everyday listening environments. Similar speaker configurations have been used in sound field speech perception studies both within the UF Department of Communication Sciences and Disorders and in other laboratories (see Hornsby & Ricketts, 2007; Kreisman & Crandell, 2002). A diagram of the HINT sound field speaker presentation setup is illustrated in Figure 3 1. Sound levels were calibrated monthly using a Quest 2700 type II sound level meter with octave band filter and a 1000 Hz calibration tone modified from the HINT test CD. The calibration tone recorded for use with the HINT is matched to the long term root mean square intensity of the sentence stimuli. Because use of a steady state tone is not appropriate for sound field calibration due to the potential of developing standing waves, the tone wa s modified using Adobe Audition 1.5. Specifically, the calibration tone was frequency modulated (warbled) by +/ 50 Hz at a modulation rate of 5 Hz, meaning the tone warbled from 950 to 1050 Hz five times per second. Measurement was conducted using Audit ion to ensure that the long term intensity of the warbled calibration tone was matched to the unmodified tone. Twenty four HINT lists were paired (list 1 with list 2, list 3 with list 4, etc.) and randomly selected for each of the two presentation conditio ns. No list was repeated for any participant. Administration of the procedure was conducted in a manner consistent with the instructions accompanying the HINT. Briefly, the first sentence of the first of each pair of lists was presented at a low intensi ty level and re presented, increasing the presentation level in 4 dB

PAGE 55

55 steps, until the sentence was repeated correctly. The second sentence was then presented at a level 4 dB below that intensity level. The presentation level was then adapted according to whether the sentence was correctly repeated (by decreasing the intensity) or incorrectly repeated (increasing the intensity) in 4 dB steps through sentence 4 and in 2 dB steps thereafter. After the twentieth sentence (tenth sentence in the second list) w as presented and a response was given by the participant, the investigator noted what the presentation level would be for a twenty first sentence if one were presented. Though no further sentences were presented, this level was noted as the presentation l evel for sentence twenty one. After HINT administration was complete, a reception threshold for sentences (RTS) in dB SPL was calculated by averaging the presentation level of the fifth through twenty first sentences. A correction factor of 1 dB was added to the presentation level to adjust for the actual was determined by the investigator prior to beginning this investigation (and verified throughout the data collection period) by measuring the intensity of the calibration tone in dB SPL at the in dB HL. A signal to noise ratio (SNR) was calculated by subtracting th e level of the noise (65 dB SPL) from the RTS. The SNR indicates the difference in intensity between a stimulus voice and a typical level of diffuse environmental noise that is required by the subject to correctly perceive speech. Noise Competition Cont inuous speech spectrum noise served as the noise competition for HINT testing. This noise is matched to the long term frequency spectrum of the HINT sentences and is included with the HINT test CDs. For this investigation, the noise competition track was recorded to the

PAGE 56

56 right and left channel of two CDs. To control for the possible confound of comodulation masking release (Grose & Hall, 1992), the noise tracks were uncorrelated by shifting the starting point of each noise track fractionally so that, if a ll four tracks were started simultaneously, the four CD channels would produce noise signals identical in spectral content but out of phase. Each CD channel was routed separately through one channel of a Crown D 75A amplifier to a Definitive BP 2X bipolar speaker (see previous section for speaker configuration). Each speaker was calibrated at the level of the amplifier to produce a noise signal at 59 dBA, measured at the head of a participant seated in the booth. Thus, when all four speakers were driven simultaneously, the noise competition at the center of the room was 65 dB SPL. Noise levels were calibrated monthly using a Quest 2700 type II sound level meter with octave band filter in the manner described above. Temporal Resolution Testing: Gaps In N oise (GIN) Test Assessment of auditory temporal resolution was conducted using the Gaps In Noise test, or GIN (Musiek et al, 2005). The GIN test is a recently developed measure of temporal resolution designed to minimize the potential confounds of stimulu s frequency, task complexity, and age in temporal resolution testing. The GIN test consists of four lists of six second segments of broadband (white) noise containing between zero and three silent gaps per segment. These gaps range in length from two to 20 milliseconds (ms). For each set of noise segments, a total of 60 gaps were presented with six gaps each of 2, 3, 4, 5, 6, 8, 10, 12, 15, and 20 ms randomly distributed within 29 to 36 total noise bursts. Listeners were instructed to press the audiomet response button whenever a gap was detected. The GIN threshold was calculated as the shortest gap length at which the participant detected correctly at least four of six gaps and at least four of six gaps at the next longest gap length.

PAGE 57

57 Two GIN lists were randomly selected from the four available lists for each participant. One list was presented to the right ear and one to the left ear in a random order. Testing was conducted monaurally, using one list per ear. The order of ear tested was alternat ed from participant to participant. No GIN list was used twice for any participant. Prior to administering the GIN test, the investigator explained the task according to the instructions provided with the test materials. Each participant was then given several stimulus noise bursts from a practice list most comfortable listening level (MCL). For all participants, MCL was between 35 and 50 dB SL re: SRT, the recommended presentation level parameters for the GIN. Recent studies have indicated that GIN threshold is insensitive to presentation level within this range (Musiek et al, 2005; Weihing et al, 2007). The GIN test was conducted in a single walled sound treated booth under supra aural headphones. Gap stimuli were presented monaurally via the headphones and responses were stimulus) channel of the test CD provides b eeps aligned with the gaps on the stimulus channel. The second channel of the CD player was routed through the second channel of the audiometer and to an absent transducer (speaker) to provide cues to the presence of the gaps to the tester, without being audible to the participant. By monitoring the non stimulus channel, the investigator could determine accurately whether a gap had occurred prior to each participant responding button press. Monaural GIN thresholds were calculated for each ear and are deno ted as RGIN (GIN threshold in the right ear) and LGIN (GIN threshold in the left ear). In addition, the better ear GIN (BETGIN), worse ear GIN (WRSGIN), and a composite binaural GIN (BINGIN) were

PAGE 58

58 calculated. The binaural GIN was created by averaging the GIN threshold in the right and left ear. A GIN scoring form and the four stimulus lists are included in Appendix B. Working Memory Testing: Digit Span and Digit Ordering Tests Working memory was assessed using a test battery consisting of forward digit s pan, backward digit span, and digit ordering tasks. The digit span tasks were taken from the Wechsler Adult Intelligence Scale, version III (WAIS III, The Psychological Corporation, 1997). The digit ordering task was adapted by Dr. Lori Altmann, head of t he University of Florida Language Over the Lifespan Lab, from a test procedure developed by Cooper and colleagues (1991). Each digit memory task required the participant to repeat lists of numbers between 1 and 10. In forward digit span, the participant w as asked to repeat the digit lists spoken by the investigator in the same sequence they were presented. In the backward digit span task, the participant was asked to repeat the numbers in reverse of the sequence they were presented. In the digit ordering task, the participant was asked to reproduce the series in ascending numerical order. Each digit memory test was conducted with digit lists ascending in length until the participant could no longer correctly reproduce the number lists. A score was obtai ned for each digit task based on the number of correctly repeated digit lists. Scores were standardized and summed to produce a total working memory score for analysis. To ensure audibility of the lists, the working memory battery was conducted binaurall y under supra aural headphones in a single walled sound digit forward (denoted as DIGFWD), digit backward (DIGBAK), and digit ordering (DIGORD) tests were standardized using the mean and standard deviation for the participant sample and added together, producing an overall standardized working memory score denoted for analysis as WM.

PAGE 59

59 Self Reported Hearing Handicap: Hear ing Handicap Inventory for Adults (HHIA) or for the Elderly (HHIE) Participants age 65 or older completed the Hearing Handicap Inventory for the Elderly (HHIE, Weinstein & Ventry, 1983), whereas participants under age 65 completed the Hearing Handicap Inve ntory for Adults or, HHIA, Newman, Weinstein, Jacobson, & Hug, 1990). The HHIA/E is a 25 item inventory that assesses the consequences of hearing loss on the lives of adults. Two sub scales comprise the HHIA/E: the HHIA/E Emotional subscale measures the mental and emotional effects of hearing loss on the individual, while the HHIA/E Social/Situational subscale measures the situational and interpersonal consequences of hearing loss. Both scales are frequently used in audiologic clinics to document the ne ed for, and benefit from, rehabilitation. The HHIE and HHIA are almost identical; three questions are slightly reworded in the HHIA to make the inventory more applicable to individuals under age 65. For each item, the inventories, the investigator calculated Emotional and Social/Situational subscale scores as well as a total score on a 100 point scale. The Emotional subscale score of the HHIA and HHIE measure s the emotional impact of hearing loss on the individual. The Social/Situational subscale score quantifies the situational impact of hearing loss (Weinstein & Ventry, 1983). The HHIA and HHIE scales were administered face to face with the answer forms mar ked by the investigator. The participant was provided with a flash card containing the possible answers. When necessary, a frequency modulated (FM) listening system was used to ensure is, the HHIA/E Emotional subscale score was denoted as HHIEMOT and the Social/Situational subscale score was denoted as HHISOC. The HHIA/E total score was denoted as HHITOT.

PAGE 60

60 Table 3 1. Summary of the University of Florida 65/55 Distortion Product Otoacou stic Emissions (DPOAE) protocol. Point f 1 (Hz) L 1 (dB SPL) f 2 (Hz) L 2 (dB SPL) f 2 / f 1 ratio DP (Hz) 1 6070 65 7277 55 1.20 4863 2 5285 65 6340 55 1.20 4230 3 4582 65 5496 55 1.20 3668 4 4020 65 4816 55 1.20 3223 5 3492 65 4184 55 1.20 2801 6 3023 65 3621 55 1.20 2426 7 2625 65 3152 55 1.20 2098 8 2297 65 2754 55 1.20 1840 9 2004 65 2402 55 1.20 1605 10 1746 65 2098 55 1.20 1395 11 1512 65 1816 55 1.20 1207 12 1324 65 1594 55 1.20 1055 13 1148 65 1371 55 1.20 926 14 984 65 1184 55 1.20 785 15 855 65 1031 55 1.20 680 16 750 65 902 55 1.20 598 17 656 65 785 55 1.20 527 18 574 65 691 55 1.20 457 19 504 65 598 55 1.20 410 f 1 = frequency of first stimulus tone in Hertz; L 1 = intensity of first stimulus tone in decibels sound pressure level ; f 2 = frequency of second stimulus tone in Hertz; L 2 = intensity of second stimulus tone in decibels sound pressure level ; DP = distortion product frequency in Hertz

PAGE 61

61 Table 3 2. Assignment of DPOAE test points into high mid and low frequency groups. Point Test Frequency (Hz) f 1 (Hz) f 2 (Hz) Group Group Range 1 7277 6070 7277 High 4000 8000 Hz 2 6340 5285 6340 High 4000 8000 Hz 3 5496 4582 5496 High 4000 8000 Hz 4 4816 4020 4816 High 4000 8000 Hz 5 4184 3492 4184 High 4000 8000 Hz 6 362 1 3023 3621 Mid 1000 4000 Hz 7 3152 2625 3152 Mid 1000 4000 Hz 8 2754 2297 2754 Mid 1000 4000 Hz 9 2402 2004 2402 Mid 1000 4000 Hz 10 2098 1746 2098 Mid 1000 4000 Hz 11 1816 1512 1816 Mid 1000 4000 Hz 12 1594 1324 1594 Mid 1000 4000 Hz 13 1371 1148 1371 Mid 1000 4000 Hz 14 1184 984 1184 Mid 1000 4000 Hz 15 1031 855 1031 Low 500 1000 Hz 16 902 750 902 Low 500 1000 Hz 17 785 656 785 Low 500 1000 Hz 18 691 574 691 Low 500 1000 Hz 19 598 504 598 Low 500 1000 Hz f 1 = fre quency of first stimulus tone in Hertz; f 2 = frequency of second stimulus tone in Hertz

PAGE 62

62 Figure 3 1. Diagram of sound field HINT test environment. BP 2X = Definitive BP 2x Bipolar Speaker (noise source) CD = Sony CDP CE375 5 CD Changer (stimulus a nd noise routing) D 75A = Crown D 75A Two Channel Amplifier (noise routing) GSI 61 = GSI 61 Clinical Audiometer (stimulus routing) T 600 = Tannoy System 600 Monitor Speaker (stimulus source)

PAGE 63

63 CHAPTER 4 RESULTS Participant Characteristics Fifty six adult s served as participants for the present investigation. The mean age for the participant group was 66.4 years (SD = 9.4). Participants ranged in age from 50 years, 1 month to 89 years, 11 months. Seventeen participants were between 50 and 59 years of ag e (inclusive), 19 participants were between 60 and 69, 15 were between 70 and 79, and 5 were between 80 and 89. Seventeen participants were male while 39 were female. Four participants (two male, two female) were previous hearing aid users. Prior to com pleting the study measures, all participants were screened for cognitive impairment using the Mini Mental State Examination (MMSE; Folstein et al, 1975). The mean MMSE score for this group was 29.6 (SD = .8) with a range of 27 to 30. The cutoff for study participation was 26. All participants were able to complete all measures administered. Statistical Analysis Participant data were compiled using Microsoft Excel 2003 for Windows. Statistical analyses were performed using SPSS for Windows, release 11.5 .0. Stepwise multiple regression analyses were performed for the dependent variables (HINT threshold in noise or HINTN; HINT threshold in quiet or HINTQ; HHIA/E Emotional subscale score or HHIEMOT; HHIA/E Social subscale score or HHISOC; and HHIA/E total score or HHITOT) to determine which predictor variables could most effectively account for variance in speech perception ability and hearing handicap score. The predictor variables examined were audibility, as measured by Articulation Index score in the b etter ear (BETAI); temporal resolution ability, as measured by the average of GIN thresholds in the right and left ears (BINGIN), and working memory capacity, as measured by a composite standardized working memory score (WM).

PAGE 64

64 The SPSS 11.5.0 software pac kage was also used to calculate descriptive statistics for the predictor variables, dependent variables, and other subject measures, such as age and gender, The following variables examined and discussed in this section were not included as predictor varia bles: GIN thresholds in the right ear (RGIN), left ear (LGIN), better ear (BETGIN), and worse ear (WRSGIN); and working memory scores on forward digit span (DIGFWD), backward digit span (DIGBAK), and digit ordering (DIGORD) tasks. All analyses were tested = .05. A glossary of abbreviations and acronyms used for the variables studied can be found in Appendix C. Distortion Product Otoacoustic Emissions (DPOAEs) Participant DPOAE data are summarized in Table 4 1 and Figure 4 1. Table 4 1 provid es a comparison table for DPOAE and audiometric status. In this table, participants demonstrating normal audiometric data and normal DPOAE status within a frequency range category were classified as normal/normal; participants with abnormal DPOAEs and aud iometric thresholds were classified as abnormal/abnormal. Participants with audiometric thresholds within normal limits but absent or abnormal DPOAEs were classified as normal/abnormal. DPOAEs are a more sensitive index of cochlear function than is the pure tone audiogram (Hall, 2000). The group of individuals with audiometrically normal hearing and abnormal DPOAEs can be interpreted as individuals who have cochlear dysfunction that is not yet detected by a decrease in pure tone threshold. Such individ uals may demonstrate other difficulties resulting from loss of outer hair cells, such as diminished cochlear fine tuning and spectral resolution ability, that are not apparent by pure tone audiometry. Four left ears demonstrated normal DPOAEs but abnorma l audiometric thresholds for the mid frequency range only. Closer inspection of these four ears revealed steeply sloping hearing loss in each of these ears, with thresholds below 2000 Hz within the normal range and thresholds

PAGE 65

65 at 3000 and 4000 Hz abnormal. For these four participants, DPOAE data were consistent with audiometric data. The mismatch in this classification is an artifact of the grouping scheme. In other words, if the mid frequency range were further separated into low mid and high mid ranges and audiometric data were compared, the low mid range (approximately 1000 to 2000 Hz) would be normal both audiometrically and in DPOAE measurement, while the high mid range (approximately 2000 to 4000 Hz) would be abnormal in both respects. Figure 4 1 pr ovides a breakdown by ear of DPOAE status as a function of test frequency range (low, mid, and high) for this participant group. Declines in cochlear status particularly affecting the high frequency area of the cochlea are evident bilaterally. As detaile d in the methods chapter, low frequency DPOAE measurement was confounded to some degree by high ambient noise levels in the test room for the first sixteen participants in this investigation. These participants were tested prior to the installation of an extension cable allowing DPOAE measurement to be conducted within the sound booth. As a result, low frequency DPOAEs may not be a reliable measure in this investigation. It is likely that some participants were classified as having absent low frequency D POAEs when responses were present but could not be measured due to the elevated noise floor. However, examination of higher frequency DPOAE data reveal declines from the mid to high frequency DPOAE test ranges. These declines are consistent with the co mmon findings of increased loss of cochlear function and high frequency hearing loss with age. Furthermore, comparison of DPOAE data to audiometric thresholds in Table 4 1 suggests that more cochlear dysfunction was present within the participant group th an was apparent in pure tone threshold testing. Audibility The composite audiogram for the fifty six participants in this investigation revealed a moderate sloping high frequency sensorineural hearing loss bilaterally (Figure 4 2). Standard

PAGE 66

66 deviation bar s illustrate the substantial variability in hearing status among this rather unselected group of adults who thought they had a hearing problem. Mean SRT and word recognition scores (conducted at MCL) are reported along with mean pure tone thresholds in Ta ble 4 2. A repeated measures within subjects comparison between right and left ear thresholds revealed a significant difference at 4000 Hz only (F 1, 55 = 9.36, p = .003), indicating that, overall, participants had slightly but significantly better hearin g sensitivity at 4000 Hz in their right ears than in their left ears. Right and left ear audiometric thresholds were not significantly different at any other test frequency. Neither the SRT nor word recognition scores were significantly different betwee n the left and right ears in this participant group. The mean AI in the sample was 0.83 in the right ear (SD = .18) and 0.78 in the left ear (SD = .23). The mean better ear AI was .85 (SD = .17). These AI scores predict generally good ability to perceiv e speech on average but with substantial variability within the sample. Variability in audibility was an expected finding given the recruiting strategy for this investigation, i.e., individuals within a wide age range and with no upper or lower limit on d egree of hearing loss. A repeated measures within subjects comparison of AI scores revealed significantly better AI in the right ears compared to the left ears in this participant sample (F 1, 55 = 4.02, p = .050). While significant, this difference was s mall, with an average difference of 0.05, equivalent to 5% speech intelligibility. correlation ( r ) between right ear AI (RAI) and age was .75 (p < .001), and the correlation between left ear AI (LAI) and age was .57 (p < .001). The principal audibility measure, better ear AI (BETAI), also was significantly correlated with age ( r = .76, p < .001). The correlations

PAGE 67

67 in this participant sample confirm the commonly encountered strong negative association between hearing sensitivity and age in adults (worsening hearing with increasing age). Of the fifty six participants in this investigation, three demonstrated a large asymmetry in hearing sensitivity, defined as a difference in AI of 0.25 or more between ears. One participant presented with a congenital unilateral moderately severe hearing loss in her left ear; one participant presented with a severe left diagnosed in the 197 0s. Another participant presented with a moderate (20 25 dB) hearing loss in her right ear of unknown etiology, with an apparent onset in childhood. Temporal Resolution Temporal resolution threshold in each ear for each participant was calculated accordin g to the prescribed threshold method for the GIN test. Threshold was defined as the shortest gap the participant was able to identify in at least four out of six trials, while also identifying the next longest gap at least four out of six trials. The ave rage GIN thresholds for the participant group were 7.6 ms in the right ear (SD = 2.1 ms) and 8.4 ms in the left ear (SD = 2.8 ms). Thresholds ranged from 5 to 15 ms in the right ear and from 5 to 20 ms in the left ear. The distribution of GIN thresholds is graphed in Figure 4 3. A repeated measures within subjects comparison between right and left ear GIN thresholds revealed that left ear thresholds were significantly longer (worse) than right ear thresholds in this group (F 1, 55 = 6.01, p = .017). Bo th right and left ear GIN thresholds were significantly correlated with age, with audibility (as measured by the AI in the same ear as the GIN threshold), and with each other. Better ear, worse ear, and binaural GIN threshold (calculated as the average o f left and right ear thresholds) were also significantly correlated with age and audibility (see Table 4 3 for a correlation matrix for these variables). Monaural GIN and AI variables were moderately correlated (LAI x LGIN: r = .64, p < .001; RAI x RGI N: r = .51, p < .001).

