W

I
1
i

J

HIllHlHHl

129

 

TfinuNE

I'M-B RIES

mil“\lll\\‘\\j\‘\l\ll§lli\m \ll \\\\l\\l\\\l !

 

 

 

 

This is to certify that the

thesis entitled

Attentional Demands of Speech Listening for People
with Normal Hearing and with Acquired Sensorineural
Hearing Loss: A Dual Task Assessment.

presented by

Margaret Evelyn Whearty

has been accepted towards fulfillment
of the requirements for

M.A. degree in Audiology

 

Major pro ssor

Date 3/29/93

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

LIBRARY
Elohim lute
Unlverolty

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Ie An Affirmative ActiorVEquel Opportunity Institution
oMWJ

ATTENTIONAL DEMANDS OF SPEECH LISTENING FOR PEOPLE WITH NORMAL HEARING
AND WITH ACQUIRED SENSORINEURAL HEARING LOSS: A DUAL TASK ASSESSMENT

BY

Margaret Evelyn Whearty

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Department of Audiology and Speech Sciences

1993

ABSTRACT

ATTENTIONAL DEMANDS OF SPEECH LISTENING FOR PEOPLE WITH NORMAL HEARING
ANDWITH ACQUIRED SENSORINEURAL HEARING LOSS: A DUAL TASK ASSESSMENT

BY

Margaret Evelyn Whearty

The largest group of hearing aid users are individuals who suffer
frmm acquired sensorineural hearing loss due to aging. Many members of
this group report that hearing'aids only'partially remediate'their speech
listening difficulties: for them, listening to speech remains effortful,
even fatiguing. The hypothesis tested here was that these listeners are
compensating for their diminished ability to hear speech cues with an

exceptional cognitive effort, or commitment of attention.

The study compared eight mild-to-moderate sensorineural hearing
impaired.subjects and eight normal hearing subjects. Their primary task
was to memorize and correctly report back digit strings in competition
with two separate listening conditions: meaningless noise, or a spoken
passage. Fourteen of the sixteen subjects forgot more digits with
speech interference. The more notable finding was that the amount of
added forgetting due to speech interference was significantly greater for
members of the hearing impaired group than for their normal hearing

counterparts.

ACKNOWLEDGEMENTS

I would like to recognize the time, commitment, and assistance of

Dr. Brad Rakerd.

iii

Introduction .

Processes of Attention
The Dual Task Paradigm

The Central Hypothesis

Cognition, Aging, 5 Hearing

Related Studies .

Methm O O O .

TABLE

Cognitive 5 Language Screening .

Experimental Procedure

Apparatus . . .

The Passages .

Scoring of the Data .
Results and Discussion
Statistical Analysis

Individual Differences

Summary . . . .

The Speech Passages

Concluding Remarks

Aging and.Memory Capacity

Future Studies

Methodological Considerations . .

Clinical Note .
Appendix A . .
Appendix B . .
Appendix C . .
Bibliography .

OF

CONTENTS

iv

GUIU'leH

13
14
15
16
18
22

O 27

33
35
39
40
42
42
44
45
49
63

65

Table

Table

Table

Table

Table

2:

LIST OF TABLES

Personal History Details for Acquired Hearing
Loss Subjects . . . . . . . . . . . . . . . . . .

Results of Analysis of Variance Performed on the
Data Presented in Figure 4 . . . . . . . . . . .

Subject Information: Experimental Performance:
Experimental Parameters; Personal History Data .

Number of Questions that a Subject Answered
Correctly Presented with Respect to the Test
Materials There were Five Passages in Each Block
and Five Questions Per Passage; Hence, the

Maximum.Possible Score in Each Block was 25 . . .

Tally of the Number of Subjects Who Missed Each
Question About Passage Content. Maximum.Possible
Tally Was 16 (8 NH subjects a 8 AL subjects) . .

11

25

31

36

38

Figure

Figure

Figure

Figure

Figure

Figure

Figure

LIST OF FIGURES

Combined Audiogram (test ear only) for the Acquired
Loss subjects (n-8, 0 O O O O I O O O O O O I O O O 0

Block by Block Performance of Normal Hearing Subject
SD. The top panel shows the number of digits

' correctly recalled from memory in each block (maximum

possible - 55). The bottom panel shows the number of
content questions about spoken passages correctly
answered per block (maximum possible - 25) . . . . .

Block by Block Performance of Acquired Loss Subject
SA. The top panel shows the number of digits
correctly recalled from memory in each block (maximum
possible - 55) . The bottom panel shows the number of
content questions about spoken passages correctly

answered per block (maximum possible - 25). . . . . .

Average Performance of the Eight Subjects in Each
Experimental Group Presented on a Block by Block
Basis. Open and filled squares represent results from
the normal hearing group: open and filled circles

represent results from the acquired loss group . . .

Average Performance of the Eight Subjects in Each
Group for the Study Overall. . . . . . . . . . . . .

Overall Performance of Individual Subjects. Reported
are the mean number of digits forgotten due to speech
interference. This value is the difference between
the number of digits recalled with no interference and
the number of digits recalled with speech

interference . . . . . . . . . . . . . . . . . . . .

Overall Performance of Individual Subjects. Reported
are the number of digits forgotten due to speech
interference (see figure 6) expressed as a percentage
of the number of digits recalled in the no

interference condition . . . . . . . . . . . . . . .

vi

10

19

20

23

26

28

29

INTRODUCTION

For individuals with normal hearing, listening to and understanding
speech is a pleasant activity that requires little conscious effort or
attention. The task only becomes cognitively effortful when the speech
is degraded as, for example, when it is transmitted.over’a poor telephone
line, or when it is masked by other co-occurring sounds. However, even
under ideal conditions some cognitive resources must be devoted to speech
listening and understanding. Factors that affect the amount of attention
that will be required include the quality of the speech signal itself and
the nature of the environment in which the speech signal is being
transmitted. Thus, varying amounts of attentional resources are needed

to fully process speech signals.

Particularly relevant to the present investigation are studies
assessing the cognitive commitment that normally hearing listeners must
make in order to understand synthetic speech. In one such study, Pisoni,
Manous & Dedina (1987; see Pisoni & Greene, 1990), asked listeners to
report whether the propositions conveyed by sentences were true or false.
Reaction times to reach aidecision.proved.to be significantly longer when
the sentences were produced. by' a high quality speech synthesizer
(Dectalk) than when they were spoken naturally. The authors attributed
this increased response time to the added cognitive effort required to

fully process the synthetic speech signal.

Luce, Feustel & Pisoni (1983) showed that encoding and maintaining
synthetic speech in one's short term memory requires more capacity than
does the memorization of comparable natural speech. They asked subjects

to memorize visually presented digit strings and thereafter to memorize

2

words presented auditorily where the latter were either synthetic or
natural. Subjects recalled fewer of the digits in the synthetic speech

condition.