PAGE 68

68 The negative r values indicated that individuals with serious hearing loss (lower AI scores) demonstrated significantly longer gap detection thresholds (higher GIN thresholds) than individuals with little or no hearing loss. Monau ral audibility measures accounted for about 26% of the variance in temporal resolution ability in the right ear and 41% in the left ear. When controlling for audibility, GIN thresholds were not significantly correlated with age. Musiek et al (2005) report ed a normal average GIN threshold of 4.9 ms in the right ear (SD = 1.0 ms) and 4.8 ms (SD = 1.0 ms) in the left ear for a group of 50 normal hearing control subjects aged 13 to 46 years. Using an upper limit for normal GIN test scores as two standard devi ations above the mean, Musiek and colleagues defined normal performance as a threshold of 7 ms or better. Of the 56 participants in the present investigation, 40 yielded at least one GIN threshold value greater than 7 ms. According to the normative crite rion suggested by Musiek et al (2005), therefore, 71% of the participants in the present study showed impaired auditory temporal processing. Figure 4 4 and Table 4 4 display comparisons between the normative GIN data and participant data collected in the present investigation. Working Memory Working memory was assessed using a battery of commonly employed digit tasks: forward digit span, backward digit span, and digit ordering. In order to create a composite working memory score, individual test score s were standardized at the conclusion of data collection and the resulting Z scores were summed. For the forward digit span test, a score of 8 approximates successful repetition of a maximum six digit string. The maximum possible score on forward digit spa n is 14. Forward digit span (DIGFWD) scores averaged 8.1 (SD = 2.1) and ranged from 5 to 13. Backward digit span (DIGBAK) scores averaged 6.8 (SD = 2.1). A score of 7 approximates successful repetition of a maximum four to five digit string, placing the digits in

PAGE 69

69 reverse order. Scores ranged from 3 to 13. The maximum possible score on backward digit span is 14. Digit ordering (DIGORD) scores averaged 16.4 (SD = 4.0). A score of 16 approximates successful repetition of a maximum five digit string, plac ing the digits in numerical order. Scores ranged from 3 to 24. The maximum possible score on digit ordering is 24. Working memory test scores were compared to normative data collected by the University of Florida Language over the Lifespan laboratory on community dwelling older adults in the Alachua County area. Data were available from this laboratory for age groups 60 64 (n=12), 65 69 (n=32), 70 74 (n=48), 75 79 (n=58), 80 84 (n=41), and 85 or older (n=17). No data for individuals between age 50 an d 59 years were available. Table 4 5 provides comparisons between the participant group in the present investigation and the normative group within these age ranges. Participants age 70 and up had mean scores slightly lower than normative means on all w orking memory measures. However, paired samples T tests on all measures confirmed that the participant group did not differ significantly from the normative group on any of the working memory tests in any age group examine investigation was over age 85. This participant scored more than two standard deviations below the mean of the normative group on the digit ordering task. In general, comparison of data from the present i nvestigation to the local normative data suggests that the participant group had working memory status within the normal range for community dwelling adults in the geographical area from which the group was sampled. Moderate collinearity was calculated amo ng all working memory measures (DIGFWD x DIGBAK: r = .49, p < .001; DIGFWD x DIGORD: r = .37, p = .006; DIGBAK x DIGORD: r

PAGE 70

70 = .60, p < .001). These relations suggest an association among these measures, but substantial variance in each unexplained by t he others. There was a significant negative correlation between age and all working memory measures: DIGFWD ( r = .33, p = .012), DIGBAK ( r = .32, p = .017), DIGORD ( r = .48, p < .001), and total WM ( r = .47, p < .001), suggesting an age related decr ease in working memory status. When controlling for age, none of the working memory variables (DIGFWD, DIGBAK, DIGORD, and WM) were significantly correlated with BETAI. Collinearity Testing Collinearity testing was conducted for each experimental variab le (GIN thresholds and working memory scores) to determine whether the variance each predicted in HINT or HHIA/E score, if any, was shared with the other experimental variables. A correlation matrix for the variables RGIN, LGIN, BETGIN, WRSGIN, BINGIN, DI GFWD, DIGBAK, DIGORD, and WM can be seen in Table 4 6. A small but significant negative correlation was seen between RGIN and DIGORD ( r = .27, p = .046). No other correlations were significant. Because independence between the temporal resolution and wor king memory measures was anticipated, the unexpected correlation between right ear GIN threshold and digit ordering ability was examined further. As described previously, all GIN and working memory variables were significantly correlated with age. When a ge was controlled, the correlation between RGIN and DIGORD disappeared ( r = .11, p = .440). It is likely that the association between RGIN and DIGORD is a spurious correlation caused by the influence of age on each of the experimental variables. Speech Perception Speech perception thresholds with and without noise competition were obtained according to the recommended procedure for the HINT as described in Chapter 3. Thresholds in

PAGE 71

71 noise (HINTN) and quiet (HINTQ) were calculated by averaging the present ation level of HINT sentences five through twenty one and adding a 1 dB correction factor for the conversion to dB SPL. A signal to noise ratio (SNR) for threshold in noise was calculated by subtracting 65 (the intensity level of the competition noise in dB SPL) from the corrected HINT threshold in noise. The average speech perception threshold in the quiet listening condition was 38.7 dB SPL (SD = 9.3). Thresholds ranged from 26.1 to 63.5 dB SPL, indicating large variance within the participant group in ability to perceive spoken speech in quiet environments. As expected, HINTQ showed a strong and significant negative correlation with BETAI ( r = .86, p <.001) indicating that the majority (about 73%) of the variance in speech perception in quiet is expla ined by loss of hearing sensitivity. With hearing sensitivity controlled, age was not a significant predictor of speech perception performance in quiet. The average speech perception threshold in the noise condition was 72.2 dB SPL (SD = 2.6). Threshol ds ranged from 68.7 dB to 81.7 dB SPL. That is, participants perceived the sentence stimuli correctly at an average SNR of 7.2 and in a range from 3.7 to 16.7 dB. A strong and significant negative correlation was seen between HINTN and BETAI ( r = .72, p < .001), indicating that hearing sensitivity accounted for about 52% of the variance in speech perception in noise. With hearing sensitivity controlled, age was not a significant predictor of speech perception performance in noise. Both HINTQ and HINTN we re positively correlated with temporal resolution threshold when audibility was not controlled. Quiet condition HINT thresholds were significantly correlated with RGIN ( r = .36, p = .006); BETGIN ( r = .32, p = .018); WRSGIN ( r = .28, p = .040); and BINGIN ( r = .31, p = .019). Noise condition HINT thresholds were significantly correlated with RGIN ( r = .42, p = .001); LGIN ( r = .39, p = .003); BETGIN ( r = .42, p = .001);

PAGE 72

72 WRSGIN ( r = .41, p = .002); and BINGIN ( r = .44, p = .001). However, when controlli ng for BETAI, no significant associations were seen between temporal resolution variables and HINT thresholds in quiet or in noise. Few significant associations were found between HINT scores and working memory variables. HINTQ was weakly but significan tly correlated with DIGORD ( r = .31, p = .018) and with WM ( r = .29, p = .034), while HINTN was weakly correlated with DIGORD only ( r = .31, p = .019). When controlling for BETAI, no significant associations were seen between working memory variables an d HINT thresholds in quiet or in noise. Because there were no significant partial correlations between any temporal resolution or working memory measure and HINTQ or HINTN, it was expected that none of those variables would contribute significantly to a mu ltiple regression on either HINTQ or HINTN in the presence of BETAI. Stepwise linear regression procedures were conducted separately for HINTQ and for HINTN to confirm this expectation. The predictor variables included were audibility (BETAI), temporal r esolution (BINGIN), and working memory (WM). The bivariate correlations between HINT scores and the various measured (RGIN, LGIN) and derived GIN variables (BETGIN, WRSGIN, BINGIN) were small; therefore the model was simplified to include only the average binaural GIN threshold. Only BETAI was retained as a predictor variable in the regression equation for HINTQ ( R = .86, F 1, 54 = 150.82, p < .001). Inclusion of BINGIN and WM increased the correlation coefficient from .858 to .863, which was not a signific ant change (F 2, 52 = .79, p = .459). Similarly, only BETAI was retained in the regression equation for HINTN ( R = .72, F 1, 54 = 58.67, p < .001). Inclusion of temporal resolution and working memory variables increased the correlation coefficient from .72 2 to .731, which was not a significant change (F 2, 52 = .74, p =

PAGE 73

73 .482). As noted above, audibility accounted for the majority of variance in HINT score: 74% in quiet and 52% in noise. The results of the regression indicate that no experimental variable contributed a significant increase in predictive power. Hearing Handicap Self reported hearing handicap was assessed using the age appropriate version of the Hearing Handicap Inventory: the Hearing Handicap Inventory for Adults (HHIA) for individuals unde r age 65 and the Hearing Handicap Inventory for the Elderly (HHIE) for individuals age 65 and up. Scores on the HHIA/E varied widely: Emotional subscale scores (HHIEMOT) ranged from zero to 44 with a mean of 10.8 and a standard deviation of 9.5. Social subscale scores (HHISOC) ranged from 2 to 40 with a mean of 10.5 and a standard deviation of 6.9. Total scores (HHITOT) ranged from 2 to 74 with a mean of 21.3 and a standard deviation of 15.1. According to published reports, total scores of 18 or highe r on the HHIE and 11 on the HHIA indicate presence of a significant hearing handicap (Jerger et al, 1990; Weinstein, 2000; Weinstein & Ventry, 1983). Of the 56 participants in the present investigation, 32 scored at or above the level indicating handicap. On the HHIA, 14 of 27 participants scored above 11 (mean = 16.4, SD = 12.1); on the HHIE, 18 of 29 participants scored above 18 (mean = 22.4, SD = 13.9). All participants reported difficulty in at least one HHIA/E domain. The most commonly reported di fficulties were in understanding when others speak in a whisper; when listening to conversation while eating at a restaurant; and when listening when at a party. Audibility was a significant predictor of both HHISOC ( r = .45, p = .001) and HHITOT ( r = .3 5, p = .008). The negative valence of these correlations indicates that decreasing AI score was correlated with increasing hearing handicap, particularly the social and situational aspect of hearing handicap. No correlation was seen between HHIEMOT and B ETAI.

PAGE 74

74 A significant positive correlation was seen between age and HHISOC ( r = .49, p < .001) and between age and HHITOT ( r = .34, p =.01), but not between age and HHIEMOT, indicating that perceived social and situational impact of hearing loss, but not the emotional consequences of that hearing loss, increased slightly with age. When controlling for BETAI, no significant correlation was seen between age and HHIA/E scores, indicating that age did not contribute significant predictive power for HHIA/E score when controlling for hearing sensitivity. Because listening in difficult, noisy environments is the principal complaint of most individuals seeking help for a hearing problem, it seems logical to assume that self report measures of hearing handicap would correlate with speech perception ability in noise. To test this assumption, bivariate correlations were calculated between each pair of speech perception measures (HINTQ and HINTN) and HHIA/E scales (HHIEMOT, HHISOC, and HHITOT). In addition, the correl ations among these variables were compared to the independent correlations between each variable and BETAI to evaluate the relative strength of associations. A correlation matrix for these variables is displayed in Table 4 7. Significant correlations were seen between HINTQ and HHISOC ( r = .37, p = .004) and between HINTQ and HHITOT ( r = .30, p = .024). Performance on the HINT in quiet was not significantly correlated with HHIEMOT (p = .128). Correlations were also seen between HINTN and HHIEMOT ( r = .33 p = .012), between HINTN and HHISOC ( r = .45, p < .001), and between HINTN and HHITOT ( r = .42, p = .001). Overall, HHIA/E scores were better predicted by HINTN ( r = .33, .45, and .42 for HHIEMOT, HHISOC, and HHITOT respectively) than by HINTQ ( r = .21, .38, and .30) or by BETAI ( r = .23, .45, and .35). There were significant correlations among all temporal resolution measures (RGIN, LGIN, BETGIN, WRSGIN, and BINGIN) and all HHIA/E subscale and total scores as displayed in the

PAGE 75

75 correlation matrix in Ta ble 4 8. Correlation coefficients for HHIA/E scales ranged from .41 to .45 for RGIN, .30 to .32 for LGIN, .34 to .39 for BETGIN, .34 to .38 for WRSGIN, and .38 to .41 for BINGIN. When controlling for audibility using BETAI, HHIEMOT was significantly cor related with RGIN ( r = .34, p = .010), WRSGIN ( r = .29, p = .034), and BINGIN ( r = .31, p = .023). Significant correlations were also calculated between HHITOT and RGIN ( r = .34, p = .012) and between HHITOT and BINGIN ( r = .28, p = .037). There was only o ne significant association between HHIA/E scores and working memory variables. Social/Situational HHI/E subscale score was weakly correlated with DIGORD ( r = .27, p = .048). Stepwise linear regression procedures were conducted separately for HHIEMOT, HHI SOC, and HHITOT. The predictor variables included were audibility (BETAI), temporal resolution (BINGIN), and working memory (WM). This procedure is identical to the procedure described above which regressed the explanatory variables on HINTQ and HINTN. Only the temporal resolution factor BINGIN was retained as a factor in the regression equation for HHIEMOT ( R = .38, F 1, 54 = 8.85, p = .004). Inclusion of BETAI and WM increased the correlation coefficient from .375 to .380, which was not a significant change (F 2, 52 = .11, p = .895). The audibility factor BETAI was retained as a factor in the regression equation for HHISOC ( R = .45, F 1, 54 = 13.56, p = .001). Inclusion of BINGIN and WM increased the correlation coefficient from .448 to .506, which was not a significant change (F 2, 52 = 1.94, p = .155). Only the temporal resolution factor BINGIN was retained as a factor in the regression equation for HHITOT ( R = .41, F 1, 54 = 11.19, p = .002). Inclusion of BETAI and WM increased the correlation coeffi cient from .414 to .453, an insignificant change (F 2, 52 = 1.09, p = .345).

PAGE 76

76 Gender A post hoc analysis of variance (ANOVA) was conducted to determine whether gender differences existed on any measure examined in the present investigation. Using SPSS, dat a for male participants (N=17) was compared to data for female participants (N=39), yielding the ANOVA results in Table 4 9. No difference was seen between the male and female participants in this investigation on any of the measures compared: age, audib ility (RAI, LAI, BETAI, WRSAI, BINAI), working memory (DIGFWD, DIGBAK, DIGORD, WM), temporal resolution (RGIN, LGIN, BETGIN, WRSGIN, BINGIN), speech perception in quiet or in noise (HINTQ, HINTN), and self reported hearing handicap (HHIEMOT, HHISOC, HHITOT ). Because gender differences were absent for all measures, no further analysis of gender effects was necessary. Analysis of Variance I: Grouping by Age As demonstrated in the present investigation, both temporal resolution (e.g., Gordon Salant & Fitzgib bons, 1993; Schneider et al, 1994; Strouse et al, 1998) and working memory (e.g., Carpenter et al, 1994; Hultsch et al, 1992; Wingfield et al, 1988) show declines in older adulthood. In addition, previous studies have detected changes in middle age for bo th temporal resolution (e.g., Grose et al, 2006) and working memory (e.g., Palladino & De Beni, 1999). The particular, some studies have found that working memory declines in early adulthood and then is stable from middle age to older adulthood until very late in life (e.g., Fisk & Sharp, 2004; Parente, de Taussik, Ferre ira, and Kristensen, 2005; Pelosi & Blumhardt, 1999), while others have found more consistent declines across the lifespan (Salthouse, 1991; Palladino & De Beni, 1999).

PAGE 77

77 Because the participant sample from the present investigation spans almost four decad es within the adult age range (50 to 89), it might be useful to examine age differences within the sample to identify possible age differences in associations among the variables studied. For this analysis, the participant sample was divided into three gr oups by age. Group 1 (late middle aged adults) was composed of participants age 50 to 59. This group had 17 total members (6 male, 11 female) with a mean age of 56.0 years (SD = 3.1) (Figure 4 5). Group 2 (young old adults) was composed of participants a ge 60 to 69. This group had 19 total members (5 male, 14 female) with a mean age of 64.7 years (SD = 2.7) (Figure 4 6). Group 3 (old old adults) was composed of participants age 70 to 89. This group had 20 total members (6 male, 14 female) with a mean a ge of 76.9 years (SD = 4.8) (Figure 4 7). Mean audiometric data for each age group is displayed in Table 4 10. A comparison of the three groups on the experimental measures and dependent variables can be seen in Table 4 11. An ANOVA was conducted to te st for group differences on each of the experimental measures and dependent variables (Table 4 12). Group 1 (late middle aged adults) had significantly higher audibility scores compared to Group 2 (young old adults) as measured by RAI, WORSAI, and BINAI ( p < .05); significantly higher DIGORD scores (p < .05); significantly shorter GIN thresholds as measured by LGIN and WRSGIN (p < .05); and significantly lower HINTQ thresholds (p < .05). No significant differences were seen between Groups 1 and 2 for LAI, BETAI, DIGFWD, DIGBAK, WM, RGIN, BETGIN, BINGIN, HINTN, HHIEMOT, HHISOC, or HHITOT. Group 1 had significantly higher audibility scores compared to Group 3 (old old adults) on all AI measures (p < .001); significantly higher working memory scores as measur ed by DIGFWD (p < .05), DIGBAK (p < .05), DIGORD (p < .01), and WM (p < .01); significantly

PAGE 78

78 shorter thresholds on all GIN measures (p < .01); significantly lower thresholds on HINTQ (p < .001) and HINTN (p < .05); and significantly lower scores on HHISOC ( p < .01). No significant differences were seen between Groups 1 and 3 for HHIEMOT or HHITOT. Group 2 had significantly higher audibility scores compared to Group 3 as measured by RAI (p < .01), LAI (p < .05), BETAI (p < .01), WORSAI (p < .05), and BINAI ( p < .01); significantly higher working memory scores as measured by DIGFWD (p < .01) and WM (p < .01); significantly lower thresholds on HINTQ (p < .01) and HINTN (p < .05); and significantly lower scores on HHISOC (p < .05). No significant differences we re seen between Groups 2 and 3 for DIGBAK, DIGORD, any GIN measure, HHIEMOT, or HHITOT. Multiple Regression, Grouping by Age Separate stepwise regression analyses were conducted for Groups 1 through 3 to determine whether the set of measures best predictin g the dependent variables (HINTQ, HINTN, HHIEMOT, HHISOC, HHITOT) differed between age subsets. Each group regression analysis was conducted in the same manner as the analyses conducted on the entire participant group as described above. HINT in Quiet Con dition For Group 1 (age 50 59), only BETAI was retained as a predictor variable in the regression equation for HINTQ ( R = .50, F 1, 15 = 4.95, df = p = .042). Inclusion of BINGIN and WM increased the correlation coefficient from .498 to .513, which was n ot a significant change (F 2, 13 = .13, p = .881). For Group 2 (age 60 69), only BETAI was retained as a predictor variable in the regression equation for HINTQ ( R = .81, F 1, 17 = 33.40, p < .001). Inclusion of BINGIN and WM increased the correlation co efficient from .814 to .839, which was not a significant change (F 2, 15 = 1.05, p = .375).