The implication of these studies is that, even under ideal
listening conditions, high quality synthetic speech must be processed
differently than natural speech by normal hearing listeners. In effect,
synthetic speech must be cognitively processed more fully which requires
more time and more cognitive resources. Pisoni and his colleagues have
argued that this added processing is needed to compensate for the
impoverished nature of the synthetic speech signal. Synthetic speech
lacks many of the phonetic cue redundancies which are present in natural

speech and the listener must work to overcome this limitation.

Individuals with sensorineural hearing impairments are at risk for
their ability to detect all of the naturally occurring speech cues even
under conditions of'optimal hearing'aid fitting (Studebaker & Bess, 1982:
Boothroyd, 1984; Summers and Leek, 1992). They are in effect, receiving
a cue impoverished version of the natural speech signal. The purpose of
the present study was to test the hypothesis that they cope in the same
way that normally hearing listeners do with synthetic speech, that is,
they cope by making an exceptional commitment of cognitive resources to

compensate.

This possibility was tested with particular focus on the largest
group of sensorineural hearing impaired listeners - those who suffer from

hearing loss due to presbycusis, or aging.

PROCESSES OP ATTENTION

This study focused on an individual's allocation of attention.

Attention is a central process that co-ordinates and controls performance

in some task environment (Logan, 1979) and it is generally accepted.that

3
human attentional capacity is limited (Shiffrin & Schneider, 1977).
Basic to the consideration of attentional demands in regards to speech
understanding is the distinction between cognitive processing that is

automatic and processing that is controlled.

WWW' Automatic processing is considered to be
self managed or effortless, requiring little active attention in order

to be carried out. For example, on hearing your own name, you cannot
help but become aware of it. The perception is so ingrained as to have
become virtually automatic. According to Schneider and Shiffrin (1977),
these effortless cognitive processes are activated by a learned sequence
of elements in long term memory, that are initiated by appropriate

inputs.

Another example of an automated process is skilled reading. A good
reader, once presented with a letter string, cannot help but decode its
content. This may best be illustrated by the Stroop Effect. With the
Stroop Effect, subjects are asked to state the physical color of words
presented to them and to ignore the word meanings. However, when the
words happen to be actual names of colors, for example the word yellow
presented in purple ink, meaning interferes with the ability to simply
state the color of the word being presented (Stroop, 1935: Logan, 1985) .
Upon viewing a familiar word, a reader cannot help but read it for

meaning .

Wage; Controlled processes contrast with automatic
processes in that the former demand more attentional resources in order
to be accomplished. .These processes are considered to be "temporary
activations" of sequential elements that, while able to be set up quickly
and easily, require additional conscious effort or attention to execute
(Schneider & Shiffr‘in, 1977) . Examples of situations in which controlled

processing is necessary are search, detection, and memory rehearsal.

Important to keep in mind is that automatic and controlled

4

processes need not be viewed strictly as opposites or dichotomies.
Rather, as Logan (1985) has pointed out, they are different points along
a continuum. This contention receives support from studies using
synthetic speech whereby listeners' performance on speech understanding
and comprehension tasks improve with practice and subsequent familiarity

with the speech signal (Humes et al., 1991: Pisoni (- Greene, 1991) .

THE DUAL TASK PARADIGM

The particular method used to measure allocation of attention is
the dual task paradigm (Anderson & Craik, 1974; Logan, 1979). Early
models of attention as a single channel or single capacity (Broadbent,
1958) have been replaced by models involving multiple resources
(Baddeley, 1992) . However, the concept of a resource limit has remained.
The fundamental premise to any dual task design is that there will be
interactions among task performances if those tasks are drawing from the
same limited resources. In this study, the dual tasks of interest were
listening to a spoken passage or to meaningless noise and memorizing a
string of digits. Storing digit strings in working memory for recall is
an attention demanding, cognitively effortful task (Shiffrin I. Schneider,
1977; Baddeley, 1992). And if listening to and understanding a spoken
passage is, likewise, attention demanding, then it can be expected to
create attentional competition, thereby degrading digit memory
performance. The magnitude of the degradation becomes an index of the
cognitive effort, or attention, required to process the speech passages
for content. Moreover, it can be expected that if hearing impaired
listeners must devote greater attention to speech processing than do
normally hearing listeners, as hypothesized earlier, then hearing
impaired listeners should experience proportionally greater degradation

under the dual task load.

5

THE CENTRAL HYPOTHESIS

The central hypothesis of the present study can be restated as
follows: Sensory-perceptual.processing'of speech is largely'automatic for
normally-hearing listeners and more controlled for individuals with an
acquired.sensorineural hearing loss, even under 'ideal' amplification.and
listening conditions. In other words, speech understanding requires
substantially more cognitive effort, or increased attention, for
listeners with sensorineural hearing impairments due to reasons such as
reduced frequency selectivity and poor temporal processing, which limit
their access to speech cues (Hannely and Dormann, 1983; Gagne, 1988;

CHABA, 1988: CHABA, 1991).

The specific prediction to be tested here is that, relative to
normally-hearing listeners, hearing-impaired listeners should suffer a
greater loss in the performance of a memory task when they must
concurrently focus their attention on speech understanding. If this
hypothesis proves correct, then it would lend support to the contention
that hearing impaired listener's speech processing requires a higher
degree of controlled cognitive processes and that different processing
strategies for effective listening may need to be employed by members of

this group.
COGNITION, AGING, & HEARING LOSS

The focal group of this study were individuals with acquired
mild-to-moderate sensorineural hearing losses who clinically demonstrated
good speech discrimination abilities. When addressed casually, all of
the hearing impaired individuals tested in this study appeared to follow
and understand the majority of conversational content with little to no
difficulty; This study sought to find out whether they were nevertheless

bearing an especially high cognitive burden when speech listening, and

6

hence were more cognitively disadvantaged than one might have realized

on casual or even clinical inspection.

Since the vast majority of hearing aid users are individuals who
suffer from presbycusis, the subjects in this study were drawn from.the
presbycusic population in order to address circumstances specific to
them. The specific interest in the presbycusic population was on
cognitive deficits that result from hearing loss, per se. However, aging
adults often experience cognitive deficits for reasons unrelated to their
hearing status (Otto & McCandless, 1987). Therefore, in an attempt to
'separate out' declining cognitive skills and cognitive effects of
hearing disorders, the hearing impaired subjects in this study were
selected so as to include two subgroups: ”younger” presbycusics and
”older” presbycusics. The control group for this study were normal
hearing young adults - a group comparable to the one used to set the
audiometric standards for normal hearing (ANSI, 1969). It was
anticipated. that the ”younger“ presbycusics ‘would, overall, behave
cognitively much like the control group and that the "older“ group would
behave less like the control group. This separation of age groups would
allow for the appropriate comparisons of cognitive changes due to aging

and for cognitive changes perhaps enhanced by presbycusis.