PAGE 79

79 For Group 3 (age 70 89), both BETAI and BINGIN were retained as predictor variables in the regression equation for HINTQ ( R = .86, F 2, 17 = 23.99, p < .001). Incl usion of WM increased the correlation coefficient from .859 to .863, which was not a significant change (F 1, 16 = .35, p = .565). HINT in Noise Condition For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) an d HINTN; therefore, no regression model was computed. For Group 2, only BETAI was retained as a predictor variable in the regression equation for HINTN ( R = .47, F 1, 17 = 4.74, p = .044). Inclusion of BINGIN and WM increased the correlation coefficient fr om .467 to .469, which was not a significant change (F 2, 15 = .02, p = .982). For Group 3, only BETAI was retained as a predictor variable in the regression equation for HINTN ( R = .90, F 1, 18 = 77.76, p < .001). Inclusion of BINGIN and WM increased the c orrelation coefficient from .901 to .911, which was not a significant change (F 2, 16 = .83, p = .452). HHIA/E Emotional Subscale For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHIEMOT; therefore, no regression model was computed. For Group 2, only BINGIN was retained as a predictor variable in the regression equation for HHIEMOT ( R = .50, F 1, 17 = 5.54, p = .031). Inclusion of BETAI and WM increased the correlation coefficient from .496 to .532, whic h was not a significant change (F 2, 15 = .39, p = .681). For Group 3, only BINGIN was retained as a predictor variable in the regression equation for HHIEMOT ( R = .45, F 1, 18 = 4.62, p = .046). Inclusion of BETAI and WM increased the

PAGE 80

80 correlation coefficie nt from .452 to .481, which was not a significant change (F 2, 16 = .29, p = .754). HHIA/E Social/Situational Subscale For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHISOC; therefore, no regression m odel was computed. For Group 2, only BETAI was retained as a predictor variable in the regression equation for HHISOC ( R = .60, F 1, 17 = 9.69, p = .006). Inclusion of BINGIN and WM increased the correlation coefficient from .603 to .639, which was not a s ignificant change (F 2, 15 = .58, p = .574). For Group 3, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHISOC; therefore, no regression model was computed. HHIA/E Total For Group 1, no significant correlation wa s seen between any predictor variable (BETAI, BINGIN, WM) and HHITOT; therefore, no regression model was computed. For Group 2, only BINGIN was retained as a predictor variable in the regression equation for HHITOT ( R = .49, F 1, 17 = 5.25, p = .035). Incl usion of BETAI and WM increased the correlation coefficient from .486 to .586, which was not a significant change (F 2, 15 = 1.23, p = .319). For Group 3, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHITOT; the refore, no regression model was computed. Analysis of Variance II: Grouping by Hearing Loss A second ANOVA was conducted after dividing the participant sample into two groups by degree of hearing loss, as measured by BETAI. This division was intended to separate those participants classified clinically as audiometrically normal from those classified as having a

PAGE 81

81 hearing loss. A cutoff BETAI of .98 was used to separate the participant sample into two groups. An AI score of .98 (.976 unrounded) or greater equates to hearing thresholds within normal limits (better than or equal to 20 dB HL) at all frequencies included in the AI calculation (250, 500, 1000, 2000, and 4000 Hz). A .98 AI also corresponds to a prediction of 98% or better speech intelligibility in quiet environments. That is, participants with BETAI greater than or equal to .98 (Group 1) would be considered to have hearing within normal limits and, according to the theory underlying AI, excellent speech intelligibility. This group contained 24 participants (5 male, 19 female) with a mean BETAI of .99 (SD = .02) (Figure 4 8). Participants with BETAI less than .98 (Group 2) would be considered to have at least a mild hearing loss in the better ear and poorer than 98% predicted speech intelligibil ity. This group contained 32 participants (12 male, 20 female) with a mean BETAI of .73 (SD = .14) (Figure 4 9). Table 4 13 summarizes mean audiometric data for each group. A comparison of the two groups on the experimental measures and dependent variab les can be seen in Table 4 14. Previous studies have suggested that listeners with SNHL are especially susceptible to environmental distortions of speech such as noise (e.g., Plomp & Mimpen, 1979). It is possible that limiting the analysis to only those participants with measurable loss of hearing sensitivity might reveal relations among between the predictor and dependent variables that are not as clear when participants with normal or near normal audiograms are included. A one way ANOVA was conducted t o test for group differences on the experimental measures and dependent variables (Table 4 15). Not surprisingly, Group 1 was significantly younger than Group 2. Group 1 had mean age 59.4 (SD = 5.6) while Group 2 had mean age 71.6 (SD = 8.2).

PAGE 82

82 Group 1, t he group with audiometrically normal or near normal hearing had significantly higher scores on all AI measures (p < .001); significantly shorter thresholds on all GIN measures (p < .01), significantly lower thresholds on HINTQ (p < .001) and HINTN (p < .01 ); and significantly lower scores on HHISOC (p < .05) than Group 2, the group with some audiometrically measurable hearing loss. There were no significant group differences on any working memory measure (DIGFWD, DIGBAK, DIGORD, and WM) or on HHIA/E Emoti onal subscale or total scores. Multiple Regression, Grouping by Hearing Loss Separate stepwise regression analyses were conducted for Group 1 and Group 2 to determine whether the set of measures best predicting the dependent variables (HINTQ, HINTN, HHIEMO T, HHISOC, HHITOT) differed between the set of participants with audiometrically measurable hearing loss (thresholds worse than 20 dB) and those without. Each group regression analysis was conducted in the same manner as the analyses conducted on the enti re participant group as described above. HINT in Quiet Condition regression equation for HINTQ ( R = .60, F 1, 22 = 12.40, p = .002). Inclusion of BETAI and WM increased the correlation coefficient from .600 to .636, which was not a significant change (F 2, 20 = .73, p = .492). For Group 2 (BETAI < .98), only BETAI was retained as a predictor variable in the regression equation for HINTQ ( R = .74, F 1, 30 = 37.18, p < .001). I nclusion of BINGIN and WM increased the correlation coefficient from .744 to .769, which was not a significant change (F 1, 28 = 1.28, p = .294).

PAGE 83

83 HINT in Noise Condition For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HINTN; therefore, no regression model was computed. For Group 2, only BETAI was retained as a predictor variable in the regression equation for HINTN ( R = .78, F 1, 30 = 45.27, p < .001). Inclusion of BINGIN and WM increased the correlatio n coefficient from .776 to .804, which was not a significant change (F 2, 28 = 1.76, p = .190). HHIA/E Emotional Subscale For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHIEMOT; therefore, no regressi on model was computed. For Group 2, only BINGIN was retained as a predictor variable in the regression equation for HHIEMOT ( R = .40, F 1, 30 = 5.71, df = 1, 30, p = .023). Inclusion of BETAI and WM increased the correlation coefficient from .400 to .451, which was not a significant change (F 2, 28 = .76, p = .476). HHIA/E Social/Situational Subscale For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHISOC; therefore, no regression model was computed. For Group 2, only BETAI was retained as a predictor variable in the regression equation for HHISOC ( R = .36, F 1, 30 = 4.43, p = .044). Inclusion of BINGIN and WM increased the correlation coefficient from .359 to .502, which was not a significant change (F 2, 28 = 2.31, p = .118). HHIA/E Total For Group 1, no significant correlation was seen between any predictor variable (BETAI, BINGIN, WM) and HHITOT; therefore, no regression model was computed.

PAGE 84

84 For Group 2, only BINGIN was retained as a predictor variable i n the regression equation for HHITOT ( R = .39, F 1, 30 = 5.28, p = .029). Inclusion of BETAI and WM increased the correlation coefficient from .387 to .493, which was not a significant change (F 2, 28 = 1.74, p = .195). Summary To review, the participant gr oup showed the expected relations among predictor and descriptor variables. Significant declines in hearing sensitivity, temporal resolution, and working memory were seen with increasing age. Hearing sensitivity was correlated with temporal resolution th reshold, with speech perception ability, and with self reported hearing handicap, but not with working memory ability. As anticipated, declines in hearing sensitivity, temporal resolution, and working memory were significantly correlated with decreased s peech perception ability and increased hearing handicap. Loss of hearing sensitivity was the strongest predictor of speech perception difficulty, accounting for the majority of the variance in HINT threshold both in quiet and in noise conditions. When he aring sensitivity was controlled, neither temporal resolution ability nor working memory capacity was significantly correlated with speech perception threshold. Self reported hearing handicap was significantly but weakly associated with declines in hearing sensitivity, temporal resolution, and speech perception ability. No association was seen between working memory and hearing handicap score. The strongest predictor of hearing handicap was speech perception ability in noise. Multiple regression procedur es revealed that, of the experimental variables, temporal resolution threshold was the best predictor of total HHIA/E score, although no combination of predictors could account for more than approximately 17% of the variance in hearing handicap.

PAGE 85

85 Analysis o f the participant sample separated into three age groups confirmed significant age related declines in all measures except HHIA/E Emotional subscale and total scores. Separate regression procedures for these three groups revealed changing associations bet ween the experimental and dependent variables with increasing age. For the two older groups, speech perception thresholds in quiet and in noise were predicted increasingly well by hearing sensitivity and, in the case of the oldest group for speech percep tion in quiet, by hearing sensitivity and temporal resolution together. For the youngest group, speech perception was not well predicted by any measure. Hearing handicap scores were predicted poorly or not at all by the experimental measures in each age group. A second grouped analysis of the sample separated participants with normal hearing from participants with hearing loss as determined by pure tone thresholds. Restricting the analysis only to those participants with audiometrically measurable hearin g loss did not significantly improve the predictive ability of the experimental variables for speech perception or hearing handicap.

PAGE 86

86 Table 4 1. Comparison of pure tone audiogram status to DPOAE status in each DPOAE frequency group. Normal DPOAE, Normal Pure Tone Normal DPOAE, Abnorm al Pure Tone Abnormal DPOAE, Normal Pure Tone Abnormal DPOAE, Abnorm al Pure Tone Total Right High 4 0 5 47 56 Left High 3 0 4 49 56 Right Mid 25 0 6 25 56 Left Mid 21 4 5 26 56 Right Low 16 0 25 15 56 Left Low 14 0 26 16 56

PAGE 87

87 Table 4 2. Mean audiometric data (N= 56). 250 Hz 500 Hz 1000 Hz 2000 Hz 4000 Hz 8000 Hz PTA SRT WRS Right Ear Mean 12.7 12.8 13.2 18.5 34.0 49.3 14.8 15.5 94.3 Right Ear SD 9.3 8.4 10.6 15.2 22.5 27.0 10.4 10.7 10.0 Left Ear Mean 12.3 13.7 14.7 20.6 39.6 48.9 16.3 17.0 93.1 Left Ear SD 11.8 13.0 16.0 18.2 21.6 24.2 14.6 14.6 12.5 PTA = pure tone average (500, 1000, 2000 Hz) in dB HL; SRT = speech reception threshold in dB HL; WRS = word recognition score in %

PAGE 88

88 Table 4 3. Correlation matrix for age, audibility, and GIN variables (N = 56). AGE RAI LAI BETAI BINAI RGIN LGIN BET GIN WRS GIN BIN GIN AGE r 1 .751 .571 .765 .715 .366 .444 .393 .447 .455 Sig. .000 .000 .000 .000 .006 .001 .003 .001 .000 RAI r .751 1 .655 .963 .888 .509 .452 .554 .440 .524 Sig. .000 .000 .000 .000 .000 .000 .000 .001 .000 LAI r .571 .655 1 .717 .929 .386 .635 .441 .629 .589 Sig. .000 .000 .000 .0 00 .003 .000 .001 .000 .000 BET AI r .765 .963 .717 1 .908 .531 .485 .562 .484 .555 Sig. .000 .000 .000 .000 .000 .000 .000 .000 .000 BIN AI r .715 .888 .929 .908 1 .484 .608 .539 .598 .615 Sig. .000 .000 .000 .000 .000 .000 .000 .000 .000 R GIN r .366 .509 .386 .531 .484 1 .632 .939 .703 .862 Sig. .006 .000 .003 .000 .000 .000 .000 .000 .000 L GIN r .444 .452 .635 .485 .608 632 1 .744 .971 .938 Sig. .001 .000 .000 .000 .000 .000 .000 .000 .000 BET GIN r .393 .554 .441 .562 .539 .939 .744 1 .733 .907 Sig. .003 .000 .001 .000 .000 .000 .000 .000 .000 WRS GIN r .4 47 .440 .629 .484 .598 .703 .971 .733 1 .951 Sig. .001 .001 .000 .000 .000 .000 .000 .000 .000 BIN GIN r .455 .524 .589 .555 .615 .862 .938 .907 .951 1 Sig. .000 .000 .000 .000 .000 .000 .000 .000 .000 All correlations are significant at the 0.01 level (2 tailed). RAI = Articulation Index in right ear; LAI = Articulation Index in left ear; BETAI = Articulation Index in better ear; BINAI = average binaural Articulation Index; RGIN = Gaps In Noise t hreshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise threshold in better ear; WRSGIN = Gaps In Noise threshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears

PAGE 89

89 Table 4 4. Comparison of pa rticipant mean Gaps In Noise threshold values to normative data collected by Musiek et al (2005) on 50 subjects (age 13 to 46 years). Participant Group Normative Group Right Left Right Left Mean GIN 7.6 ms 8.4 ms 4.9 ms 4.8 ms SD 2.1 3.0 1.0 1.0 Max GIN 15 ms 20 ms 8 ms 8 ms Min GIN 5 ms 5 ms 4 ms 3 ms Mean PTA 14.8 dB 16.3 dB 3.1 dB 4.1 dB N 56 50 Mean Age 66.4 24.6 SD = standard deviation of mean; Max GIN= highest GIN threshold in group; Min GIN= lowest GIN threshold in group; Mean PTA = group mean pure tone average (500, 1 k, 2 kHz); N = number of participants in sample

PAGE 90

90 Table 4 5. Comparison of participant working memory scores to UF Language Over the Lifespan Laboratory normative data. Participant Group Normative Data Age Mean SD N Me an SD N t df DIG FWD 50 54 8.63 2.26 8 -----55 59 8.33 2.06 9 -----60 64 8.20 1.81 10 7.92 2.70 12 0.2793 20 65 69 9.78 2.17 9 8.75 2.70 32 1.0499 39 70 74 6.78 1.92 9 8.02 2.50 48 1.4081 55 75 79 7.50 1.22 6 8.02 2.20 58 0.5672 62 80 84 7.00 1.63 4 7.63 2.40 40 0.5105 42 85 and up 5.00 n/a 1 7.44 2.50 16 n/a n/a DIG BAK 50 54 7.38 2.72 8 -----55 59 7.56 1.67 9 -----60 64 6.70 2.26 10 6.08 1.80 12 0.7168 20 65 69 7.33 1.41 9 7.63 2.70 32 0.3193 39 70 74 6.00 1.87 9 7.13 2.50 48 1.2862 55 75 79 6.00 2.61 6 7.00 2.30 58 1.0023 62 80 84 6.00 1.83 4 6.35 1.80 40 0.3703 42 85 and up 4.00 n/a 1 6.88 1.70 16 n/a n/a DIG ORD 50 54 18. 63 3.81 8 -----55 59 18.44 2.70 9 -----60 64 16.60 2.95 10 16.50 3.20 12 0.0756 20 65 69 16.33 2.24 9 18.37 3.40 32 1.6914 39 70 74 15.44 3.84 9 17.23 3.00 48 1.5713 55 75 79 14.83 6.79 6 17.34 3.30 58 1.5796 62 80 84 14.25 2.22 4 16.27 3.80 41 1.0390 43 85 and up 5.00 n/a 1 15.41 4.70 17 n/a n/a DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; SD = standard deviation of mean; N = number of participants in sample; t = two tail ed t statistic for difference between participant and normative data; df = degrees of freedom

PAGE 91

91 Table 4 6. Correlation matrix for GIN and working memory variables (N = 56). DIG FWD DIG BAK DIG ORD WM RGIN LGIN BET GIN WRS GIN BIN GIN DIG FWD r 1 492 ** 366 ** 764 ** 033 095 078 068 077 Sig. 000 006 000 807 486 569 620 572 DIG BAK r 492 ** 1 598 ** 859 ** 023 131 072 103 096 Sig. 000 000 000 865 336 599 450 481 D IG ORD r 366 ** 598 ** 1 808 ** .268 .146 .222 .185 .215 Sig. 006 000 000 046 285 100 172 111 WM r 764 ** 859 ** 808 ** 1 .087 033 .030 .006 .017 Sig. 000 000 000 524 809 827 965 900 RGIN r 033 023 .268 087 1 632 ** 939 ** 703 ** 862 ** Sig. 807 865 046 524 000 000 000 000 LGIN r 095 131 .146 033 632 ** 1 744 ** 971 ** 938 ** Sig. 486 336 285 809 000 000 000 000 BET GIN r 078 072 .222 .030 939 ** 744 ** 1 733 ** 907 ** Sig. 569 599 100 827 000 000 000 000 WRS GIN r 068 103 .185 .006 703 ** 971 ** 733 ** 1 951 ** Sig. 620 450 172 965 000 000 000 000 BIN GIN r 077 096 .215 .017 862 ** 938 ** 907 ** 951 ** 1 Sig. 572 481 111 900 000 000 000 000 = correlation is significant at the .05 level. ** = correlation is significant at the .01 level. DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; WM = composite working memory score ; RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise thresho ld in better ear; WRSGIN = Gaps In Noise threshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears

PAGE 92

92 Table 4 7. Correlation matrix for HINT, HHIA/E, and audibility variables (N = 56). HINTQ HINTN HHI EMOT HHISOC HHITOT BET AI HINTQ r 1 660** 206 375** 302* .858** Sig. 000 128 004 024 000 HINTN r 660** 1 332* 452** 417** .722** Sig. 000 012 000 001 000 HHI EMOT r 206 332* 1 679** 942** .231 Sig. 128 012 000 000 087 HHISOC r 375** 452** 679** 1 886** .448** Sig. 004 000 000 000 001 HHITOT r 302* 417** 942** 886** 1 .351** Sig. 024 001 000 000 008 BETAI r .858** .722** .231 .448** .351** 1 Sig. 000 000 087 001 008 = correlation is significant at the .05 level. ** = correlation is significant at the .01 level. HINTQ = HINT threshold in quiet; HINTN = HINT threshold in noise; HHIEMOT = HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale score; HHITOT = HHIA/E total score; BETAI = Articulation Index score in the better ear

PAGE 93

93 Table 4 8. Correlation matrix for GIN and HHIA/E variables (N = 56). RGIN LGIN BETGIN WRS GIN BINGIN HHI EMOT HHISOC HHITOT RGIN r 1 632** 939** 703** 862** 406** 432** 454** Sig. .000 .000 .000 000 002 001 000 LGIN r 632** 1 744** 971** 938** .295* .297* .322* Sig. 000 000 000 000 027 026 015 BET GIN r 939** 744** 1 733** 907** 342** 389** 394** Sig. 000 000 000 000 010 003 003 WRS GIN r 703** 971** 733** 1 951** 356** .341* 381** Sig. 000 000 000 000 007 010 004 BIN GIN r 862** 938** 907** 951** 1 375** 388** 414** Sig. 000 000 000 000 004 003 002 HHI EMOT r 406** .295* 342** 356** 375** 1 679** 942** Sig. 002 027 01 0 007 004 000 000 HHI SOC r 432** .297* 389** .341* 388** 679** 1 .886** Sig. 001 026 003 010 003 000 000 HHI TOT r 454** .322* 394** 381** 414** 942** 886** 1 Sig. 000 015 003 004 002 000 000 = correlation is significant at the .05 level. ** = correlation is significant at the .01 level. RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise threshold in better ear; WRSGIN = Gaps In Noise threshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears; HHIEMOT = HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale score; HHITOT = HHIA/E total score

PAGE 94

94 Table 4 9. ANOVA table for comparison of male and female participant data on age, audibility, GIN, working memory, HINT, and HHIA/E variables. Female (N = 39) Male (N = 17) ANOVA Results Mean SD Mean SD F Sig. AGE 66.16 8.99 66.91 10.44 .074 .787 RAI 85 .18 .78 .19 1.342 .252 LAI .78 .25 .77 .18 .042 .839 BETAI .87 .16 .81 .18 1.228 .273 WRSAI .76 .26 .74 .18 .101 .752 BINAI .82 .19 .78 .18 .474 .494 DIGFWD 8.13 2.13 8.00 2.10 .043 .836 DIGBAK 6.92 2.23 6.35 1.69 .886 .351 DIGORD 16.31 4.22 16.59 3.66 .057 .813 WM .08 2.58 .18 2.11 .139 .711 RGIN 7.56 2.14 7.71 1.93 .055 .815 LGIN 8.33 2.77 8.47 3.63 .024 .877 BETGIN 7.28 2.20 7.12 1.90 .072 .790 WRSGIN 8.62 2.60 9.06 3.40 .284 .596 BINGIN 7.95 2.25 8.09 2.50 .043 .837 HINTQ 38.72 9.29 38.53 9.64 .005 .946 HINTN 72.03 2.31 72.65 3.27 .664 .419 HHIEMOT 11. 08 9.85 10.24 8.97 .091 .764 HHISOC 10.05 5.53 11.41 9.48 .455 .503 HHITOT 21.13 14.16 21.65 17.48 .014 .907 RAI = Articulation Index in right ear; LAI = Articulation Index in left ear; BETAI = Articulation Index in better ear; WRSAI = Art iculation Index in the worse ear; BINAI = average binaural Articulation Index; DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; WM = composite working memory score ; RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise threshold in better ear; WRSGIN = Gaps In Noise threshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears; HINTQ = HINT threshold in quiet; HINTN = HINT threshold in nois e; HHIEMOT = HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale score; HHITOT = HHIA/E total score

PAGE 95

95 Table 4 10. Mean audiometric data, separated by age group. Right Ear Age 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz PTA SRT WRS AI 50 59 8.5 7.7 7.9 8.5 16.2 26.8 8.0 9.4 98.8 .95 SD 6.8 5.9 5.6 9.0 13.6 16.8 5.6 6.1 2.7 .09 60 69 10.0 11.1 11.1 16.6 29.7 47.4 12.9 13.4 95.1 .86 SD 7.6 7.0 9.8 13.1 17.2 22.4 9.0 10.0 6.5 .16 70 89 18.8 18.8 19.8 28.8 43.3 53.3 22.4 22.8 89.8 .68 SD 9.7 7.9 11.6 15.3 17.9 18.3 10.2 10.5 14.2 .18 Left Ear Age 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz PTA SRT WRS AI 50 59 7.1 10.0 7.1 9.7 25.3 32.1 8.9 11.5 97.8 .92 SD 5.9 6.4 5.9 7.6 12.8 17.5 5.7 7.2 4.0 .09 60 69 12.9 13.7 14.5 18.2 36.1 46.3 15.4 15.0 95.7 .80 SD 15.0 14.8 19.1 17.4 23.3 23.7 16.2 16.3 6.1 .24 70 89 16.3 16.8 21.5 32.3 46.5 55.0 23.5 23.8 86.8 .64 SD 1 0.8 14.9 16.3 19.2 19.7 15.9 15.4 15.6 18.2 .22 PTA = pure tone average (500, 1 k, 2 kHz); SRT = speech reception threshold; WRS = word recognition score (in %); AI = articulation index

PAGE 96

96 Table 4 11. Comparison across participant age groups for descript ive factors, experimental measures, and dependent variables. Age Male Female N DIG FWD DIG BAK DIG ORD RGIN LGIN HINT Q HINT N HHI EMOT HHI SOC HHI TOT 50 59 56.0 6 11 17 8.5 7.5 18.5 6.8 6.7 31.9 71.3 10.2 7.2 17.4 SD 3.1 2.1 2.2 3.2 1.4 1.6 4.1 1.8 9.9 5.0 13.7 60 69 64.7 5 14 19 9.0 7.0 16.5 7.6 8.4 37.4 71.6 9.5 9.6 19.1 SD 2.7 2.1 1.9 2.6 2.6 3.0 7.9 1.8 6.5 5.4 10.7 70 89 76.9 6 14 20 7.0 5.9 14.5 8.3 9.9 45.6 73.5 12.6 14.1 26.7 SD 4.8 1.6 2.0 4.9 1.8 3.3 9.2 3.4 11.6 8.1 18.5 N = number of participants in group; DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; HINTQ = Hearing In Noise Test threshold in quiet condition; HINTN = Hearing In Noise Test threshold in noise condition; HHIEMOT = Hearing Handicap Inventory for Adults / for the Elderly Emotional subscale score; HHISOC = Hearing Han dicap Inventory for Adults / for the Elderly Social/Situational subscale score; HHITOT = Hearing Handicap Inventory for Adults / for the Elderly Emotional aggregate score

PAGE 97

97 97 Table 4 12. ANOVA table for comparison of participant data, grouped by age, on ag e, audibility, working memory, HINT, and HHIA/E variables. Age 50 59 (N = 17) Age 60 69 (N = 19) Age 70 89 (N = 20) ANOVA Results Mean SD Mean SD Mean SD F Sig. AGE 55.99 3.10 64.67 2.73 76.86 4.83 148.19 .000** RAI .95 .09 .8 6 .16 .68 .18 16.09 .000** LAI .92 .09 .80 .24 .64 .22 9.53 .000** BETAI .96 .07 .89 .14 .71 .16 18.19 .000** WRSAI .92 .10 .77 .24 .61 .22 10.51 .000** BINAI .94 .08 .83 .16 .66 .18 15.94 .000** DIGFWD 8.47 2.10 8.95 2.10 6.95 1.64 5.62 .006** DIGBAK 7.47 2.15 7.00 1.89 5.90 2.00 3.03 .057 DIGORD 18.53 3.17 16.47 2.57 14.50 4.94 5.34 .008** WM 1.06 2.34 .55 1.83 1.42 2.42 6.64 .003** RGIN 6.76 1.44 7.63 2.59 8.30 1.75 2.71 .076 LGIN 6.65 1.58 8.37 3.04 9.85 3.25 6.14 .004** BETGIN 6.29 1.40 7.26 2.66 8.00 1.72 3.30 .045* WRSGI N 7.12 1.50 8.74 2.83 10.15 3.08 6.22 .004** BINGIN 6.71 1.38 8.00 2.56 9.08 2.20 5.69 .006** HINTQ 31.91 4.08 37.35 7.88 45.64 9.18 15.79 .000** HINTN 71.32 1.80 71.63 1.74 73.54 3.38 4.55 .015* HHIEMOT 10.24 9.90 9.47 6.49 12.60 11.59 .56 .573 HHISOC 7.18 4.95 9.58 5.44 14.10 8.07 5.68 .006** HHITOT 17.41 13.73 19.05 10.65 26.70 18.53 2.14 .127 = difference is significant at the .05 level. ** = difference is significant at the .01 level. RAI = Articulation Index in right ear; LAI = Articulation Index in left ear; BETAI = Articulation Index in better ear; WRSAI = Articulation Index in the worse ear; BINAI = average binaural Articulation Index; DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; WM = composite working memory score ; RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise threshold in better ear; WRSGIN = Gaps In Noise thr eshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears; HINTQ = HINT threshold in quiet; HINTN = HINT threshold in noise; HHIEMOT = HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale score; HHITOT = H HIA/E total score

PAGE 98

98 98 Table 4 13. Mean audiometric data, separated by hearing loss group. Right Ear 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz PTA SRT WRS AI Group 1 7.3 7.1 6.3 7.5 13.5 29.2 6.9 7.3 98.7 .99 6.6 5.3 4.2 6.4 7.9 13.8 3.9 5.5 2.6 .03 Group 2 16.7 17.0 18.4 26.7 49.4 66.4 20.7 21.7 91.1 .71 9.0 7.7 11.0 14.7 16.8 25.4 9.9 9.3 12.1 .16 Left Ear 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz PTA SRT WRS AI Group 1 8.5 10.8 9.2 10.4 21.9 29.6 10.1 11.9 98. 0 .92 12.1 13.0 16.1 14.4 15.2 16.4 14.0 15.0 4.6 .20 Group 2 15.2 15.8 18.9 28.3 52.8 67.5 21.0 20.9 89.5 .68 10.8 12.7 14.9 17.1 15.2 20.9 13.4 13.3 15.1 .19 PTA = pure tone average (500, 1 k, 2 kHz); SRT = speech reception threshold; WRS = word recognition score (in %); AI = articulation index

PAGE 99

99 Table 4 factors, experimental measures, and dependent variables. Age Male Fema le N DIG FWD DIG BAK DIG ORD RGIN LGIN HINT Q HINT N HHI EMOT HHI SOC HHI TOT Group 1 59.4 5 19 24 8.4 6.9 17.4 6.6 7.1 31.2 71.0 9.5 7.8 17.3 SD 5.6 2.3 2.2 3.6 1.3 2.2 3.7 1.4 8.8 4.8 12.1 Group 2 71.6 12 20 30 7.8 6.6 15.6 8.3 9.3 44.2 73.1 11.8 12.4 24.3 SD 8.2 1.9 2.0 4.3 2.2 3.3 8.3 3.0 10.0 7.6 16.6 N = number of participants in group; DIGFWD = forward digit span; DIGBAK = backward digit span; DIGORD = digit ordering; RGIN = Gaps In N oise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; HINTQ = Hearing In Noise Test threshold in quiet condition; HINTN = Hearing In Noise Test threshold in noise condition; HHIEMOT = Hearing Handicap Inventory for Adults / for the Elder ly Emotional subscale score; HHISOC = Hearing Handicap Inventory for Adults / for the Elderly Social/Situational subscale score; HHITOT = Hearing H andicap Inventory for Adults / for the Elderly Emotional aggregate score

PAGE 100

100 Table 4 15. ANOVA table for compa rison of participant data, grouped by hearing loss (Group HINT, and HHIA/E scale scores. Group 1 (N = 24) Group 2 (N = 30) ANOVA Results Mean SD Mean SD F Sig. AGE 59.43 5.61 71.60 8.16 39.35 .000** RAI .99 .03 .71 .16 75.42 .000** LAI .92 .20 .68 .19 21.40 .000** BETAI 1.00 .01 .74 .14 82.49 .000** WRSAI .91 .20 .64 .19 25.17 .000** BINAI .95 10 .69 .16 51.33 .000** DIGFWD 8.42 2.30 7.84 1.94 1.02 .317 DIGBAK 6.92 2.17 6.63 2.04 .27 .609 DIGORD 17.38 3.56 15.66 4.25 2.57 .115 WM .36 2.39 .48 2.46 1.66 .204 RGIN 6.62 1.28 8.34 2.24 11 .35 .001** LGIN 7.12 2.15 9.31 3.26 8.14 .006** BETGIN 6.17 1.05 8.03 2.34 13.25 .001** WRSGIN 9.63 3.06 7.58 2.06 7.97 .007** BINGIN 6.88 1.35 8.83 2.52 11.80 .001** HINTQ 31.24 3.68 44.23 8.31 51.01 .000** HI NTN 71.03 1.44 73.11 2.96 10.02 .003** HHIEMOT 9.50 8.81 11.81 10.04 .81 .373 HHISOC 7.83 4.79 12.44 7.63 6.73 .012* HHITOT 17.33 12.07 24.25 16.57 2.99 .090 = correlation is significant at the .05 level. ** = correlation is sig nificant at the .01 level. RAI = Articulation Index in right ear; LAI = Articulation Index in left ear; BETAI = Articulation Index in better ear; WRSAI = Articulation Index in the worse ear; BINAI = average binaural Articulation Index; DIGFWD = forward di git span; DIGBAK = backward digit span; DIGORD = digit ordering; WM = composite working memory score ; RGIN = Gaps In Noise threshold in right ear; LGIN = Gaps In Noise threshold in left ear; BETGIN = Gaps In Noise threshold in better ear; WRSGIN = Gaps In Noise threshold in worse ear; BINGIN = average Gaps In Noise threshold in right and left ears; HINTQ = HINT threshold in quiet; HINTN = HINT threshold in noise; HHIEMOT = HHIA/E Emotional subscale score; HHISOC = HHIA/E Social/Situational subscale score; H HITOT = HHIA/E total score

PAGE 101

101 Figure 4 1. Distribution of distortion product otoacoustic emissions results within low, middle, and high frequency ranges. Right Ear DPOAE 0 5 10 15 20 25 30 35 40 500 1000 Hz 1000 4000 Hz 40 00 8000 Hz Test Frequency Range Normal Abnormal Absent # of Participants Left Ear DPOAE 0 5 Normal Abnormal Absent Test Frequency Range 40 35 30 10 15 20 25 500 1000 Hz 1000 4000 Hz 4000 8000 Hz # of Participants

PAGE 102

102 Figure 4 2. Mean audiogram with standard deviation bars. dB HL Freq uency

PAGE 103

103 Figure 4 3. Comparison of Gaps In Noise test thresholds between right and left ears. 0 5 10 15 20 25 2 3 4 5 6 8 10 12 15 20 GIN Threshold (ms) # of Participants Right Left

PAGE 104

104 Figure 4 4. Comparison of participant mean Gaps In Noise threshold values (with s tandard deviation bars) to normative data collected by Musiek et al (2005) on 50 subjects (age 13 to 46 years). 0.0 2.0 4.0 6.0 8.0 10.0 12.0 Participant Group Normative Group Right Left GIN threshold (ms)

PAGE 105

105 Figure 4 5. Mean audiogram for participants age 50 59 (N = 17). dB HL Frequency

PAGE 106

106 Figure 4 6. Mean audiogram for participants age 60 69 (N = 1 9). dB HL Frequency

PAGE 107

107 Figure 4 7. Mean audiogram for participants age 70 89 (N = 20). dB HL Frequency

PAGE 108

108 Figure 4 8 Mean audiogram for HL Group 1 ) (N = 24 ). dB HL Frequency

PAGE 109

109 Figure 4 9. Mean audiogram for HL Group 2 (BETAI < .98 ) (N = 32 ). dB HL Frequency

PAGE 110

110 CHAPTER 5 DISCUSSION The inability of the standard audiometric test battery to account well for speech perception ability in difficult listening situations and self reported hearing handicap is well known. Audiologists commonly encounter patients whose hearing problems are not we ll explained by the audiogram, or for whom certain listening situations are more difficult than would be expected based on measurable loss of hearing. Two patients evaluated by the same audiologist, and with similar audiometric profiles, may report and ex perience everyday listening difficulties that are quite different in nature and in severity. Without appropriate test criteria to distinguish these difficulties, the audiologist may manage both patients similarly. The outcome for each patient may, however vary substantially. It is likely that both audiologic and non audiologic factors contribute to hearing handicap and success with rehabilitation. An audiologist may be able to identify some of the differences among patients using good, sensitive tests o f complex auditory capabilities. However, factors that are not dependent on hearing status (such as attention and memory are generally) may be more difficult to measure. In addition, the influence of both audiologic and non audiologic factors may be very different from person to person. A construction foreman whose job requires him to communicate in very noisy situations may rely heavily on his eyesight to compensate for his hearing loss by speechreading. A receptionist at a busy law practice may depend on attention and working memory to listen and take messages rapidly on the phone. An elderly retiree might have severe difficulty keeping up in evening conversations with her husband due to her inability to focus on a single speech signal in the presence of auditory and visual distractions from the television.

PAGE 111

111 To what degree can an audiologist account for the uniqueness of each of his or her account for all individual variations using any combination of tests or questionnaires, there is certainly room for improvement over standard clinical methods by examining factors that are not tested in the typical clinical battery. The present investigation examined the roles of two of these potential factors, auditory temporal resolution and working memory capacity, in prediction of speech perception performance and self reported hearing handicap. Previous investigations had suggested that both of these factors may exp lain part of the variance in speech perception performance which is not accounted for by loss of hearing sensitivity. In this investigation, both factors were tested using clinically applicable measures. That is, each measure could be added easily to a clinical audiologic test battery with minimal additional training to the audiologist or costly equipment. The fundamental aim of this study was to determine whether the GIN test and a simple working memory battery could replicate or approximate the explan atory value of the factors shown in previous studies using more complex testing. If so, these simple tests might be valuable tools both for audiologic diagnosis and rehabilitation planning. Sampling Method Fifty six community dwelling adults age 50 to 8 9 who were aware of a hearing problem served as participants. This group represented a fairly unselected sample of upper middle aged and older adults with a range of hearing losses. The mean audiogram (Figure 4 2) shows that, on average, this group had a generally symmetrical moderate sloping hearing loss bilaterally, with substantial variance within the group. The hearing loss profile for participants in the investigation was representative of an older adult population. Unlike most previous studies of hearing impairment and hearing handicap, the sample in the present investigation was not taken

PAGE 112

112 from a clinical population. Perhaps for this reason, participants had better hearing thresholds and less evidence of auditory dysfunction than patients in a typi cal clinic population. Many similar studies examining the role of suprathreshold factors or set of factors in y separated age ranges. For example, as cited several times in this dissertation, Humes and Floyd (2005) studied the role of working 21 or 22 years, and a group o terms of size, gender makeup, age, width of age range, and hearing status), enhances detecti on of group differences. With the two group design, however, it is difficult to determine which factors contribute contributing to the easily detected differences since the groups differ along so many dimensions. In contrast, in the present study data wer e examined for a single group with age treated as a continuous variable. All participants were within a relatively limited range of approximately 39 years, spanning late middle and old age. This strategy allowed for analysis of change in hearing sensitiv ity, auditory temporal resolution, and working memory within a period when deficits in all of these abilities are known to appear. Using a single group sample and treating age as a continuous variable has one immediate disadvantage. Unless hearing loss an d age are independent in the sample, the influences of these variables can be difficult to separate. Hearing loss and age are actually correlated in most populations, including an unselected volunteer group. As a result, variable interactions may be hard to detect or obscured completely. Using statistical controls for hearing loss and/or age,

PAGE 113

113 rather than grouping a sample categorically, can reveal changes or trends across a continuous range of ages or increments of hearing loss. This data analysis method also allows for examination of the way variable associations change across age ranges, for example, determining whether the association between speech perception and audibility is comparable for individuals in their 50s as it is for individuals in their 7 0s and 80s. Gaps in Noise (GIN) Test Findings Results of GIN testing were consistent with previous studies finding that individuals with serious hearing loss (lower AI scores) demonstrate significantly longer gap detection thresholds (higher GIN thresholds ) than individuals with little or no hearing loss (Arlinger & Dryselius, 1990; Buus & Florentine, 1985; Madden & Feth, 1992; Moore et al, 1989). Similar to findings of most auditory temporal resolution studies, in the present study declining temporal reso lution ability was associated with decreased hearing, but much of the variance in temporal resolution ability was unexplained by hearing sensitivity. Monaural audibility measures accounted for about 26% of the variance in temporal resolution ability in th e right ear and 41% in the left ear. The apparent independence of GIN threshold and age, with hearing loss controlled, was also consistent with previous research (e.g., He et al, 1999). Importantly, analysis of the GIN test data revealed that more than 70 % of the participant group (40 of 56) demonstrated temporal thresholds outside of the normative range in at least one ear. Of these 40 participants, 19 had worse ear GIN thresholds of 8 ms, the gap iteration just above the 7 ms norm (Musiek et al, 2005). Gap lengths evaluated with the GIN test are 2, 3, 4, 5, 6, 8, 10, 12, 15, and 20 ms. There is no 7 ms gap. The omission of the 7 ms gap increment leaves open the possibility that individuals with borderline normal temporal resolution might be classified as abnormal. For example, in this investigation participant #2 correctly identified 6 ms gaps two out of six times (33%) and 8 ms gaps five out of six times (83%) in his right ear.