RELATED STUDIES

A hearing loss present with the elderly person does not necessitate
a cognitive decline. Likewise, it also true that a cognitive decline
need not be accompanied by a hearing impairment. As it so happens the
group of people at highest risk for hearing losses are the same group of
people who are at greatest risk for cognitive deficits (CHABA, 1988) .
In fact, the sort of generalized cognitive decline that accompanies aging
seems to have relatively little to do with speech perception by the

elderly, at least in other contexts. From their research, van Rooij and

7

Plomp (1991), developed the "Auditive Hypothesis" which states that
speech perception is a skill that is relatively impervious to the effects

of aging.

In another study, Humes et a1. (1991) came to sbmilar conclusions
regarding results from speech listening tasks performed by elderly
people with sensorineural hearing losses and younger individuals with
either real or masked hearing losses. Hearing loss determined
performance on tasks, not age differences. This design allowed for a
comparison of performance based on hearing loss and not cognitive

concerns per se.

To reiterate, the focus of the present study was like that of Humes
et al (1991) on cognitive effects that accompany hearing loss, not on

more general cognitive changes of other origin.

METHODS

Subjects; Eight normal hearing adults and eight adults with
acquired sensorineural hearing loss participated in this study. These
subjects were paid for their participation. The subjects with hearing
loss were identified from clinical files from the Oyer
Speech—Language-Hearing Clinic and invited.to participate based.on their

audiological profiles.

AudielogicaLSelestionfiriteriuNomaljearinm The selection

criterion for normal hearing participants was audiometric pure tone
thresholds at or below 20 dB HL bilaterally from 250 Hz to 8,000 Hz. In
addition, subjects had to perform satisfactorily on the individual tasks
that, combined, comprised the dual task paradigm (see below). The eight
subjects in this group ranged in age from 20 - 29 years with a mean age

of 23.6 years.

AudiologicaLSelectionfiriteriammmineuraLLoau A11
hearing impaired subjects had full audiological evaluations within the

previous two years. The audiological criterion for participants with
acquired sensorineural hearing loss was pure tone speech frequency (0.5,
1 and 2 KHz) averages (PTA) indicating bilateral mild-to-moderate hearing
loss; that is, mean PTA's between 30 and 50 dBHL. Another criterion to
be met by these individuals were word recognition scores of 80% or better

bilaterally.

In addition, hearing losses for this group had to be bilaterally
symmetrical, with ear differences at each speech frequency not to exceed

15 dB HL. .Also, a listener's thresholds at each of the speech

9

frequencies had to be similar; specifically differences among the
thresholds at 0.5KHz, 1KHz, and 2KHz could not exceed 20 dB HL. Figure
1 presents a group audiogram for the acquired loss subjects. Within the
speech range, these subjects had essentially flat losses. With this
pattern of loss, the amplification provided by an audiometer (see
apparatus section) can be considered a reasonable approximation of the

target gain needed for individuals within the speech frequency range.

The eight subjects of the hearing impaired group were
counterbalanced, as closely as possible, with respect to age, sex, years
of hearing aid use, and use of binaural or monaural amplification
devices. The subjects had varying degrees of higher frequency
sensorineural loss. Hearing aid experience among the subjects ranged
from hearing aid candidacy to 10 years of use. Subjects with acquired
losses ranged in age from 42 - 68 years with a mean age of 57.0 years.

Details regarding the individual subjects are presented in Table 1.

As previously mentioned, general cognitive deficits are known to
accompany aging (CHABA, 1988) . To permit a statistical dissociation of
these effects for hearing impaired subjects, subject groups were chosen
to represent two age groups: a 'younger' group aged 42 to 52 and an
'older' group aged 60 to 68 years. There were four members in each

group.
COGNITIVE & LANGUAGE SCREENING

On the initial visit, hearing impaired subjects were administered
a subset of the Peabody Picture Receptive Vocabulary test (PPVT) as a
screening measure to ensure that their receptive vocabulary skills
exceeded the grade level of the story passages presented to them.
Specifically, all stories were intended for students at the Junior High
School level and below. All subjects passed this screening. To test the
level of their receptive vocabulary, they were presented 25 items from

the PPVT. Those items began at the median basal item for 15 year olds

lO

HEARING LEVEL (r‘e ANSI - 1989)

00 OD #3 DD
CD CD CD CD

I'llllllll

081

 

(2H) ADNEDDEIHJ
0003 0001 009 098
l ! ! !
+ 0
"' .H._rhhnu,fl..._rh..nlfllw.3.H
+ ' o i
+ o 9
3 i i' i
i i i i

000v
7
2
i“
l

0009
F
9
+
i

 

 

 

Figure 1: Combined Audiogram (test ear only) for the Acquired Loss
Subjects (n=8).

11

memo» 0H Honsmcoz ucmumflmcou ooow monumooz m mm «m
use» m Honsocoz ancoflmnooo ucoaaooxm paw: m mm mm
new» m v amusmcflm HMGOAmmooo oooo monumooz 2 mm on
whom» m downscam ocoumwmcou ooow ououopoz z om mo
mums» OH Housocoz ucmumemcoo ooow oumuoooz m mm <>
use» H v downscam ucmumflmcou unmaaawoxm oaflz 2 mm 08
mumowocmu ucoaaooxm mumuoooz m Hm mu
umo> H v Housmnem ucoumemcou ucoaaooxm monsoooz m me me
haaououoaam

coauooauaagaa nouoom anon

Hensonam .u> omD cad doauwnoooom endeavoum
onD uo nuns» Housndoz unease: who: nooonm Rom mod uoonndm
nuuuuuuuuuuannunuuluuuunuuuuuuuuuuuuuuuuuuunuunnuuuuunuuuunuquuunuuunuuuuunnununuuuuu

.nuoonndm neon unease: mouaevod

How naaouon whounam announce "H wands

12
(Item # 105, Form M). In order to be retained in the study a subject was
obliged to: (1.) Correctly identify a string of eight consecutive items
within the set of 25, i.e., to establish a personal basal that was
greater than that for an average 18 year old and, (2.) To miss no more
than five consecutive items within the list of 25 i.e., not to have

reached a ceiling by the end of the list.

Mirauiamasuamiliarizatimndfirmmncsi On their
first visit, all subjects were administered exercises designed to
familiarize them.with the test procedures. At this time, subjects also
chose their preferred ear for monaural listening and indicated a most
comfortable listening level (MCL) for each stimulus. These levels were
maintained. constant thereafter» JFinallyy a subject's ability' and
willingness to participate were evaluated. Individuals who performed the
individual tasks in an unacceptable manner were to be eliminated;

however, no one was, in fact, eliminated from participation.