PAGE 114

114 hold of 8 ms is outside of the normal range. However, his ability to identify correctly some gaps at 6 ms indicates that his actual gap threshold may be closer to 7 ms and, according to the normative criterion, within normal limits. It appears, therefore that the current version of the GIN needs to be revised to test gaps closer to the normative cutoff if it is to be used as a diagnostic tool. The high percentage of participants with GIN thresholds outside of the 7 ms cutoff suggests that the normative d ata published by Musiek et al may not be appropriate for adults above age 50 or for adults with SNHL. As of this writing, no GIN test normative data have been published for individuals in the age range recruited for this investigation or for any other sam ple of older adults. The youngest participant recruited, at 50 years, 1 month of age, was still four years older than the oldest participant in the normative group. Perhaps more importantly, no normative data have been published for individuals with conf irmed SNHL. As noted above, it appears that the elevated GIN thresholds in this sample are a result of hearing loss and not of age. However, no firm conclusion can be drawn from these data because, in comparison to the normative subject sample (Musiek et al, 2005), the sample in the present study had greater hearing loss and was older. As expected based on the recruitment criteria (individuals who believed they had a hearing problem), few participants presented with audiometrically normal hearing. In all six tone sensitivity at 20 dB HL or better at ms on the right (mean age 55.1), and 5.8 ms on the left (mean age 58.8). These data, although clearly limited, suggest that the average elevation in GIN threshold in the whole sample is primarily a function of hearing loss. However, mean gap thresholds for these individual ears

PAGE 115

115 ll elevated compared to the normative averages of 4.9 and 4.8 ms, suggesting the possibility of a small age effect. Further study is needed to separate the effects of age and hearing loss on GIN threshold. Larger samples of age matched groups with varyin g levels of hearing loss, or hearing loss matched groups differing in age, might provide insight into the role of each factor. Age and hearing loss appropriate normative criteria are needed for the GIN test. The number of participants who would be classi fied abnormal under the original norms is quite high over 70% of a group of a non clinical population. In fact, no participant had GIN threshold in either ear better than 5 ms, which is marginally above the average thresholds seen in the normative group. A second indication of a need for expanded normative data is the finding of an apparent age related left ear disadvantage on the GIN. This increase is most apparent in Table 4.12, where right and left ear differences in hearing sensitivity (RAI vs. LAI) and GIN threshold (RGIN vs. LGIN) are displayed for groups of participants in their 50s (Group 1), 60s (Group 2), and 70s and 80s (Group 3). Left ear GIN threshold showed a more rapid decline with age than did right ear threshold: Inter ear differences were 0.11 ms in Group 1, 0.64 in Group 2, and 1.55 in Group 3. While the inter ear difference in GIN threshold was significant only between Group 2 and Group 3, an increase in asymmetry is evident across the age range of the sample. This increase in asym metry is apparently unrelated to audibility: the inter ear difference in AI was stable and did not change significantly between groups. If this ear asymmetry is seen consistently in future GIN investigations, development of both clinical norms that are b oth age appropriate and ear appropriate would be indicated. Asymmetry favoring the right ear is common in hearing studies on older adults. Research confirms enhanced function for the right ear, i.e., a left ear disadvantage, in aging for selected

PAGE 116

116 measures of hearing including hearing sensitivity (Kannan and Lipscomb, 1974), otoacoustic emissions (Khalfa, Morlet, Michey, Morgon, & Collet, 1997), and dichotic speech perception (Bellis & Wilber, 2001; Jerger et al, 1994; Musiek, Wilson, & Pinheiro, 1979). On e explanation offered for auditory asymmetry is noise exposure. Examples of unbalanced noise exposures include rifle shooting, which, for right handed individuals, places the left ear closer to the barrel than the right ear, or driving a car with the wind ow down, where the left ear is closer to the window and thus more susceptible to wind noise. However, asymmetry in auditory abilities within and beyond the peripheral hearing system can not always be explained by noise exposure. Numerous studies have ide ntified a left ear disadvantage in dichotic listening that increases with age even when controlling for changes in hearing sensitivity (Bellis & Wilber, 2001; Clark & Knowles, 1973; Jerger & Jordan, 1992; Jerger et al, 1990; Jerger et al, 1994; Johnson, Co le, Bowers, Foiles, Nikaido, Patrick, & Wolliver, 1979). Auditory system asymmetry is likely related to physiological changes with age that affect the right and left auditory structures and pathways unequally. One hypothesis to explain auditory asymmetry for most individuals the centers for language and verbal processing are found, provides that ear with privileged access to speech perception resources (Jerger et al, 1994). More over, neuro imaging studies show greater atrophy in the right cerebral hemisphere (to which the left ear has superior access) compared to the left hemisphere (to which the right ear has access) (Goldstein & Shelly, 1981; Levy Agresti & Sperry, 1968). Det erioration in function of the corpus callosum in older adults may also affect auditory pathways by reducing inter hemispheric communication (Duffy, McAnulty, & Albert, 1996; Goldstein & Braun, 1974; Hellige, 1993; Jerger et al, 1994). This deterioration c an reduce the

PAGE 117

117 efficiency of left ear auditory pathways, which decussate at the level of the brainstem and then must pass through the pathways of the corpus callosum to reach the left hemisphere auditory processing centers. The increasing asymmetry seen o n the GIN test in the present investigation may be related to physiological changes that affect the right and left auditory pathways differently. Most of the previous studies using gap detection paradigms have focused on a single number, binaural measure of temporal resolution ability. It is likely that, in these studies, only the best ear is responding, making ear differences undetectable. Further research is needed to confirm and explain the apparent asymmetrical aging of gap detection ability. Working Memory Findings Performance on all working memory tasks declined with age, consistent with much of the previous literature on memory and aging (Carpenter et al, 1994; Hultsch et al, 1992; Salthouse, 1991; Wingfield et al, 1988). Interestingly, the rate o f change for each working memory task was different within the range of this age sample. For example, age sub group comparison showed significant declines in forward digit span ability only from the seventh decade on; that is, no performance difference be tween participants in their 50s and participants in their 60s. Digit ordering ability, by contrast, declined between the sixth and seventh decades, but not afterward. Some recent working memory studies have similarly shown different age related decline p atterns among working memory measures (Meguro, Fujii, Yamadori, Tsukiura, Suzuki, Okuda, & Osaka, 2000; Parente, de Taussik, Ferreira, and Kristensen, 2005). Inconsistencies in age effects may be explained by differences in the cognitive abilities accesse d for different working memory tasks. Declines in digit span ability late in the aging process are likely to be related to loss of short term memory capacity (Palladino & De Beni, 1999). Digit ordering, which requires more manipulation of information tha n do digit span tasks, may be more sensitive to declines in

PAGE 118

118 executive function, one of the earliest cognitive processes to show changes with aging (Fisk & Sharp, 2004; Meguro et al, 2000). The varying patterns of age related change are notable in that the y suggest that the method of combining the three measures into a single working memory score may have helped to balance these differing age effects across the sample. One concern in designing the present investigation was the possible collinearity among th e predictor variables of audibility, temporal resolution, and working memory. While an association between temporal resolution ability and hearing loss was anticipated (as described above), no association was expected between working memory and either aud ibility or temporal resolution. When controlling for age, there was no correlation between hearing loss and working memory change. Previous research has also found independence between hearing status and working memory when memory testing is conducted at an audible level (Gordon Salant & Fitzgibbons, 2001; Lyxell et al, 2004). All temporal resolution variables were independent of working memory variables when age was controlled. The lack of association among the predictors confirms that the GIN and dig it tasks accessed different abilities, and that testing their partial correlation with a separate independent variable should not be confounded by any spurious collinear ity, (e.g., correlation of both variables with overall intelligence). A review of the literature was conducted to identify reported relations among temporal resolution and working memory variables. Three articles, all from a single research group, with language impairment (Fernell, Norrelgen, Bozkurt, Hellberg, & Lwing, 2002; Norrelgen, Lacerda, & Forssberg, 2001, 2002). Like the present study design, each of these investigations tested the separate predictive ability of a temporal resolution tas k and a working memory task on

PAGE 119

119 speech perception. None of the three articles published by Norrelgen and colleagues reported any association among these predictor variables, supporting the finding of the present investigation that temporal resolution and w orking memory measures are independent. Speech Perception The association between audibility and speech perception in the present investigation was remarkably consistent with the results of previous studies. In the quiet test condition, audibility (as mea sured by the better ear AI) accounted for about 72% of the variance in HINT threshold. The strength of this correlation was consistent with previous speech perception research (Divenyi & Haupt, 1997; Glasberg & Moore, 1989; Humes et al, 1994; Plomp & Mimp en, 1979). In a slightly older age group (63 to 83), Humes and colleagues (1994) found that audibility accounted for between 70 and 75% of the variance on a series of speech recognition tasks. In the noise condition, audibility explained about 52% of th e variance in HINT threshold. Other speech perception studies also found that loss of hearing sensitivity explains about half of the variance in speech perception in some noisy environments (Crandell, 1991; Middelweerd et al, 1990; Smoorenburg, 1992). Pa rticipants in the present investigation required a signal to noise ratio (SNR) of 3.7 to 16.7 dB (mean = 7.2) for perception of the HINT sentences in noise. Previous research has indicated that listeners with sensorineural hearing loss, the type of hearin g loss demonstrated by the individuals participating in this study, require SNRs from about 4 to 12 dB for speech perception (Crandell & Smaldino, 2002; Killion, 1997; Moore, 1997). When hearing sensitivity was statistically controlled for the full partici pant sample, none of the temporal resolution or working memory predictor variables accounted for additional variance in HINT threshold (in quiet or in noise). The absence of an association was confirmed when the participant sample was separated into two g roups by audiometric hearing status, as well as when the sample was divided into age groups, with one exception. When the oldest participant group

PAGE 120

120 (age 70 to 89 years) was analyzed separately, binaural average GIN threshold was a significant partial predi ctor of HINT threshold in the quiet condition only. That is, temporal resolution ability explained an additional 12% of the variance in speech perception above the 62% explained by hearing sensitivity. The overall lack of association between HINT and GI N variables suggests that the GIN does not significantly contribute to prediction of HINT threshold except for a modest explanatory contribution for the oldest participants examined. Because previous research indicated an association between temporal res olution and speech perception in noise, a partial correlation between GIN threshold and HINTN was anticipated in the present study. However, other investigations also have found weak or no correlations between temporal processing and speech perception whe n hearing sensitivity was controlled (Dubno & Dirks, 1990; Festen & Plomp, 1983; Snell & Frisina, 2000; Strouse et al, 1998; van Rooij & Plomp, 1991). While it seems likely that auditory temporal resolution ability is a factor for certain populations when listening in difficult environments, the present investigation produced little ev idence of such an association As noted in the review of literature, previous attempts to find an association between working memory and difficult speech perception have prod uced widely varying results (Gordon Salant & Fitzgibbons, 1997; Humes & Floyd, 2005; Lunner, 2003; Pichora Fuller & Singh, 2006; Pichora Fuller et al, 1995; Tun et al, 1991; Vaughan et al, 2006). A correlation between working memory and HINT threshold was cautiously anticipated, but the conflicting findings in the cited studies made this association far from certain. As with temporal resolution ability, the present investigation produced no evidence of any role for working memory in speech perception.

PAGE 121

121 Ana lyses of the data set by age group revealed changing regression models from group to group. The age related changes suggest that the degree to which speech perception performance can be predicted by audibility (or any other predictor variable) is different for adults in their sixth, seventh, and eighth decade, and above. In fact, for the youngest participants (age 50 to 59 years), HINT thresholds in noise could not be predicted at all using audibility, temporal resolution, or working memory variables. Aud ibility increased in strength as a predictor of HINT scores as mean group age increased. For speech perception testing in quiet, the correlation between HINT threshold in quiet and audibility was .50 for participants in their fifties and .81 for participa nts in their sixties. The group of participants in their seventies and eighties showed a correlation of .86 between HINT threshold in quiet and a combination of audibility and temporal resolution threshold. For speech perception testing in noise, the co rrelation between HINT threshold and audibility was not significant for participants in their fifties (0.10) but was significant for participants in their sixties (0.47), and for participants in their seventies and eighties (0.90). A clear trend is eviden t in these data: the role of audibility in speech perception performance, both in quiet and in noise, increased with age. Notably, age was not significantly correlated with speech perception performance in quiet or in noise when hearing sensitivity was co ntrolled. This was a somewhat surprising finding given that many other speech perception investigations found an effect of age for listening in noisy or otherwise difficult environments. However, in some of the other studies examining the role of age in speech perception, comparisons were made between groups of listeners chosen as defined as over age 60 or 65 years. For example, Divenyi and Haupt (1997) compared a

PAGE 122

122 compared two groups with age ranges of 17 to 40 and 64 to 77 years. It is possible that age effects are only detectable when comparing such disparate age groups, and that age effects are minimized when data are compared within a group of older adults. Treating age as a continuous variable within a narrow age range, as in the present investigation, may obscure these effects. In all likelihood, age influences speech p erception performance, particularly in difficult listening situations. Hearing Handicap Not surprisingly, none of the predictor variables predicted self reported hearing handicap well. Both age and audibility were significantly, albeit weakly, correlated with HHIA/E total score and Social/Situational subscale score, but not correlated with Emotional subscale score. The associations among age and Social/Situational measures were quite weak, explaining only 11 to 24% of the variance in HHIA/E scores. Dimi nished hearing sensitivity (lower AI) was associated with increased hearing handicap (higher HHIA/E score), as would be expected. However, as with age, the relationship was weak, with audibility accounting for 12 to 20% of the variance in hearing handicap score. In fact, there was no difference in HHIA/E Emotional subscale or total score when comparing audiometrically normal participants with those having a hearing loss. The absence of an association suggests that the emotional impact of hearing loss on the individual is so variable that it can not even be distinguished between individuals with and without clinically significant hearing loss. When controlling for hearing sensitivity, the correlation between HHIA/E score and age disappeared, indicating tha t the association between objective and subjective measures of hearing loss was similar for younger and older participants. Notably, the correlations between age and HHIA/E score and between audibility and HHIA/E score were almost identical in strength. In

PAGE 123

123 other words, a clinical audiologist would appear to have just as much success in predicting a conducting an audiometric examination. Within this group of adults, there was no evidence that older participants reported a greater social or emotional impact of hearing loss than younger (late middle aged) participants. Previous studies have yielded varying results on the relationship of age to hearing self rep ort. Some researchers found that older adults report more disabling effects of hearing loss than younger individuals with similar audiograms (e.g. Lutman et al, 1987), whereas in other studies older adults tended to underrate their hearing difficulties in comparison to the ratings of hearing difficulties by younger adults (e.g. Gordon Salant & Fitzgibbons, 1994). Data from the present investigation suggest that the association between hearing loss and hearing handicap is stable above age 50. Overall, audi ometric hearing sensitivity only weakly predicted the situational and social impact of hearing loss, and did not predict the emotional consequence of the hearing loss. Early validation studies of the HHIE confirmed that hearing sensitivity only explained about half of the variance in HHIE total score (Weinstein & Ventry, 1983). Low correlations between objective and subjective measures of hearing loss, as found in the present investigation, are common in studies of hearing (Brainerd & Frankel, 1985; Matth ews et al, 1990; McKenna, 1993; Newman et al, 1997). The well established disconnect between audiometric hearing status and hearing handicap by self report may be related to the weak association between hearing sensitivity and speech perception in the pres ence of noise or other distortion. Difficulty when listening in noisy and reverberant environments is the most common complaint from individuals reporting a hearing

PAGE 124

124 loss (CHABA, 1988; Kim et al, 2005; Koehnke & Besing, 2001). Speech perception problems a re particularly common in older adults. Older adults demonstrate disproportionately more difficulty listening in noisy environments when compared to younger adults with similar degrees and configurations of hearing loss (Pichora Fuller & Souza, 2003; Souz a et al, 2000; Tun et al, 2002). In the present investigation, HINT threshold in noise was more closely related to HHIA/E score than were HINT threshold in quiet or audibility. That is, speech perception in noise was a better predictor of hearing handi cap than hearing sensitivity or speech perception in quiet. This important finding supports the contention that speech perception in noise is a primary factor in de speech perception in noise measures, such as the HINT, in a clinical test battery, particularly for adults who may be candidates for hearing aids. Hearing handicap score was weakly correlated with measures of auditory function. Temporal resolution th reshold accounted for about 14% of the variance in HHIA/E Emotional subscale score and about 17% of the variance in total score. Audibility accounted for about 20% of the variance in Social/Situational subscale score. Both GIN threshold and hearing sensi tivity were less closely associated with HHIA/E scores than the HINT threshold in noise. The finding that GIN threshold was a better predictor than audibility for HHIA/E Emotional subscale and total scores was unexpected. One possible explanation is that HHIA/E perceive speech in background noise, which has been shown to be related to temporal resolution (Glasberg & Moore, 1989; Glasberg et al, 1987; Irwin & M cAuley, 1987; Snell et al, 2002; Tyler et al, 1982). Several of the questions on the HHIA/E ask either directly or indirectly about

PAGE 125

125 Hearing loss accompanied by loss of temporal resolution ability is particularly detrimental to speech perception in the presence of fluctuating noise, such as many voices speaking simultaneousl y (Gustafsson & Arlinger, 1994; Hygge, Rnnberg, Larsby, & Arlinger, 1992). The Emotional subscale of the HHIA/E was significantly correlated with GIN threshold and with HINT threshold in noise but not with audibility or HINT threshold in quiet. Emotion al subscale items are framed to reflect the way the individual feels about his or her hearing problems. Many of these questions ask whether certain situations make the answerer feel em cause you to feel left when listening in noisy situations or when many people are speaking at once. Therefore, the emotional impact of hearing loss may be most salient for listening in noise. Limitations Lack of variance in HINT thresholds in the noise condition in the present investigation may have obscured some variable relations, specifically the influence of temporal resolution on speech perception ability. The standard deviation for thresholds in the noise condition was only 2.6 dB, compared to 9.3 dB in the quiet condition. It is not surprising that HINT threshold was more invariant when in noise than in quiet. In the quiet condition, HINT threshold was closely related to hearing sensitivity, and hearing sensitivity was highly variable among participants in the present investigation. However, the HINT presentation level in noise was well above hearing threshold. In order to be audible to a participant i n the present investigation, with or without hearing loss, the speech stimulus needed to be presented at the intensity level of the noise (65 dB SPL) or higher. Average HINT threshold in noise in this investigation was 72.2 dB SPL. At this

PAGE 126

126 intensity leve l, without the noise competition, the sentences were clearly audible to almost all of the participants. Thus, elevation of HINT threshold was presumably due to some interference effect from the noise, and not a function of the audibility of the sentences. However, because thresholds in noise f ell in a very narrow range, an interference effect may have been obscured. In addition, u sing the steady state HINT competition noise rather than a modulated masker may have contributed to threshold invariance. Seve ral speech perception studies have noted an increase in temporal resolution effects when the intensity of the competing noise is modulated, either using some fixed modulation rate or by employing a multi talker noise that contains random fluctuations in in tensity (Bacon et al, 1998; Dubno, Horwitz, & Ahlstrom, 2002; Eisenberg, Dirks, & Bell, 1995; Festen & Plomp, 1990; Glasberg et al, 1987; Gustafsson & Arlinger, 1994; Howard Jones & Rosen, 1993; Jin & Nelson, 2006; Miller & Licklider, 1950; Nelson & Freyma n, 1987). Individuals who have normal hearing and normal temporal masking) in background noise. These periods of low intensity are common when listening t o several people who are speaking at once. Individuals with hearing loss and impaired temporal resolution lose the ability to take advantage of these modulations and do not show speech perception improvement when modulated and steady noises are compared ( Bronkhorst & Plomp, 1990; Jin & Nelson, 2006; Middelweerd et al, 1990). It is possible that temporal resolution differences within this sample of adults with hearing loss were obscured by the use of the standard HINT competition, which remains at a consta nt intensity throughout testing. A modulated competition noise, such as a multi talker babble or a modified version of the HINT competition noise, may have increased the variance in HINT threshold and made temporal resolution effects on speech perception more apparent.

PAGE 127

127 Little is known about the association between speech perception in noise and working memory capacity. As detailed in the introduction, studies examining the role of working memory in speech perception, using widely varying measurement metho ds, have not surprisingly yielded widely varying results. Following the advice of Lunner (2003), sentence length speech stimuli were used in the present investigation in an attempt to identify working memory effects that are not evident for simpler, word or syllable level stimuli (e.g., Humes et al, 1994; Humes and Floyd, 2005). However, no working memory effects were apparent in HINT performance in quiet or in noise. It is possible that the HINT sentences did not tax memory ability sufficiently because they are, by design, simple in vocabulary and structure and fairly high in context (i.e., (2003) and Vaughan and colleagues (2006) support this possibility. These r esearchers reported working memory effects for speech perception using low context and nonsense sentences. Furthermore, Tun and colleagues (1991) noted apparent working memory effects for recall of portions of lengthy expository passages. Use of more com plex or lower context speech stimuli may have revealed working memory effects not seen using the simpler HINT sentences. A major potential sample limitation in the present investigation was the gender imbalance seen among the participant group (17 male, 39 female). Because participants were recruited as young as the sixth decade, it is unlikely that the gender imbalance reflects a survival effect. Moreover, the direct recruitment method was not overtly gender biased. Participants were recruited from th e Alachua county area principally using flyers posted on campus and at recruitment talks given at locations around Gainesville. The best explanation for the high proportion of female participants is that women enrolled early during the data collection per iod were more likely to ask their friends and neighbors to participate. Approximately one half of the

PAGE 128

128 participant pool was obtained by word of mouth recruitment, and a large majority of both the recruiters and recruits were female. At least three female participants recruited from the membership of their social groups, skewing this secondary recruitment population. Regardless of the reason, no gender differences were evident on any measure, suggesting that the gender imbalance did not distort the data se t. It should be noted that, while the sample of fifty six adults was adequate for statistical analysis of the group as a whole, sample sizes for post hoc group tests were as small as seventeen participants. The separate analyses on these groups certainly are limited by their small group size and no strong evidence should be inferred from these data alone. In addition, reducing the participant sample in this way exacerbated the gender imbalance problem noted above. In the examination of age groups, the nu mber of male participants was reduced to five or six per group. Future investigations examining group differences will attempt to recruit larger participant samples with more even gender distribution. Recruitment of participants was limited to individua ls who could transport themselves to the University of Florida campus for one appointment for completion of all study measures. Caution should be exercised in extrapolating any findings to populations who might not have been able to satisfy these requirem ents, such as individuals with dementia or other cognitive impairment, or whose health prevented them from traveling to the test site. Participants were also required to speak English as a first language in order to eliminate possible linguistic confounds in HINT testing or speech audiometry. The findings of the present investigation cannot, therefore, be generalized to individuals who speak a language other than English or for whom English is a second language.

PAGE 129

129 Finally, while the GIN test appears to have good face validity as a gap detection measure, it is a new tool with limited published data on reliability and accuracy. As noted above, the current version of the test appears to lack gap stimuli of appropriate length (specifically 7 ms) for precise dia gnosis of impaired temporal resolution according to published norms. In addition, GIN normative data are lacking for populations other than young adults with normal hearing. Future versions of the GIN test, and future clinical investigations, may address these issues. Future Directions This investigation provides a starting place for future research into the factors influencing speech perception in older adults. First, the associations between both of the predictor variables (temporal resolution and work ing memory) should be tested with updated speech perception measures. Specifically, the steady state HINT competition noise should be replaced by modulated noise in an attempt to identify temporal resolution effects. The present investigation focused on the role of a single distortion (noise) of the speech signal. Follow up studies might use reverberation effects to further distort the stimulus signal and/or the noise. Previous researchers have suggested that a combination of noise and reverberation is not only more reflective of real world environments, but also may affect speech perception to a greater degree than either of the distortions alone (Danhauer and Johnson, 1991; Gelfand and Silman, 1979; Helfer, 1992, 1994; Helfer and Huntley, 1991; Johnson 2000). Second, working memory effects should be examined using more complex speech stimuli, such as low context sentences or extended passages from which the participant must recall information heard early in the reading. Low context or nonsense sentenc es are likely to be more difficult to recall because linguistic knowledge can not be used to compensate for sounds or words that are not heard. Long speech passages may be a more effective test of working memory than are the shorter HINT sentences as they would require the listener to retain

PAGE 130

130 information while continuing to listen and process auditory input. There are few published studies examining working memory effects on speech perception and, within those few, variable measures are inconsistent (i.e., n back recall, reading and digit span, visual sequence learning). Substantial future research is needed to reconcile conflicting findings among published studies and to more accurately measure the role that working memory plays in everyday listening. Sum mary As in many prior investigations loss of hearing sensitivity was the principal factor explaining speech perception performance in quiet and in noise. However, substantial variance in speech perception was unexplained by loss of signal audibility, par ticularly for subjects below 70 years of age. No strong evidence was found to support the hypothesis that temporal resolution and/or working memory accounted for any of this additional variance when controlling for loss of hearing sensitivity. This findi ng of no additional variance explained by these central auditory and cognitive measures runs contrary to much of the prevailing literature, particularly in the case of temporal resolution research. It is likely that the hypothesized variable relations, wh ich have been demonstrated robustly and repeatedly in prior studies, were obscured by invariance in HINT threshold in the noise condition, as well as by the simplicity and brevity of the HINT sentences. While the sound field HINT protocol employed in thi s investigation was intended to represent listening in a diffuse noisy environment, the testing method clearly falls short of replicating real life listening situations. For a listener with hearing loss attempting to carry on conversation in a difficult e nvironment such as a busy restaurant, speech perception and comprehension are likely to be affected by several factors that the HINT can not simulate. Two such factors are attention span and attention switching, which may be influenced by multiple modalit y distractors (i.e., both visual and auditory stimuli). Another factor is the presence of

PAGE 131

131 multiple talkers, as in a group conversational setting, which creates fluctuating and varying noise competition that masks not only acoustically by obscuring speech signals, but also informationally in that the noise has verbal content. Further investigation is needed to determine whether more complex speech perception measures, such as tasks using more difficult stimuli, modulated noise maskers, or reverberant test environments, may reveal variable associations not seen here. In addition, speech perception tasks that purport to simulate real world listening environments should better account for factors described above such as multi modal distraction and increased l oad on speech processing in the presence of multiple verbal signals. Clearly factors beyond peripheral hearing sensitivity contribute to the commonly seen speech perception difficulties of older adults. The failure to find all of the hypothesized variable relations in the present investigation should by no means be considered a definitive refutation of the associations hypothesized and demonstrated in prior research. In all likelihood, limitations both in the sampling methodology and in the speech percept ion measure made the anticipated and previously demonstrated correlations more difficult to detect. The current findings are instructive in that they demonstrate the difficult of simulating everyday listening environments in a laboratory setting, or even many clinical settings. Indeed, a primary goal of the present investigation was to attempt to replicate significant laboratory findings of the influence of temporal resolution and working memory on speech perception using simpler clinical measures. That is, this investigation attempted to bridge the gap between laboratory results, frequenly obtained on selected samples and with complex equipment and test methods, and clinical testing, which must necessarily be brief, easy to implement (with minimal cost, equipment installation, or training to clinicians), and easy to interpret. The difficulty in

PAGE 132

132 accomplishing that goal is illustrated well by the finding that the GIN and digit span tests could not successfully translate laboratory findings to a clinical p rotocol. An important, although hardly surprising finding of this investigation was that s elf reported hearing handicap was better predicted by HINT threshold in noise than by loss of hearing sensitivity. This suggests that listening in difficult environm ents was a primary element reinforcing the need for speech perception in noise testing as a standard part of the audiologic clinical battery. Routine use of a test such as the HINT may provide the audiologist with valuable information to aid in providing appropriate rehabilitation servic es for adults with hearing loss. The finding of an apparent asymmetrical aging effect on GIN threshold is an important one that requires future study. Central audi tory changes that affect right and left ear auditory input processing differently are likely to contribute substantially to dichotic listening and auditory abilities that rely upon binaural input such as localization and binaural squelch. Difficulties af fecting all auditory input right and left ear are a consequence of the overall decline in central auditory and cognitive function seen in aging. These difficulties may be seriously compounded if central processing capabilities age differently between ears and if the relative efficiency with which auditory signals are processed is in flux with age. That is, auditory aging exists multidimensionally: overall worsening of auditory function with age affecting both ears is exacerbated by decreasing symmetry i n auditory processing. As a result, additional strain is likely to be placed upon central and cognitive processing to reconcile incoming signals that arrive and are processed at different rates, and which become more and more out of synchrony as with incr easing age. In this respect, disordered central auditory processing in older adults may differ from the pattern of disorder in children who do not experience age related loss of

PAGE 133

133 interhemispheric communication or deterioration of neural pathways. Asymmetr ic loss of central auditory function is not a new concept in hearing research. The present findings emphasize the need to account for this unique feature of older adult auditory processing when attempting to explain the nature of hearing loss with age.

PAGE 134

134 A PPENDIX A INFORMED CONSENT

PAGE 135

135

PAGE 136

136

PAGE 137

137

PAGE 138

138 APPENDIX B GAPS IN NOISE TEST

PAGE 139

139

PAGE 140

140

PAGE 141

141

PAGE 142

142

PAGE 143

143

PAGE 144

144 REFERENCE LIST Alain, C. & Woods, D. L. (1999). Age related changes in processing auditory stimuli during visual attention: Evidence for defi cits in inhibitory control and sensory memory. Psychology and Aging, 14 507 519. American Medical Association (AMA), Council on Physical Medicine (1947). Tentative standard procedure for evaluating the percentage loss of hearing in medicolegal cases. Journal of the American Medical Association, 133 396 397. American National Standards Institute (ANSI). (1986). ANSI S3.5 1969 R 1986: American national standard methods for the calculation of the articulation index. New York: ANSI. American Speech Lan guage Hearing Association (ASHA) (1978). Guidelines for manual pure tone audiometry. ASHA 20 (Suppl 2) 25 30. Appolonio, I., Carabellese, C., Magni, E., Frattola, L., & Trabucchi, M. (1995). Sensory impairments and mortality in elderly community popul ation: A six year follow up study. Age and Ageing, 24 30 36. Arlinger, S. & Dryselius, H. (1990). Speech recognition in noise, temporal and spectral resolution in normal and impaired hearing. Acta Otolaryngologica (Suppl), 469 30 37. Baddeley, A. D (1992). Working memory. Science, 255 556 559. Baddeley, A. D. & Hitch, G. J. (1974). Working Memory. In G.A. Bower (Ed.), Recent advances in learning and motivation (Vol. 8) New York: Academic Press, 47 90. Bellis, T. J. & Wilber, L. A. (2001). Eff ects of aging and gender on interhemispheric function. Journal of Speech, Language, and Hearing Research, 44 246 263. Bergman, M. (1980). Ageing and the perception of speech Baltimore: Baltimore University Press. Brainerd, S. H. & Frankel, B. G. (1 985). The relationship between audiometric and self report measures of hearing handicap. Ear and Hearing, 6 89 92. Brebion, G. (2001). Language processing, slowing, and speed/accuracy trade off in the elderly. Experimental Aging Research, 27 137 150 Bronkhorst, A. W. & Plomp, R. (1990). A clinical test for the assessment of binaural speech perception in noise. Audiology, 29(5) 275 285. Buus, S. & Florentine, M. (1985). Gap detection in normal and impaired listeners: The effect of level and fr equency. In A. Michelson (Ed.), Time Resolution in Auditory Systems (pp 159 179) London: Springer Verlag.

PAGE 145

145 Cameron, S., Dillon, H., & Newall, P. (2006). Development and evaluation of the listening in spatialized noise test. Ear and Hearing, 27(1) 30 42. Carhart, R. & Jerger, J. (1959). Preferred method for clinical determination of pure tone thresholds. Journal of Speech and Hearing Disorders, 16 340 345. Carpenter, P. A., Miyake, A., & Just, M. A. (1994). Working memory constraints in comprehens ion: Evidence from individual differences, aphasia, and aging. In M.A. Gernsbacher (Ed.), The Handbook of Psycholinguistics San Diego, CA: Academic Press. Chermak, G. D. & Lee, J. (2005). Comparison of children's performance on four tests of temporal resolution. Journal of the American Academy of Audiology, 16 554 563. Clark, L. & Knowles, J. (1973). Age differences in dichotic listening performance. Journal of Gerontology, 28 173 178. Cohen, G. (1987). Speech comprehension in the elderly: The effects of cognitive changes. British Journal of Audiology, 21 221 226. Committee on Hearing, Bioacoustics and Biomechanics (CHABA). (1988). Speech understanding and aging. Journal of the Acoustical Society of America, 83 859 920. Cooper, J. A., Sage r, H. J., Jordan, N., Harvey, N. S., & Sullivan, E. V. (1991). Cognitive Brain, 114 2095 2122. Crandell, C. (1991). Individual differences in speech recogniti on ability: Implications for hearing aid selection. Ear and Hearing 12 (Suppl 6) 100s 107s. Crandell, C. & Needleman, A. (1999). Modeling hearing loss via masking: Implications for hearing aid selection. The Hearing Journal, 52(11) 58 62. Crandell C. & Smaldino, J. (2002). Room acoustics and auditory rehabilitation technology. In J. Katz (Ed.), Handbook of Clinical Audiology, Fifth Edition (pp. 607 630) Baltimore: Lippincott Williams & Wilkins. Danhauer, J. L. & Johnson, C. E. (1991). Perce ptual features for normal listeners' phoneme recognition in a reverberant lecture hall. Journal of the American Academy of Audiology, 2 91 98. DeDe, G., Caplan, D., Kemtes, K. & Waters, G. (2004). The relationship between age, verbal working memory, and language comprehension. Psychology and Aging, 19 601 616. Divenyi, P. L. & Danner, W. F. (1977). Discrimination of time intervals marked by brief acoustic pulses of various intensities and spectra. Perception and Psychophysics, 21(2) 125 142.

PAGE 146

146 Diveny i, P. L. & Haupt, K. M. (1997). Audiological correlates of speech understanding deficits in elderly listeners with mild to moderate hearing loss: Age and lateral asymmetry effects. Ear and Hearing, 18 42 61. Divenyi, P. L. & Simon, H. J. (1999). Hear ing in aging: Issues old and new. Current Opinions in Otolaryngology, 7 282 289. Dixon Ward, W. (1983). The American Medical Association / American Academy of Otolaryngology formula for determination of hearing handicap. Audiology, 22 313 324. Dobi e, R. A. & Megerson, S. J.D. Royster, D.P. Driscoll, & M. Layne, M. (Eds.), The Noise Manual, 5th Ed (pp. 698 710) American Industrial Hygiene Association: Fairfax, VA. Dorman, M., Mart on, K., & Hannley, M. (1985). Phonetic identification by elderly normal and hearing impaired listeners. Journal of the Acoustical Society of America, 77 664 670. Dubno, J. R. & Dirks, D. D. (1989). Auditory filter characteristics and consonant recogni tion for hearing impaired listeners. Journal of the Acoustical Society of America, 85(4) 1666 1675. Dubno, J. R. & Dirks, D. D. (1990). Associations among frequency and temporal resolution and consonant recognition for hearing impaired listeners. Acta Otolaryngologica (Suppl), 469 23 29. Dubno, J. R., Horwitz, A. R., & Ahlstrom, J. B. (2002). Benefit of modulated maskers for speech recognition by younger and older adults with normal hearing. Journal of the Acoustical Society of America, 111 2897 2907. Dubno, J. R., Horwitz, A. R., & Ahlstrom, J. B. (2005). Word recognition in noise at higher than normal levels: decreases in scores and increases in masking. Journal of the Acoustical Society of America, 118 914 922. Duffy, F. H., McAnulty, G. B ., & Albert, M. S. (1996). Effects of age upon interhemispheric EEG coherence in normal adults. Neurobiology of Aging, 17(4) 587 599. Duquesnoy, A. J. & Plomp, R. (1980). Effect of reverberation and noise on the intelligibility of sentences in cases o f presbyacusis. Journal of the Acoustical Society of America, 68(2) 537 544. Eddins, D. A., Hall, J. W., & Grose, J. H. (1992). The detection of temporal gaps as a function of frequency region and absolute noise bandwidth. Journal of the Acoustical So ciety of America, 91(2) 1069 1077. Eisenberg, L. S., Dirks, D. D., & Bell, T. S. (1995). Speech recognition in amplitude modulated noise of listeners with normal and listeners with impaired hearing. Journal of Speech Language and Hearing Research, 38 222 233.