IaskJamiliarizatMcitJemizationz A practice block of
five trials was administered to familiarize subjects with the task of

memorizing digit strings. On each trial a unique seven digit string was
presented visually (on a computer screen) for memorization for a period
of 15 seconds. Digits were then blanked from the screen and replaced by
underlines at each digit position. The underlines remained on the screen
for 60 seconds. During this interval the subjects were instructed to
maintain the memory of the digit string presented in any manner they
chose. By report, preferred manners included repeating the digits aloud
or subvocally. At the end of the interval the subjects were required to
report, on paper, the content of the digit strings, with each digit

presented in its correct serial position.

W A practice block of five
trials was run to familiarize subjects with the task of speech listening.

A subject listened to one minute speech passages (presented at MCL) and

13
immediately thereafter answered, in writing, five questions pertaining
to the story content. These questions were taken directly from the
passage and in most instances required one word or short phrase answers.

There was no time constraint imposed on the answering of these questions.

EXPERIMENTAL PROCEDURE

The actual experimental procedure was a more difficult combination
of the two previously described familiarization tasks. It required,
first, that a subject perform the digit memorization task according to
the following protocol: A randomized digit string was presented visually
for memorization for a period of 15 seconds. For most subjects the digit
string length was fixed at 11 digits. This length was established as a
result of a pilot study at Gallaudet University that revealed a ceiling
effect for college aged subjects with the use of 9 digit strings (Rakerd,
Personal Communication) . For a few subjects (n-3) the digit string
length departed from 11 digits in later runs to adjust task difficulty
to match a subject's skill level. In these instances, the fewest digits
presented was 9 and the most was 13. For two subjects the runs began at
7 digits, a modest number, and the number was gradually increased as
dictated by performance. One of these subjects was incremented to 11
digits by the end of the fourth run and kept at that level thereafter.
The other subject did most runs at 11 digits but, in the final two, was

incremented to 13.

Digit memorization was followed by a one minute retention interval
during which the subject performed one of the two listening tasks. Task
NI involved listening to a one minute segment of speech noise, presented
at an individually selected comfortable listening level. This condition
will be referred to as the No Interference condition. Task SI required
listening to a one minute spoken passage about which the listener was

required to answer questions immediately after reporting the digits.

14
This condition will be referred to as the Speech Interference condition.
A complete test consisted of 10 blocks of 5 trials in each listening

condition.

Test sessions were arranged at subject convenience. Sessions
involved four separate meeting times of approximately one and a half
hours each. The first session dealt primarily with familiarization tasks
(see above) and.the completion of one block pair, where a pair comprised
a five trial block with no interference and a five trial block with
speech interference. The completion of the remaining nine block pairs
was evenly distributed over the rest of the sessions. Breaks were given
halfway through a session, and as needed. There was no time constraint
for reporting either the digits or for answering content questions.
Typically three experimental block pairs were completed per visit.
Listening conditions were alternated throughout, with half the subjects
starting in the No Interference condition and half beginning with the

Speech Interference condition.

APPARATUS

During testing a subject was seated in one of two adjacent sound
booths that were connected by a facing patch panel and a double-plated
glass window. All equipment, including the computer screen which
displayed the digit strings, was housed in the experimenters booth. The
computer screen was viewed through the window from.a comfortable viewing

distance.

High fidelity recordings of the spoken passages and speech noise
were played from a cassette tape deck (Kenwood KX46C), amplified by an
audiometer (Grassen Stadler 1710) and presented monaurally via headphones
(TDH-49 Telephonic). Presentation levels ranged from.60 - 80 dB SPL for
hearing impaired subjects and 45'- 65 dB SPL for subjects with normal

hearing.

15

The digit strings were produced.by a random.number generator, and
displayed, in large font, on the screen of a personal computer (AT&T
6300). The computer also served as a timer. After the fifteen second
study interval the digit string was replaced by individual place holders
(underline characters). Sixty seconds after that, the screen displayed
REPORT DIGITS and simultaneously a tone was presented to alert both the
subject and experimenter to the end of a trial; the subject would then
report digits and the experimenter would.prepare for the presentation of

the next trial.

THE PASSAGES

The speech passages were tape recorded in a sound treated room.
They were read comfortably by a speaker who had previously rehearsed the
specific text. A passage was re-recorded if either the reader or the
recording 'engineer' detected a speaking error. Speech passages were
taken from.the 1959 edition of the Junior Britannica Encyclopedia which
is intended for an audience of Junior High School students and below.
Over the last three seconds of recording a speech passage, the intensity
level was 'faded out' to zero. This served the dual purpose of cuing
both the subject and the experimenter to the end of a trial. A single

sample passage and its content questions are presented below. Additional

samples appear in Appendix A.

Passage: The Opossum

Opossums belong to a group of animals called
Marsupial. The females of this group have pouches on the
underside of the body in which the young develop. The
Kangaroos of Australia and the opossums of the United
States are best known of this groupu Opossums are from
9 to 20 inches long. Their fur is grayish white in
color. Their round ears, long narrow tails, and the
palms of their feet are hairless. The inside toe of the
hind foot can be bent like a thumb to meet any of the
other toes. The opossum.uses his hind feet as hands.
They help him climb trees. His long flexible tail is
also used in tree climbingu Opossumm spend alot of time

16

in trees, hunting and eating. They like to eat upside

down. To do this, they wrap their tails around a
branch, hang down, and grasp their food by all four
feet.

Content Questions and Representative Accepted Answers:

1. Which group of animals do opossums belong to?
Marsupials

2. Name another member of the Marsupial family.
Kangaroo

3. What are the colors of the opossum? Grayish White

4. What does the opossum use his hind feet as? Hands

5. How does the opossum like to eat? Upside Down

SCORING OF THE DATA

W The dependent variable in this experiment was the
accuracy of recall for the digit string presented at the beginning of each
trial. Digit strings were scored on a position by position basis, i.e., ‘
to be correct a digit had to appear in its specific location within the
string. Subjects indicated digit position by reporting on an

appropriately marked answer sheet (See Appendix B).

W In a listening task pilot study at Michigan

State University, each recorded passage was played to at least three
subjects under casual listening conditions. Listeners were required to
answer questions, developed by the experimenter, regarding the passages.
No time limit was imposed on the completion of this task. Answers were
scored against an original answer list prepared by the experimenter.
Questions that were responded to correctly by most listeners were kept as
acceptable questions. Questions that were not answered correctly were

either dropped or modified. Participants were encouraged to add questions

17

they felt might be appropriate. A final list, used in the study, was thus

derived.

WW Subjects were instructed to immediately report

digits at the completion of listening to a story passage and then to
respond to the content questions. The questions were provided for
subjects typed out in a separate notebook they could read from. ‘They’were
instructed to record their answers to these questions in the spaces
provided on the back of the digit string answer sheet (See Appendix 8).
They were informed also that for responding to most questions, one or two
word and short phrase answers would be appropriate. Data collected on
speech passages in the study were scored by the experimenter and by a

second judge. Conflicts in scorer judgment were resolved by conference.