PAGE 147

147 Emanuel, D. (2002). The auditory processing battery: Survey of common practices. Journal of the American Academy of Audiology, 13 93 117. Fabry, D. & Van Tasell, D. (1986). Masked and filtered simulation of hearing loss: Effects on consona nt recognition. Journal of Speech and Hearing Research, 29 170 178. Fernell, E., Norrelgen, F., Bozkurt, I., Hellberg, G., & Lwing, K. (2002). Developmental profiles and auditory perception in 25 children attending special preschools for language impa ired children. Acta Paediatrica, 91(10) 1108 1115. Festen, J. M. & Plomp, R. (1983). Relations between auditory functions in impaired hearing. Journal of the Acoustical Society of America, 73 652 662. Festen, J. M. & Plomp, R. (1986). Speech recept ion threshold in noise with one and two hearing aids. Journal of the Acoustical Society of America, 79 465 471. Festen, J. M. & Plomp, R. (1990). Effects of fluctuating noise and interfering speech on the speech reception threshold for impaired and nor mal hearing. Journal of the Acoustical Society of America, 88(4) 1725 1736. Fisk, J. E. & Sharp, C. A. (2004). Age related impairment in executive functioning: updating, inhibition, shifting, and access. Journal of Clinical and Experimental Neuropsych ology, 26(7) 874 890. Fitzgibbons, P. J. (1983). Temporal gap detection in noise as a function of frequency, bandwidth, and level. Journal of the Acoustical Society of America, 74(1) 67 72. Fitzgibbons, P. J. & Gordon Salant, S. (1987). Temporal gap resolution in listeners with high frequency sensorineural hearing loss. Journal of the Acoustical Society of America, 81(1) 133 137. Fitzgibbons, P. J. & Gordon Salant, S. (1994). Age effects on measures of auditory duration discrimination. Journal o f Speech and Hearing Research, 27 662 670. Fitzgibbons, P. J. & Gordon Salant, S. (1995). Duration discrimination with simple and complex stimuli: Effects of age and hearing sensitivity. Journal of the Acoustical Society of America, 98 3140 3145. Fitz gibbons, P. J. & Gordon Salant, S. (1996). Auditory temporal processing in elderly listeners. Journal of the American Academy of Audiology, 7 183 189. Fletcher, H. & Galt, R. H. (1950). The perception of speech and its relation to telephony. Journal of the Acoustical Society of America, 22 89 151.

PAGE 148

148 Folstein, M., Folstein, S., & McHugh, P. (1975). Mini Mental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12 189 198. Foos, P. & Wright, L. (1992). Adult age differences in the storage of information in working memory. Experimental Aging Research, 18(1 2) 51 57. Formby, C. & Forrest, T. (1991). Detection of silent temporal gaps in sinusoidal markers. Journal of the Acoust ical Society of America, 89(2) 830 837. Formby, C., Gerber, M. J., Sherlock, L. P., and Magder, L. S. (1998). Evidence for an across frequency, between channel process in asymptotic monaural temporal gap detection. Journal of the Acoustical Society of America, 103(6) 3554 3560. French, N. R. & Steinberg, J. C. (1947). Factors governing the intelligibility of speech sounds. Journal of the Acoustical Society of America, 19 90 119. Frisina, D. R. & Frisina R. D. (1997). Speech perception and presbya cusis: Relations to possible neural mechanisms. Hearing Research, 106 95 104. Fry, A. F. & Hale, S. (2000). Relationships among processing speed, working memory, and fluid intelligence in children. Biological Psychology, 54(1 3) 1 34. Gatehouse, S. (1994). Components and determinants of hearing aid benefit. Ear and Hearing, 15 30 49. Gatehouse, S. (1998). Speech tests as measures of outcome. Scandinavian Audiology (Suppl), 49 54 60. Gelfand, S. A. & Silman, S. (1979). Effects of small room reverberation upon the recognition of some consonant features. Journal of the Acoustical Society of America, 66 22 29. Gennis, V., Garry, P., Haaland, K., Yeo, R., & Goodwin, J. S. (1991). Hearing and cognitive status in the elderly. Archives of Inter nal Medicine, 151 2259 2264. George, E. L. J., Zekveld, A. A., Kramer, S. E., Goverts, S. T., Festen, J. M., and Houtgast, T. (2007). Auditory and nonauditory factors affecting speech reception in noise by older listeners. Journal of the Acoustical Soc iety of America, 121(4) 2362 2375. Gilinsky, A. S. & Judd, B. B. (1994). Working memory and bias in reasoning across the adult life span. Psychology and Aging, 9 356 371. Glasberg, B. R. & Moore, B. C. J. (1989). Psychoacoustic abilities of subjects with unilateral and bilateral cochlear impairments and their relationship to the ability to understand speech. Scandinavian Audiology (Suppl), 32 1 25.

PAGE 149

149 Glasberg, B. R., Moore, B. C. J., & Bacon, S. P. (1987). Gap detection and masking in hearing impai red and normal hearing subjects. Journal of the Acoustical Society of America, 81 1546 1556. Goldstein, S. & Braun, L. (1974). Reversal of expected transfer as a function of increased age. Perceptual and Motor Skills, 38 1139 1145. Goldstein, S. & S helly, C. (1981). Does the right hemisphere age more rapidly than the left? Journal of Clinical Neuropsychology, 3 67 78. Gordon Salant, S. (2005). Hearing loss and aging: New research findings and clinical implications. Journal of Rehabilitation an d Research Development 42 (Suppl 4) 9 24. Gordon Salant, S. & Fitzgibbons, P. J. (1993). Temporal factors and speech recognition performance in young and elderly listeners. Journal of Speech and Hearing Research, 36(6) 1276 1285. Gordon Salant, S. & Fitzgibbons, P. J. (1995). Comparing recognition of distorted speech using an equivalent signal to noise ratio index. Journal of Speech and Hearing Research, 38(3) 706 713. Gordon Salant, S. & Fitzgibbons, P. J. (1997). Selected cognitive factors and speech recognition performance among young and elderly listeners. Journal of Speech, Language, and Hearing Research, 40(2) 423 431. Gordon Salant, S. & Fitzgibbons, P. J. (1999). Profile of auditory temporal processing in older listeners. Journal of S peech, Language, and Hearing Research, 42 300 311. Gordon Salant, S. & Fitzgibbons, P. J. (2001). Sources of age related recognition difficulty for time compressed speech. Journal of Speech, Language, and Hearing Research, 44 709 719. Gordon Salant, S., Lantz, J., & Fitzgibbons, P. J. (1994). Age effects on measures of hearing disability. Ear and Hearing, 15(3) 262 265. Greenberg, S. (1996). Auditory processing of speech. In N.J. Lass (Ed.), Principles of Experimental Phonetics (pp. 362 407). S t. Louis: Mosby Year Book. Grose, J. H., Eddins, D. A., Hall, J. W. (1989). Gap detection as a function of stimulus bandwidth with fixed high frequency cutoff in normal hearing and hearing impaired listeners. Journal of the Acoustical Society of Americ a, 86(5) 1747 1755. Grose, J. H. & Hall, J. W. (1992). Comodulation masking release for speech stimuli. Journal of the Acoustical Society of America, 91(2) 1042 1050.

PAGE 150

150 Grose, J. H., Hall, J. W., & Buss, E. (2001). Gap duration discrimination in liste ners with cochlear hearing loss: Effects of gap and marker duration, frequency separation, and mode of presentation. Journal of the Association for Research in Otolaryngology, 2(4) 388 398. Grose, J. H., Hall, J. W., & Buss, E. (2006). Temporal proces sing deficits in the pre senescent auditory system. Journal of the Acoustical Society of America, 119(4) 2305 2315. Grose, J. H., Hall, J. W., Buss, E., & Hatch, D. (2001). Gap detection for similar and dissimilar gap markers. Journal of the Acoustica l Society of America, 109(4) 1587 1595. Gustafsson, H. A. & Arlinger, S. D. (1994). Masking of speech by amplitude modulated noise. Journal of the Acoustical Society of America, 95, 518 529. Haggard, M. P. & Hall, J. W. (1982). Forms of binaural summa tion and the implications of individual variability for binaural hearing aids. Scandanavian Audiology, 15 47 63. Hall, J. W. (2000). Handbook of Otoacoustic Emissions San Diego: Singular Publishing Group. Halling, D. & Humes, L. (2000). Factors af fecting the recognition of reverberant speech by elderly listeners. Journal of Speech, Language, and Hearing Research, 43(2) 414 431. Hargus, S. E. & Gordon Salant, S. (1995). Accuracy of speech intelligibility index predictions for noise masked young listeners with normal hearing and for elderly listeners with hearing impairment. Journal of Speech and Hearing Research, 38 234 243. Harris, R. W. & Reitz, M. L. (1985). Effects of room reverberation and noise on speech discrimination by the elderly. Audiology, 24(5) 319 324. Haubert, N., & Pichora Fuller, M. K. (1999). The perception of spoken language by elderly listeners: Contributions of auditory temporal processes. Canadian Acoustics, 27 96 97. Hawkins, D., Prosek, R., Walden, B., & Montgomery A. (1987). Binaural loudness summation in the hearing impaired. Journal of Speech and Hearing Research, 30(1) 37 43. Hawkins, D. & Yacullo, W. (1984). Signal to noise advantage of binaural hearing aids and directional microphones under different lev els of reverberation. Journal of Speech and Hearing Disorders, 49 278 286. He, N., Horwitz, A., Dubno, J., & Mills, J. (1999). Psychometric functions for gap detection in noise measured from young and aged subjects. Journal of the Acoustical Society o f America, 106(2) 966 978. Heinrich, A. & Schneider, B. (2006). Age related changes in within and between channel gap detection using sinusoidal stimuli. Journal of the Acoustical Society of America, 119(4) 2316 2326.

PAGE 151

151 Helfer, K. S. (1992). Aging an d the binaural advantage in reverberation and noise. Journal of Speech and Hearing Research, 35(6) 1394 1401. Helfer, K. S. (1994). Binaural cues and consonant perception in reverberation and noise. Journal of Speech and Hearing Research, 37(2) 429 4 38. Helfer, K. S. & Huntley, R. A. (1991). Aging and consonant errors in reverberation and noise. Journal of the Acoustical Society of America, 90(4) 1786 1796. Helfer, K. S. & Wilber, L. A. (1990). Hearing loss, aging, and speech perception in rever beration and noise. Journal of Speech and Hearing Research, 33(1) 149 155. Hellige, J. (1993). Hemispheric asymmetry Cambridge, MA: Harvard University Press Hirsh, I. J. (1950). Relation between localization and intelligibility. Journal of the Aco ustical Society of America, 22 196 200. Hornsby, B. & Ricketts, T. (2007). Directional benefit in the presence of speech and speechlike maskers. Journal of the American Academy of Audiology, 18 5 16. Howard Jones, P. A. & Rosen, S. (1993). The percep tion of speech in fluctuating noise. Acustica, 78 258 272. Hulme, C., Roodenrys, S., Brown, G., & Mercer, R. (1995). The role of long term memory mechanisms in memory span. British Journal of Psychology, 86 527 536. Hultsch, D. F., Hertzog, C., Smal l, B. J., McDonald Miszczak, L., & Dixon, R. A. (1992). Short term longitudinal change in cognitive performance in later life. Psychology and Aging, 7 571 584. Humes, L. E. (1996). Speech understanding in the elderly. Journal of the American Academy of Audiology, 7 161 167. Humes, L. E., Dirks, D., Bell, T., & Kincaid, G. (1987). Application of the articulation index and the speech transmission index to the recognition of speech by normal hearing and hearing impaired listeners. Journal of Speech a nd Hearing Research, 29 447 62. Humes, L. E., Espinoza Varas, B., & Watson, C. (1988). Modeling sensorineural hearing loss. Journal of the Acoustical Society of America, 83 188 202. Humes, L. E. & Floyd, S. (2005). Measures of working memory, sequen ce learning, and speech recognition in the elderly. Journal of Speech, Language, and Hearing Research, 48 224 235. Humes, L. E. & Jesteadt, W. (1991). Models of the effects of threshold on loudness growth and summation. Journal of the Acoustical Socie ty of America, 90 1933 1943.

PAGE 152

152 Humes, L. E. & Roberts, L. (1990). Speech recognition difficulties of the hearing impaired elderly: The contributions of audibility Journal of Speech and Hearing Research, 33, 726 735. Humes, L. E., Watson, B. U., Christ ensen, L. A., Cokely, C. G., Halling, D. C., & Lee, L. (1994). Factors associated with individual differences in clinical measures of speech recognition among the elderly. Journal of Speech and Hearing Research, 37 465 474. Hygge, S., Rnnberg, J., Lar sby, B., & Arlinger, S (1992). Normal hearing and hearing impaired subjects' ability to just follow conversation in competing speech, reversed speech, and noise backgrounds. Journal of Speech and Hearing Research, 35, 208 215. Irwin, R. J. & McAuley, S. F (1987). Relations among temporal acuity, hearing loss, and the perception of speech distorted by noise and reverberation. Journal of the Acoustical Society of America, 81(5) 1557 1565. Jerger, J. (1973). Audiological findings in aging. Advances in Otorhinolaryngology, 20 115 24. Jerger, J. (1992). Can age related decline in speech understanding be explained by peripheral hearing loss? Journal of the American Academy of Audiology, 3 33 38. Jerger, J., Chmiel, R., Allen, J., & Wilson, A. (1994). Effects of age and gender on dichotic sentence identification. Ear and Hearing, 15 274 286. Jerger, J. & Hayes, D. (1977). Diagnostic speech audiometry. Archives of Otolaryngology, 103 216 22. Jerger, J., Jerger, S., Oliver, T., & Pirozzolo, F. (1 989). Speech understanding in the elderly. Ear and Hearing, 10 79 89. Jerger, J., Jerger, S., & Pirozzolo, F. (1991). Correlational analysis of speech audiometric scores, hearing loss, age, and cognitive abilities in the elderly. Ear and Hearing, 12( 2) 103 109. Jerger, J. & Jordan, C. (1992). Age related asymmetry on a cued listening task. Ear and Hearing, 13 272 277. Jerger, J. & Musiek, F. (2000). Report of the consensus conference on the diagnosis of auditory processing disorders in school a ged children. Journal of the American Academy of Audiology, 11 467 474. Jerger, J., Oliver, T., & Pirozzolo, F. (1990). Impact of central auditory processing disorder and cognitive deficit on the self assessment of hearing handicap in the elderly. Jou rnal of the American Academy of Audiology, 1 75 80.

PAGE 153

153 Jerger, J., Stach, B., Johnson, K., Loiselle, L., & Jerger, S. (1990). Patterns of abnormality in dichotic listening. In J. Jensen, J. (Ed.), Presbyacusis and Other Age Related Aspects Copenhagen: Stougaard Jensen. Jesteadt, W., Bacon, S. P., & Lehman, J. R. (1982). Forward masking as a function of frequency, masker level, and signal delay. Journal of the Acoustical Society of America, 71(4) 950 962. Jin, S. & Nelson, P. (2006). Speech percept ion in gated noise: the effects of temporal resolution. Journal of the Acoustical Society of America, 119(5) 3097 3108. John, A. & Kreisman, B. (in preparation). Validity of hearing impairment calculation methods for prediction of self reported hearing handicap. Johnson, C. (2000). Children's phoneme identification in reverberation and noise. Journal of Speech Language and Hearing Research, 43(1) 144 157. Johnson, R., Cole, R., Bowers, J., Foiles, S., Nikaido, A., Patrick, J. & Wolliver, R. (1979). Hemispheric efficiency in middle and later adulthood. Cortex, 15 109 119. Kannan, P. M. & Lipscomb, D. M. (1974). Bilateral hearing asymmetry in a large population. Journal of the Acoustical Society of America, 55(5), 1092 1094. Kaplan, H. & Picket t, J. M. (1981). Effects of dichotic/diotic versus monotic presentation on speech understanding in noise in elderly hearing impaired listeners. Ear and Hearing, 2(5) 202 207. Keith, R. (2000). Random Gap Detection Test (RGDT). St. Louis: Auditec. K emper, S. (1992). Language and aging. In F. Craik & T. Salthouse (Eds.) The handbook of aging and cognition (pp. 213 270) Hillsdale, NJ: Erlbaum. Khalfa, S., Morlet, T., Michey, C., Morgon, A., & Collet, L (1997). Evidence of peripheral hearing asy mmetry in humans: clinical implications. Acta Otolaryngologica, 117 192 196. Hearing Review, 4(12) 10 14. Kim, S., Frisina, R. D., Mapes, F. M., Hickman, E. D ., & Frisina, D. R. (2005). Effect of age on binaural speech intelligibility in normal hearing adults. Speech Communication, 48(6) 591 597. Kjellberg, A. (2004). Effects of reverberation time on the cognitive load in speech communication: Theoretical considerations. Noise and Health, 7 11 21. Kock, W. E. (1950). Binaural localization and masking. Journal of the Acoustical Society of America, 22 801 804.

PAGE 154

154 Koehnke, J. & Besing, J. (2001). The effects of aging on binaural and spatial hearing. Semi nars in Hearing, 22(3) 241 254. Koenig, W. (1950). Subjective effects in binaural hearing. Journal of the Acoustical Society of America, 22 61 62. Koenig, A. H., Allen, J. B., Berkley, D. A., & Curtis, T. H. (1977). Determination of masking level di fferences in a reverberant environment. Journal of the Acoustical Society of America 61 1374 1376. Kreisman, B. & Crandell, C. (2002). Frequency modulation (FM) systems for children with normal hearing [Healthy Hearing]. Available at http://www.healthyhearing.com/library/article_content.asp?article_id=160 Accessed June 1, 2007. Larsby, B. & Arlinger, S. (1998). A method for evaluating temporal, spectral and combined temporal spectral resolution of hearing. Scandinavian Audiology, 27(1) 3 12. Larsby, B., Hallgren, M., Lyxell, B., & Arlinger, S. (2005). Cognitive performance and perceived effort in speech processing tasks: effects of different noise backgro unds in normal hearing and hearing impaired subjects. International Journal of Audiology, 44(3) 131 143. Lee, L. W. & Humes, L. E. (1992). Factors associated with speech recognition ability of the hearing impaired elderly. Journal of the American Spee ch Language Hearing Association, 34(10) 212. Letowski, T. & Poch, N. (1996). Comprehension of time compressed speech: Effects of age and speech complexity. Journal of the American Academy of Audiology, 7 447 457. Levy Agresti, J. & Sperry, R. (1968) Differential perception capacities in major and minor hemispheres. Proceedings of the National Academy of Sciences of the United States of America, 61 1151. Lindenberger, U. & Baltes, P. (1994). Sensory functioning and intelligence in old age: A st rong connection. Psychology and Aging, 9 339 355. Lister, J., Besing, J., & Koehnke, J. (2002). Effects of age and frequency disparity on gap discrimination. Journal of the Acoustical Society of America, 111 2793 2800. Lister, J., Koehnke, J., & Bes ing, J. (2000). Binaural gap duration discrimination in listeners with impaired hearing and normal hearing. Ear and Hearing, 21(2) 141 150. Lister, J., Roberts, R. A., Shackelford, J., & Rogers, C. J. (2006). An adaptive clinical test of temporal reso lution. American Journal of Audiology, 15 133 140.