RESULTS AND DISCUSSION

WW Figures 2 and 3
represent the complete data records of two subjects: Subject BD, who is a

representative member of the Normally Hearing group (NH): and subject SA,
a representative member of the Acquired Sensorineural Hearing Loss group
(AL). Comparable figures presenting data for the remaining subjects are
included in Appendix C. Symbols in these two figures indicate the number
of digits that a subject correctly recalled in a block of 5 test trials,
where the maximum.possible number was 55 (11 digits per trial x 5 trials
per block). Open squares represent performance in the No Interference
condition, filled squares performance in the Speech Interference
condition. Data are shown for each of the 10 blocks of trials that

together comprised a full test.

Some personal history details such as age and hearing status are
found at the top of the figures. Both subjects generally recalled more
digits in the No Interference condition - open symbols generally lie above
the filled symbols in the figures. This pattern of recall is consistent
with the resource theory of attention, which holds that multiple demands
on attention will cause a decline in collective performance provided that
the tasks are actually tapping into the same resource. In this case, the
task of speech listening combines with the task of digit memorization to
create a larger demand on workingHmemory resources, which in turn leads to
poorer performance on digit recall than occurs with digit memorization

alone.

W The difference between

18

19

 

 

 

 

 

 

 

 

 

Group: NH
Subject: 80
Gender: M
Age(yrs): 23
l I l I l l l l l l
65 - .—
D—--D NO Interference
H SPEECH Interference
50 '- ................ Maximum Possible .—
55 - .—
B
:33 50 — _
'(r
8
m 45- .—
L1
S2
2
35 _ -
30 - _
25}- v
{ I I I I I I I I I I If
I l l l l l I l l l
9.— 25 - ._
90 23 - -
*‘w 21 - _
(DEE
L5”: 19- _
08 17- ..
L I I I I I I I I I I 1
l 2 3 4 S 6 7 8 9 10
TEST BLOCK
Figure 2: Block by Block Performance of Normal hearing Subject BD. The
top panel shows the number of digits correctly recalled from
memory in each block (maximum possible = 55) . The bottom

panel shows the number of content questions about spoken
passages correctly answered per block (maximum possible = 25) .

65

60

55

50

45

40

N DIGITS CORRECT

35

N QUESTIONS
CORRECT
2

Figure 3:

20

 

Group: AL LPTA(dBHL): 41* HAStotus: MON(R)
Subject: SA RPTA(dBHL): 41 HAYrs: 10
Gender: F SRS(%): 88 HAUse: CONSIST.
Age(yrs): 68
l I I I I I I I I I
EI---CI N0 Interference
I—I SPEECH Interference _
— ................ MCXImum POSSIbIe

 

 

 

he

 

5
1 %? I l l I l l l

 

 

(\‘IIIII

1 2 3 4 5 6 7 8 9 10
TEST BLOCK

Block by Block Performance of Acquired Loss Subject SA. The
top panel shows the number of digits correctly recalled from
memory in each block (maximum possible = 55). The bottom
panel shows the number of content questions about spoken
passages correctly answered.per block (maximunlpossible = 25).

21

mean recall in the No Interference condition (NI) and the Speech
Interference condition (SI) represents an estimate of the amount of
forgetting a subject experienced overall due to the attentional demands of
speech listening. Normal hearing subject BD forgot an average of 2.4
digits per block due to speech listening (50.3 NI - 47.9 SI). Acquired
loss subject SA forgot an average of 9.4 digits per block (40.1 NI - 30.7

SI), three and a half times as many her normal hearing counterpart.

A second point to note in Figures 2 and 3 is the overall higher
baseline performance of subject BD relative to subject SA. The former
recalled an average of 50.3 digits in the baseline NI condition, the
latter just 40.1 digits. This difference of 10.2 digits in the No
Interference condition for these two subjects is likely due to the
cognitive changes that accompany aging - subject BD was 23 years old,
subject SA was 68 years old. Older adults tend to perform more poorly
than younger adults on a variety of cognitive tasks, including memory

tasks (Moore, 1982).

The entire pattern of results exhibited by subjects BD and SA is
consistent with the hypothesis that motivated this study, namely, that
individuals with an acquired sensorineural hearing loss are disadvantaged
for speech listening, even when listening with amplification.
Specifically, they are disadvantaged in that listening is more cognitively
effortful with a hearing loss, and hence draws more attention away from
other concurrent cognitive activities which, in the present case, leads to

added forgetting.

" " ” " - - As previously
mentioned, the presbycusic subjects were divided into 'younger' and
'older' subgroups. When these two subgroups baseline performance on digit
recall without the competition of speech listening was compared, the data
did not support the anticipated result of 'younger' subjects having digit

recall performance that was better and hence more like that of normally

22
hearing young adults. In fact, there was no difference between baseline
performance for the two subgroups of presbycusic subjects. (”Younger"
presbycusic mean baseline performance NI condition - 39.9 out of 55:
'older" presbycusics baseline performance NI condition - 39.7 out of 55) .
Given the near identical performance of the two subgroups, their data were

pooled in all analyses.

W Figure 4 shows
the average performance of the eight subjects in each experimental group.

Specifically, the figure shows means for the eight subjects in a group on
a block by block basis. As in Figures 2 and 3, open symbols represent
performance under the No Interference condition and filled symbols
performance under the Speech Interference condition. Squares in Figure 4
(both open and filled) show results for the normally hearing subjects,

circles show results of the acquired loss subjects.

STATISTICAL ANALYSIS

Statistical differences among the means shown in Figure 4 were
assessed with a three factor (1 between, 2 within) mixed model analysis of

variance (Meyers, 1972), as detailed in Table 2.

WWW GROUP, a two-level (NH-normal hearing,
AL-acquired loss) between-subjects factor indexed the cognitive effect of

aging remarked earlier. On average, the younger, normally hearing
subjects were able to remember 9 more digits than their elderly hearing
impaired counterparts (mean Nil-47.4, mean AL-38.4): this difference was

Significant [F(l,14)-7.7l,° p-0.0l4].

Attentiomanifipeechjisfsninm INTERFERENCE was a within-subjects

factor with two levels (NI-No Interference, SI-Speech Interference). On
average, the sixteen subjects recalled 44.6 digits per block with no

interference and 41.2 digits with speech interference. The difference of

23

 

60-

55-

45h-

N DIGITS CORRECT

35-

30*-

25

‘I

 

Figure 4:

1 l I I I I I I I I

'D——EJ NH -- No Interfer. O—-O AL -- No Interfer. _
H NH -- Spch Interfer. H AL -- Spch Interfer.
-------------- Maximum Possible

 

 

,
I l l I I I [J I If

Average Performance of the Eight Subjects in Each Experimental
Group Presented on a Block by block Basis. Open and filled
squares represent results from the normal hearing group; open
and filled circles represent results from the acquired loss
group.

24

3.4 digits, which was significant [F(1,14)=16.8; p-.001], represents the
forgetting 'from. memory that can be attributed to the attentional

requirements of speech listening in the experiment overall.