PAGE 155

155 Lister, J. & Tarver, K. (2004). Effect of age on silent gap discrimination in synthetic speech stimuli. Journal of Speech, Language, and Hearing Research, 47(2) 257 268. Litovksy, R. Y. & Macmillan, N. A. (1994). Sound localization precision under conditions of the precedence effect: effects of azimuth and standard stimuli. Journal of the Acoustical Society of America, 96 752 758. Lunner, T. (2003). Cognitive function in relation to hearing aid u se. International Journal of Audiology, 42 (Suppl 1) S49 S58. Lutman, M. E. (1991). Hearing disability in the elderly. Acta Otolaryngologica (Suppl) 476, 239 248. Lutman, M. E., Brown, E. J., & Coles, R. R. A. (1987). Self reported disability and h andicap in the population in relation to pure tone threshold, age, sex and type of hearing loss. British Journal of Audiology, 21 45 58. Lyxell, B., Andersson, U., Borg, E., & Ohlsson, I. S. (2003). Working memory capacity and phonological processing i n deafened adults and individuals with a severe hearing impairment. International Journal of Audiology (Suppl) 42, S86 S89. Madden, J. P. & Feth, L. L. (1992). Temporal resolution in normal hearing and hearing impaired listeners using frequency modulat ed stimuli. Journal of Speech and Hearing Research, 35 436 442. Matthews, L. J., Lee, F S., Mills, J. H., & Schum, D. J. (1990). Audiometric and subjective assessment of hearing handicap. Archives of Otolaryngology, Head and Neck Surgery, 116 1325 13 30. McCoy, S. L., Tun, P. A., Cox, L. C., Colangelo, M., Stewart, R. A., & Wingfield, A. (2005). Quarterly Journal of Experimental Psychology A, 58(1) 22 33. M cCroskey, R. & Keith, R. (1996). Auditory Fusion Test, Revised (AFT R). St. Louis: Auditec. MacKeith, N. W. & Coles, R. R. A. (1971). Binaural advantages in the hearing of speech. Journal of Laryngology and Otology, 85 213 232. McKenna, L. (1993). Some psychological aspects of deafness. In J. Ballantyne, M.C. Martin, & A. Martin (Eds.), Deafness, Fifth Edition (pp. 237 38). London: Whurr. Meguro, Y., Fujii, T., Yamadori, A., Tsukiura, T., Suzuki, K., Okuda, J., & Osaka, M. (2000). The nature of age related decline on the reading span task. Journal of Clinical and Experimental Neuropsychology, 22(3) 391 398.

PAGE 156

156 Middelweerd, M. J., Festen, J. M., & Plomp, R. (1990). Difficulties with speech intelligibility in noise in spite of a normal pure tone audiogram. Audiology, 29(1) 1 7. Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63 81 97. Miller, G. A. & Licklider, J. C. R. (1950). The intelligibil ity of interrupted speech. Journal of the Acoustical Society of America, 22 167 173. Moncur, J. & Dirks, D. (1967). Binaural and monaural speech intelligibility in reverberation. Journal of Speech and Hearing Research, 10 1557 1565. Moore, B. C. J. (1997). An Introduction to the Physiology of Hearing San Diego: Academic Press. Moore, B. C. J. & Glasberg, B. R. (1988). Gap detection with sinusoids and noise in normal, impaired, and electrically stimulated ears. Journal of the Acoustical Society of America, 83 1093 1101. Moore, B. C. J., Glasberg, B. R., Donaldson, E., McPherson, T., & Plack, C. J. (1989). Detection of temporal gaps in sinusoids by normally hearing and hearing impaired subjects. Journal of the Acoustical Society of America, 8 5 1266 1275. Moscovitch, M. & Winocur, G. (1992). The neuropsychology of memory and aging. In F. Craik & T. Salthouse (Eds.), The Handbook of Aging and Cognition (pp. 315 372). Hillsdale, NJ: Erlbaum. Mueller, H. G. & Killion, M. C. (1990). An easy method for calculating the articulation index. The Hearing Journal, 9 14 17. Musiek, F. E., Shinn, J. B., Jirsa, R., Bamiou, D. E., Baran, J. A., & Zaidan, E. (2005). GIN (Gaps In Noise) test performance in subjects with confirmed central auditory ner vous system involvement. Ear and Hearing, 26 608 618. Musiek, F. E., Wilson, D. W., & Pinheiro, M. Ear and Hearing, 5 25 29. K. (1988). Identification of vowels in quiet noise, and reverberation: Relationships with age and hearing loss. Journal of the Acoustical Society of America, 84(2) 476 484. K. & Robinson, P. K. (1982). Monaural and binaural speech perception in reverberation for listeners for vario us ages. Journal of the Acoustical Society of America, 71(5) 1242 1248. Neils, J., Newman, C. W., Hill, M., & Weiler, E. (1991). The effects of rate, sequencing and memory on auditory processing in the elderly. Journal of Gerontology, 46(2) 71 75.

PAGE 157

157 N elson, D. A., & Freyman, R. L. (1987). Temporal resolution in sensorineural hearing impaired listeners. Journal of the Acoustical Society of America, 81, 709 720. Newman, C. W., Jacobson, G. P., Hug, G. A., & Sandridge, S. A. (1997). Perceived hearing handicap of patients with unilateral or mild hearing loss. Annals of Otology, Rhinology, and Laryngology, 106(3) 210 214. Newman, C. W., Weinstein, B. E., Jacobson, G. P., & Hug, G. A. (1990). The Hearing Handicap Inventory for Adults: Psychometric adeq uacy and audiometric correlates. Ear & Hearing, 11 430 433. Nilsson, M., Soli, S. D., & Sullivan, J. A. (1994). Development of the Hearing in Noise Test for the measurement of speech reception thresholds in quiet and in noise. Journal of the Acoustical Society of America, 95(2) 1085 1099. Noble, W. & Gatehouse, S. (2006). Effects of bilateral versus unilateral hearing aid fitting on abilities measured by the Speech, Spatial, and Qualities of Hearing Scale (SSQ). International Journal of Audiology, 4 5(3) 172 181. Norman, S., Kemper, S., & Kynette, D. (1992). Adults' reading comprehension: Effects of syntactic complexity and working memory. Journals of Gerontology: Psychological Sciences, 47 258 265. Norrelgen, F., Lacerda, F., & Forssberg, H. (2001). Temporal resolution of auditory perception in relation to perception, memory, and language skills in typical children. Journal of Learning Disabilities, 34(4) 359 369. Norrelgen, F., Lacerda, F., & Forssberg, H. (2002). Temporal resolution of auditory perception and verbal working memory in 15 children with language impairment Journal of Learning Disabilities, 35(6) 539 545. Ochs, M. T. (1990). Revising the routine audiologic test battery to examine sources of interpatient variability. Jo urnal of the American Academy of Audiology, 1 217 226. Palladino, P. & De Beni, R. (1999). Working memory in aging: maintenance and suppression. Aging,11(5) 301 306. Parente, M., de Taussik, I., Ferreira, E., & Kristensen, C. (2005). Different paren ts of prospective, retrospective, and working memory decline across adulthood. Interamerican Journal of Pscyhology, 39(2) 231 238. Pelosi, L. & Blumhardt, L. D. (1999). Effects of age on working memory: an event related potential study. Cognitive Brai n Research, 7(3) 321 334.

PAGE 158

158 Phillips, D. P. & Hall, S. E. (2000). Independence of frequency channels in auditory gap detection. Journal of the Acoustical Society of America, 108(6) 2957 2963. Phillips, D. P. & Hall, S. E. (2002). Auditory temporal ga p detection for noise markers with partially overlapping and non overlapping spectra. Hearing Research, 174(1 2) 133 141. Phillips, D. P., & Smith, J. C. (2004). Correlation among within channel and between channel auditory gap detection thresholds in normal listeners. Perception, 33 371 378. Phillips, S. L, Gordon Salant, S., Fitzgibbons, P. J., & Yeni Komshian, G. (2000). Frequency and temporal resolution in elderly listeners with good and poor word recognition. Journal of Speech Language and Hear ing Research, 43 217 228. Pichora Fuller, M. K., Schneider, B. A., Benson, N. J., Hamstra, S. J., & Storzer, E. (2006). Effect of age on detection of gaps in speech and nonspeech markers varying in duration and spectral symmetry. Journal of the Acousti cal Society of America, 119(2) 1143 1155. Pichora Fuller, M. K., Schneider, B. A., & Daneman, M. (1995). How young and old adults listen to and remember speech in noise. Journal of the Acoustical Society of America, 97 593 608. Pichora Fuller, M. K., Schneider, B. A., MacDonald, E., Pass, H., & Brown, S. (2007). Temporal jitter disrupts speech intelligibility: A simulation of auditory aging. Hearing Research, 223(1 2) 114 121. Pichora Fuller, M. K. & Singh, G. (2006). Effects of age on auditory and cognitive processing: Implications for hearing aid fitting and audiologic rehabilitation. Trends in Amplification, 10(1) 29 59. Pichora Fuller, M. K. & Souza, P. E. (2003). Effects of aging on auditory processing of speech. International Journal of Audiology 42 (Suppl 2) 2S11 2S16. Plomp, R. (1978). Auditory handicap of hearing impairment and the limited benefit of hearing aids. Journal of the Acoustic Society of America, 63(2) 533 549. Plomp, R. (1986). A signal to noise ratio model for th e speech reception threshold of the hearing impaired. Journal of Speech and Hearing Research, 29 146 154. Plomp, R. & Mimpen, A. M. (1979). Speech reception threshold for sentences as a function of age and noise level. Journal of the Acoustic Society of America, 66(5) 1333 1342. Plyler, P. H. & Hedrick, M. S. (2002). Effects of stimulus presentation level on stop consonant identification in normal hearing and hearing impaired listeners. Journal of the American Academy of Audiology, 13(3) 154 159.

PAGE 159

159 Rabbitt, P. (1991). Mild hearing loss can cause apparent memory failures which increase with age and reduce with IQ. Acta Otolaryngologica (Suppl) 476, 167 176. Rakerd, B., Hartmann, W. M., & Hsu, J. (2000). Echo suppression in the horizontal and med ian sagittal planes. Journal of the Acoustical Society of America, 107 1061 1064. Ricketts, T. (2000). The impact of head angle on monaural and binaural performance with directional and omnidirectional hearing aids. Ear and Hearing, 21(4) 318 328. R osen, S. (1992). Temporal information in speech: Acoustic, auditory, and linguistic aspects. Philosophical Transactions of the Royal Society B: Biological Sciences, 336 367 373. Ryan, E. B., Giles, H., Bartolucci, G., & Henwood, K. (1986). Psycholin guistic and social psychological components of communication by and with the elderly. Language and Communication, 6 1 24. Sands, L. P. & Meredith, W. (1989). Effects of sensory and motor functioning on adult intellectual performance. Journal of Geront ology: Psychological Sciences, 44 P56 P58. Salthouse, T. A. (1991). Mediation of adult age differences in cognition by reductions in working memory and speed of processing. Psychological Science, 2 179 183. Salthouse, T. A. (1992). Influence of proc essing speed on adult age differences in working memory. Acta Psychologica, 79(2) 155 170. Salthouse, T. A. (1994). Aging associations: influence of speed on adult age differences in associative learning. Journal of Experimental Psychology, Learning, Memory, and Cognition, 20(6) 1486 1503. Schneider, B. A. & Hamstra, S. (1999). Gap detection thresholds as a function of tonal duration for younger and older listeners. Journal of Acoustical Society of America, 106(1) 371 380. Schneider, B. A. & Picho ra Fuller, M. K. (2000). Implications of perceptual processing for cognitive aging research. In F. Craik & T. Salthouse (Eds.), The Handbook of Aging and Cognition, Second Edition (pp. 155 219) New York: Erlbaum. Schneider, B. A. & Pichora Fuller, M. K (2001). Age related changes in temporal processing: Implications for speech perception. Seminars in Hearing, 22(3) 227 240. Schneider, B. A., Pichora Fuller, M. K., Kowalchuk, D., & Lamb, M. (1994). Gap detection and the precedence effect in young and old adults. Journal of the Acoustical Society of America, 95 980 991. Schneider, B. A., Speranza, F., & Pichora Fuller, M. K. (1998). Age related changes in temporal resolution: Envelope and intensity effects. Canadian Journal of Experimental Psy chology, 52 184 191.

PAGE 160

160 Schum, D. J., Matthews, L. J., & Lee, F. S. (1991). Actual and predicted word recognition performance of elderly hearing impaired listeners. Journal of Speech and Hearing Research, 34 636 642. Siegel, J. H. & Kim, D. O. (1982). Cochlear biomechanics: Vulnerability to acoustic trauma and other alterations as seen in neural responses and ear canal sound pressure. In D. Hamernik, D. Henderson, & R. Salvi (Eds.), New Perspectives on Noise Induced Hearing Loss (pp. 137 151) New Yo rk: Raven Press. Smoorenburg, G. F. (1992). Speech reception in quiet and in noisy conditions by individuals with noise induced hearing loss in relation to their tone audiogram. Journal of the Acoustical Society of America, 91(1) 421 437. Snell, K. B (1997). Age related changes in temporal gap detection. Journal of the Acoustical Society of America, 101 2214 2220. Snell, K. B. & Frisina, D. R. (2000). Relationships among age related differences in gap detection and word recognition. Journal of the Acoustical Society of America, 107 1615 1626. Snell, K. B., Mapes, F. M., Hickman, E. D., & Frisina, D. R. (2002). Word recognition in competing babble and the effects of age, temporal processing, and absolute sensitivity. Journal of the Acoustical Society of America, 112(2) 720 727. Souza, P. E., Boike, K. T., Withrell, K., and Tremblay, K. (2007). Prediction of speech recognition from audibility in older listeners with hearing loss: Effects of age, amplification, and background noise. Journal of the American Academy of Audiology, 18 54 65. Souza, P. E. & Turner, C. W. (1999). Quantifying the contribution of audibility to recognition of compression amplified speech. Ear and Hearing, 20(1) 12 20. Souza, P. E., Yueh, B., Sarubbi, M., & Loov is, C. F. (2000). Fitting hearing aids with the Articulation Index: Impact on hearing aid effectiveness. Journal of Rehabilitation Research and Development, 37 473 481. Stach, B. A., Jerger, J. F., & Fleming, K. A. (1985). Central presbyacusis: A lo ngitudinal case study. Ear and Hearing, 6 304 306. Stach, B. A., Loiselle, L. H., & Jerger, J. F. (1991). Special hearing aid considerations in elderly patients with auditory processing disorders. Ear and Hearing (Suppl), 12 131s 137s. Stine, E. L. & Wingfield, A. (1987). Process and strategy in memory for speech among younger and older adults. Psychology and Aging, 2(3) 272 279. Stine, E. L., Wingfield, A., & Poon, L. W. (1986). How much and how fast: Rapid processing of language in older adul thood. Psychology and Aging, 1(4), 303 311.

PAGE 161

161 Stephens, D. & Htu, R. (1991). Impairment, disability, and handicap in audiology: Towards a consensus. Audiology, 30 185 200. Strayer, D. L., Wichkens, C. D., & Braune, R. (1987). Adult age differences i n the speed and capacity of information processing: II. An electrophysiological approach. Psychology and Aging, 2 99 110. Strouse, A., Ashmead, D. H., Ohde, R. N., & Grantham, D. W. (1998). Temporal processing in the aging auditory system. Journal o f the Acoustical Society of America, 104 2385 2399. The Psychological Corporation (1997). Wechsler Adult Intelligence Scale Third Edition (WAIS III). San Antonio, TX. Tillman, T. W. & Carhart, R. (1966). An expanded test for speech discrimination u tilizing CNC monosyllabic words. Northwestern University Auditory Test No. 6, Technical Report, SAM TDR 62 135 Brooks Air Force Base, TX: USAF School of Aerospace Medicine, Aerospace Medical Division. Trehub, S., Schneider, B., & Henderson, J (1995). Gap detection in infants, children, and adults. Journal of the Acoustical Society of America, 98 2532 2541. Tun, P., O'Kane, A. G., & Wingfield, A. (2002). Distraction by competing speech in young and older adult listeners. Psychology and A ging, 17 453 467. Tun, P., Wingfield, A., & Stine, E. L. (1991). Speech processing capacity in young and older adults: A dual task study. Psychology and Aging, 6(1) 3 9. Turner, C. W. & Henry, B. A (2002). Benefits of amplification for speech recog nition in background noise. Journal of the Acoustical Society of America, 112(4) 1675 1680. Tyler, R. S., Summerfield, Q., Wood, E. J., & Fernandes, M. A. (1982). Psychoacoustic and phonetic temporal processing in normal and hearing impaired listeners. Journal of the Acoustical Society of America, 72 740 752. Uhlmann, R. F., Larson, E. B., Rees, T. S., Koepsell, T. D., and Duckert, L. G. (1989). Relationship of hearing impairment to dementia and cognitive dysfunction in older adults. Journal of the American Medical Association, 261(13) 1916 1919. van Boxtel, M. P. J., van Beijsterveldt, T., Houx, P. J., Anteunis, L. J. C., Metsemakers, J. F. M., & Jolles, J. (2000). Mild hearing impairment can reduce verbal memory performance in a healthy adult po pulation. Journal of Clinical and Experimental Neuropsychology, 22(1) 147 154.

PAGE 162

162 van Rooij, J. C. & Plomp, R. (1990). Auditive and cognitive factors in speech perception by elderly listeners II: Multivariate analyses. Journal of the Acoustical Society of America, 88 2611 2624. van Rooij, J. C. & Plomp, R. (1991). Auditive and cognitive factors in speech perception by elderly listeners. Acta Otolaryngologica (Suppl), 476 177 181. van Rooij, J. C., Plomp, R., & Orlebeke, J. F. (1989). Auditive and cognitive factors in speech perception by elderly listeners I: Development of test battery. Journal of the Acoustical Society of America, 86 1294 1309. Vaughan, N. & Letowski, T. (1997). Effects of age, speech rate, and type of test on temporal audito ry processing. Journal of Speech, Language, and Hearing Research, 40(5) 1192 1200. Vaughan, N., Storzbach, D., & Furakawa, I. (2006). Sequencing versus non sequencing working memory in understanding of rapid speech by older listeners. Journal of the A merican Academy of Audiology, 17(7) 506 518. Verhaeghen, P. (1999). The effects of age related slowing and working memory of asymptotic recognition performance. Aging Neuropsychology and Cognition, 6 201 213. Weihing, J. A., Musiek, F. E., & Shinn, J B. (2007). The effect of presentation level on the Gaps In Noise test. Journal of the American Academy of Audiology, 18 141 150. Weinstein, B. (2000). Geriatric Audiology New York: Thieme Medical Publishers. Weinstein, B. E. & Ventry, I. M. (198 3). Audiometric correlates of the Hearing Handicap Inventory for the Elderly. Journal of Speech and Hearing Disorders, 48(4) 379 384. Wingfield, A. (1996). Cognitive factors in auditory performance: Context, speed of processing, and constraints of me mory. Journal of the American Academy of Audiology, 7 175 182. Wingfield, A., Poon, L., Lombardi, L., & Lowe, D. (1985). Speed of processing in normal aging: Effects of speech rate, linguistic structure, and processing time Journal of Gerontology, 4 0 579 585. Wingfield, A., Stine, E. A., Lahar, C. J., & Aberdeen, J. S. (1988). Does the capacity of working memory change with age? Experimental Aging Research, 14(2 3) 103 107. Wingfield, A. & Tun, P. (2001). Spoken language comprehension in older adults: Interactions between sensory and cognitive change in normal aging. Seminars in Hearing, 22 287 301.

PAGE 163

163 BIOGRAPHICAL SKETCH Andrew Barnabas John was born in Oklahoma City, Oklahoma Andrew received his bachelor of arts in Communication S ciences and Disorders with a minor in L inguistics from the University of Florida in 2002. He began his doctoral work at the University of Florida in Fall 2003 and was awarded his Ph.D. in Summer 2007. In August 2007, Andrew will join the faculty of the Departme nt of Communication Sciences and Disorders at the University of Oklahoma Health Science Center in Oklahoma City as an assistant professor.