Of particular relevance to the present study is whether the
magnitude of this forgetting was different for the two subject groups.
The relevant means are shown in Figure 5. Normally hearing subjects
forgot an average of just 1.6 digits due to speech interference, whereas

acquired loss subjects forgot 5.1 digits, more than three times as many.

This interaction between GROUP and INTERFERENCE was significant
[F(1,14) - 4.6; p< 0.05]. This finding supports the hypothesis under
test. The acquired loss group was influenced by the added cognitive
demand of speech listening to a greater extent than the normally hearing
group. However, the normally hearing listeners were not completely imune
to the cognitive demands of speech listening and, although to a much
lesser extent, were consistently influenced by the presence of story

passages.

‘Eractige; BLOCKS was a second within subjects factor included to
test the possibility that there were practice effects across the 10 block
pairs of trials comprising an experiment. The generally upward slope of
the functions shown in Figure 4 suggests that there was a practice effect
and, indeed, the main effect of Blocks was significant [F(9,126)-14.84:

p-0.000].

Importantly, however, neither'of the other two factors significantly
interacted with BLOCKS: nor was there a three way interaction (p > 0.05).
Hence the practice effects were roughly equal for the two subject groups,

both when they were listening with and without speech interference.

Note that the performance of the acquired loss group never
approached that of the normally-hearing group at any point. Also note

that neither group had reached "ceiling" by the end of the study.

25

o.em m.Hmom mmH 4\umm
mms.o mm.o m.mH m.msH a one
mme.o sm.o m.m~ H.¢H~ m on
m.m~ m.Hssm mmH «\om
mHm.o sm.H m.os m.mem m as
ooo.o em.eH m.mev s.mmmm m u soon
m.mm H.mms «H «\mm
omo.o e.s m.mmm «.mmm H m\¢
Hoo.o m.mH m.msm m.msm H m mosaumuumunH
uuumnnsw
sHsuHs
m.e~m o.msmHH «H _<\m
mHo.o Hs.s m.vmmm G.smmm H a ozone
.mmmmfimmw
“0030”“
HuHHHHosooum oHusmIs ms mm as wousom oHouHuus

.v ounces as vouduuowm nude on» do
cuawouwom ouaoawob no awakened no mudsmom «N Hausa

26

WHEARTY (1992)

 

50

4:.
CD
I

on
01
I

NUMBER OF DIGITS RECALLED (per block)
a.
CD
I

 

CI No Interference
I Speech Interference

 

 

 

 

 

 

 

 

 

 

 

30

Figure 5:

NORMAL ACQUIRED
HEARING LOSS

Average Performance of the Eight Subjects in Each Group for
the Study Overall.

27

INDIVIDUAL DIFFERENCES

Figures 6 and 7 show the amount of forgetting due to speech
interference that was measured for each individual subject. Figure 6
presents the "raw" scores, i.e., the mean difference between the number of
digits remembered under no interference and speech interference condi-
tions. In Figure 7, the data have been scored relative to a subjects'
baseline performance level in the NI condition to adjust for the cognitive
differences between subject groups. Specifically, the number of digits a
subject forgot due to speech interference has been «expressed. as a

percentage of the No Interference mean, as follows:

% Forgotten = W
Mean NI

The "raw" and percentage analyses (Figures 6 and 7) tell a similar
story about the performance of individuals in the two subject groups,
although the between group differences are more apparent in percentage

terms.

ANQrmally_flearing_$nbjegtsi_ Under both forms of analysis, six of the

eight normally hearing subjects performed in a comparable fashion, with
forgetting of about 1 to 5 digits or 2 to 10%. The two remaining NH
subjects had negligible amounts of forgetting (JS), or measurable
"negative" forgetting (AR). Possible reasons for these results are

discussed later in this paper.

tAcquired_Loss_Sub1ecrsi, Regarding the percentage digits forgotten
(Figure 7), six of the eight acquired loss subjects jperformedmmore poorly

than even the worst normally hearing subject. The data reported in Tables
1 and 3 regarding personal and audiological information were analyzed to
determine whether any of that information suggested a basis for distin-

guishing between these six acquired loss subjects who had notable

Figure 6:

28

 

 

 

 

 

SA YA CB DD EE TC 08 GS

 

 

 

 

 

INDIVIDUAL SUBJECT PERFORMANCE

[:3 NORMAL HEARING SUBJECTS
m ACQUIRED LOSS SUBJECTS
80 AH AR JS AT NT LW TW

 

F ORGETTING DUE TO SPEECH INTERFERENCE:

 

I__

 

 

 

I L l

N O) (O m C 0')

 

M3018 83d Nauosuos sumo NVBW

Overall Performance of Individual Subjects. Reported are the
number of digits forgotten due to speech interference. This
value is the difference between the number of digits recalled
with no interference and the number of digits recalled with
speech interference.

FORGETTING DUE TO SPEECH INTERFERENCE
INDIVIDUAL SUBJECT PERFORMANCE

Figure 7:

29

 

 

 
 

    
 
 
   
       
        

 

   

   

       
   
    

 

  
  
  

  
     

 

 
 
       

              
    
    
    
 

        

   

 

  
         
   
 

      

      

  

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

U)
‘ (9
"O‘.'.'.'.'.m'.’.'
.o:ozozozozozozozozozozozo:o w
&&&&&&&&&&&&&3' ‘3
(9
F-
: same: a:
. v . v v . v . . . :‘3‘3‘3‘3‘3‘3‘3‘3‘3‘3‘3‘3‘:
fl9”UVHNRRRRPUUVNNRRRRRRRVH C)
JVVNNNflflflfififlfifififififififi§$993 c)
%%&P%fififififlfififi§$i an
VVNQflfiflflfififlfihflflu
mosssssivsssss. U
!dflflfififififi§95dVfl-flfifififififi$fl9¥¥¥¥¥¥¥~¥$¥% <t
UVVVVVVVVVVVVVVVVVVVVUVVVVV”-VVVVV%- )~
”flflflflflfififififififlfiflflfiflhufiflflgflu%% <(
>dVV?UVVVVVVWVVVVVVVVNNWVUVV§ (n
U) &&flaafififi£flEMHHUHUNMMQ&&&&&&&&L
'5 fl
ILICJ
OD")
3% __
Z 3 I '2
H
9% 3 [__ :2
DJ
:IZCD U)
m R
4
<(II: a:
s—s <
E D
0: C! l 3;:
C)()
2 < O
[] DD
0 V" 0 N (O O (D
m N -—o -- I

NELLLOSHOJ 81.1910 :10 BOVINBOHBd

Overall Performance of Individual Subjects.

Reported are the

number of digits forgotten due to speech interference (see
Figure 6) expressed as a percentage of the number of digits

recalled in the no interference condition.

30

difficulty and the two who did not. No basis could be found. Factors
that were considered included: age, sex, PTA for the test ear, higher
frequency loss (i.e. > ZKHz), hearing loss configuration, word recogni-
tion score for the test ear, selected comfortable listening levels for
both conditions, frequency of hearing aid use, hearing aid configuration
(i.e. monaural vs. binaural amplification), years of hearing aid use,
type of hearing loss (i.e. purely' sensorineural oerixed): average number
of questions answered correctly over all test blocks of speech listening
and, case histories. This lack of correlation suggests that the
underlying pathology of a nuldrto-moderate sensorineural hearing
impairment, common in all subjects, played the key role in determining

task performance.

Two subjects of interest are members from.the smaller acquired loss
subgroup having overlapping performance with certain individuals from.the
normally hearing group. Subject TG shared similarities in task perfor-
mance with normally hearing group member JS to some extent, but more so
with subject AR, in that the speech interference task did not influence
digit recall performance as would be expected,- the presence of story
passages did not adversely influence digit recall ability under that
condition. To the contrary, it led to a slightly ”negative“ effect. The
only detectable similarities between these two subjects comes by way of
anecdotal remarks. Both said it was easier to remember digits when a
story passage was presented because it allowed them to focus their
attention on the task at hand, digit memorization, and both felt the

presentation of speech noise alone to be distracting.

Subject GS can be compared to another member of the normally'hearing
group, TW. Both GS and TW were presented with longer digit strings of 13
on later trials due to ceiling effects with 11 digits. These two subjects
share a common background regarding work with numbers. GS was an

accountant who reported that he encoded the 11 digit string as an 1 - 800

31

\conH> e.mH on me muoumooz 0H“: Ammv OB o.HI wb.mn
musmOme unwoz

\msuHGGHB
\Cmew> b.0N ow om mum>mm ODMHOGOS Aomv m0 o.N wv.a

CCCCCCCCCCCGCCCCCCCCCCCCC“.CCCC{.¥.C.CCCCCCC§CCCCCCUiCCCCCCCCCCC{‘QCCCl.i..§*l.t.iizlz¢1*.##fifl

mcoeuommcH new

\cofime> m.mH cm or monumooz Amvv ma m.v wm.HH
mcoHuomucH mom
\coemw> m.eH mm me paw: monumooz Heme mu m.m wm.mH
GOemH> m.>H om ms mumumooz paw: Home mm m.m wo.mH
cone> m.Hm on mm mumumooz Ammv «m e.m wv.mm
musmonxm unwoz m.bH on me mum>mm wuouwooz Ammv an m.m we.mH
mauHccHB
\cowme> mH on mm muu>mm muouuooz Ammv «M m.oH w>.mm
uuowuoo new an: new an: anon
Snooper unadunosa ouwoz monsoon hunmsvouh “made uouuou souuom
menu a use: nuoonm Shaun Adowmbo voohndm nuHu«a# moanenx

.uuea hhouudm Huaouuom
«pneumaduom Houaoawwmnfiu «muddauouuom Hounoaewonxu «noeuufiwoudn uoohnsm an Handy

32

number and subject TW was a student studying mathematics. While GS may
not have been as cognitively taxed as those who did not devise such
effective memorization strategies, the presence of speech passages

nevertheless impaired his ability to correctly report digits under that

condition.

The two members of the acquired loss group, TG and GS, have already
been discussed in relation to members of the normally hearing group but
have not been compared.to one another; No compelling similar characteris-
tics between the two subjects could be found. The only characteristic
that could be considered a similarity between subjects TG and GS was use
of binaural amplification, however, this use is not exclusive to these two
subjects and the years of use discredits that similarity due to the fact
that TG is a new user of binaural hearing aids and GS has been using

binaural amplification consistently for eight years.

Subjects SA, YA, CB, DD, EE, and DS comprise the larger portion of
the acquired loss group - those who were cognitively burdened by speech
listening. Their heterogeneity is apparent in the following summary of
personal variables. The ages of these subjects range from 42 (DS) to 68
(SA) years, one member of this group was male (DD). All hearing losses
for speech frequencies except one, which was mild (BE), are characterized
as moderate with higher frequency losses ranging from mild to severe
including one subject (DD) with a conductive component at that level.
Word recognition scores ranged from 80% to 100%. Dissimdlar patterns,
years and type (monaural vs. binaural) of hearing aid use ranging from
hearing aid candidacy (no use at present time C8) to 10 years of use (SA)

are also characteristic of this group.

Considering performance on story passage questions does not reveal
anything either. While TG and Gs performed comparably with each other on
story content, their scores fall in the mddrange of the other members of

the acquired loss group who performed more poorly on digit recall

33

measures .

In a review of case histories, all subjects except DD reported a
visual impairment. At least one member of each subgroup had a history of
noise exposure, ear infections and.tinnitus. Two subjects (SA and.EE) had
no related audiological complaints in their case histories other than
hearing loss. Both of these individuals did report visual impairments and

were part of the cognitively affected subgroup.

From these findings, there is no strong evidence that any one of
these, or any combination of these factors are in and of themselves
critical in the determination of attentional requirements for speech
listening. Rather the attention demand appears to be a quite general
phenomena that is likely to affect most individuals who have at least a

moderate acquired sensorineural hearing loss.

SUMMARY

This study was designed so that conditions would at least
approximate those of hearing aid listening for the ear under test: A
relatively flat loss over the speech frequency range was corrected by
amplification to MCL with an audiometer. It is apparent that amplifica-
tion of this type was insufficient to overcome the speech processing
difficulties by elderly hearing impaired individuals who were obliged to

listen to a lengthy and unfamiliar sample of speech.

The findings reported here are supportive of the hypothesis that
individuals with acquired sensorineural hearing loss, are placed at a
cognitive disadvantage when speech listening, particularly when listening
in conjunction with another attention demanding task. Consistent with
this hypothesis, subject SA made verbal reports throughout the
experimental procedure that either the digit string, after the presenta-

tion of a story passage, was "completely gone" or "it was as if it never

34
happened". Due to an increased effort, or greater attentional demand on
the part of the listener, or due to the presence of speech that is not
easily decoded by the individual with a compromised auditory system, this
is the dilemma in which these individuals are placed. SA also commented
that "it was one or the other"; that is, she found it difficult to divide
her attention between digit recall and attending to a speech passage.
Again, this may be related to speech signals requiring a greater amount of
processing time due to a lack of accessible phonetic cues and a lack of
natural language redundancy provided by those cues to facilitate the

decoding of the message.

That subjects with normal hearing generally forgot more digits under
the Speech Interference condition than No Interference condition,
supports the premise that some cognitive resources must be devoted to
speech listening and understanding even under the most ideal circumstanc-
es. As suggested by Logan (1979), this indicates that speech understand-

ing is in fact only automatized by degree, along a continuum.

THE SPEECH PASSAGES

WWW As noted above. some

care was taken in the selection of passages for the speech listening
condition of the study and in the development of content questions
assessing listeners' understanding of those passages. This section

reports the details of subject performance regarding these materials.

Table 4 shows the number of questions answered correctly out of 25
by each subject in each test block. It shows that members of the acquired
hearing loss group perform more poorly on answering content questions than
do normally hearing listeners under the dual task of digit memorization.
The combined.mean number of questions answered correctly for the normally
hearing group was 19.5, compared to 17.6 for the acquired loss group.
Also, the overall pattern of normally' hearing individuals having a higher
level of performance or being influenced to a lesser extent on digit
recall by additional cognitive demands is clear. On average, per block,
the acquired loss group never out-performed the normally hearing

individuals on this task.

Interestingly, a similar pattern of content question errors is
followed by both subject groups. Story blocks that proved difficult for
the normally hearing individuals, also proved difficult for the

individuals with acquired hearing losses.

The_Eragtige_Blgck; All subjects answered at least 19 of the 25
practice block questions correctly. This indicates that the content of
the passages was accessible to all subjects when listening without dual

task pressures.

35

36

b.ba .m.mH ¢.mH m.mH m.mH v.mH m.mH N.mH m.ma m.mH m.bH new:

s.oN NH NH HN MN ON NN HN mH «N NH mo
m.mH sH mH MH «H 5H 5H m mH mH NH mo
s.mH sH ON ON MN 0N HN NH sH 5H ON o9
a.sH NH mH pH NH pH HN 5H mH ON NH mm
m.sH mH mH sH ON mH NH mH mH mH NH on
m.eH sH .OH mH NH sH mH MH NH mH MH mo
o.mH sH mH mH mH mH NH eH mH mH NN as
m.HN NH mN mN mN HN sN ON mH mN NH 4m

Andy
someone

m.mH m.mH m.mH H.0N m.mH G.ON m.ON s.mH m.sH m.0N m.mH sums
o.HN HN NN NN ON HN mN HN mH MN mH as
e.ON NH mH eN HN mH mN pH HN ON ON so
o.NH NN mH NH HN NN HN mH mH GH mH ez
s.mH HN HN ON mH HN ON 5H mH HN NH ea
m.sH sH pH NH GH mH mH HH HN NH mH mu
0.0N NN HN sH NN HN ON mH 5H mN NH ms
N.mH mH HN mH mH NN ON NH sH NH NN as
o.ON sH mH sN HN HN 0N NH NH NN HN on

.szs
sums 0H m m s m m e m N H someone

mmuflmmdm NO XUOHQ

.mm was 3035 noon 5" ououm mane—Human 553sz on» .0233 32353 new aficionado
95h odd 300.3. noun a.“ seasoned 95h owes swoon. .uaoaumuoz been. one 0» bosom—om
use? ombudsman hauueuuoo coughed boonnHHm a some. escapees no Monasz ”v ”ABE...

37

Response—Styles; Normally hearing subjects, who were stu-

dents,tended to respond to all of the questions presented to them. The
lack of response to all questions for a given passage on the part of the
acquired hearing loss group may possibly be attributed to their reduced
exposure to 'test taking' situations or perhaps an uneasiness about

writing a potentially incorrect answer.

Ind11idnal_§gntent_flnestignai Content questions were intended to
pose a roughly equal amount of difficulty per block of stories. Table 5
reports question difficulty as a tally of the number of times each
question was misseda It also shows that the distribution of difficult and
easier questions was fairly even across the passage blocks, with the
exception of blocks 3 and 4 which were somewhat more difficult than the
others. However, it is not apparent that particular story passages
influenced digit recall ability for subjects on particular blocks. It
appears to be the case that the added task of attending to a speech
passage was in itself sufficient in almost all cases to distract the

listener from the digit recall task regardless of: the story content.

Maximum Possible
10

Tally Was 16 (8 NH subjects + 8 AL subjects).

38

Block of Passages

2

Tally of the number of Subjects Who Missed Each
Passage# Question#

Question About Passage Content.

1

TABLE 5:

1

07306

48120

13805

02201

20124

02283

12345

06161

02961

18365

01022

05131

02312

13918

06436

01695

12345

11
10
15

01703

40204

01431

21726

2924.7

05161

01453

12345

00001
09830
90242
23934
09907
06533
01311

14445
1

12345

22062
06053
01922
1
58958
51020
02800

13284

16531

01161

12345

CONCLUDING REMARKS

WWW Individuals with acquired

sensorineural hearing loss due to aging - by far the largest hearing
impaired group and the largest group of hearing aid users - were the focal
group of this study. Members of this group, particularly those of advanced
age, may be expected to have some cognitive deficits, as well as hearing
deficits (CHABA, 1988) . Both of these deficits, general cognitive
decline and hearing loss, have the potential to influence the perception
of speech. Cognitive factors influence 'top down” processing and
peripheral sensory factors influence “bottom up" processing. An original
intent of this paper was to compare the performance on digit recall among
the 'younger' and 'older' members of the acquired hearing loss group to
index cognitive effects. This failed, in the end, because the digit
recall performance between these two subgroups did not differ. Therefore,
we were unable to 'factor out' cognitive influences due to aging. This is
particularly true since control subjects with normal hearing actually
learned to some extent over the course of this experiment, as evidenced by
their overall improved performance on digit recall in conjunction with a
speech listening task. This means that normally hearing subjects were not
operating exclusively under 'automatic' processes but also were
influenced by controlled processes. Speech processing in and of itself is
an attention demanding endeavor for all individuals, and just how this
interacts with cognitive deficits due to aging remains a matter of

question.

One thing that can be said, unequivocally, is that the present
findings show elderly individuals with at least mild-to-moderate
sensorineural hearing losses to be “at risk“ when they must listen to

39

40

speech in complicated circumstances, or when involved with competing

tasks. Whether the explanation for that finding will ultimately turn on
their hearing loss or on aging factors cannot be so confidently stated.
There are, though, several reasons to think that hearing loss will play

the greater role.

AGING AND MEMORY CAPACITY

Smith and Fullerton (1981), have reported that working memory
capacity changes little with aging. Therefore the results obtained here
cannot be explained through an overall deficit in cognition due to aging,

leaving the presence of hearing loss a factor.

 

Listeners; In a study by Humes et al. (1991), performance on speech
perception tasks by elderly hearing impaired subjects was compared to
performance by normally hearing college age subjects, where the latter
listened under normal conditions and under conditions of a simulated
presbycusic hearing loss. The experimental design of this study is
important because it allows for two separate: comparisons; the first being
the natural comparison of performance among the elderly and young subjects
which indexes the general effects of aging, and the second, due to the
intact cognitive differences in the face of matched peripheral losses,
allows a comparison of performance based on hearing loss and.not cognitive

concerns per 86.

Similar performance was observed between the two groups with hearing
loss, real and masked, despite age differences. In fact, the older
subjects with the actual sensorineural loss performed.slightly'better than
the younger 'hearing impaired' group. This could be attributed to the
lack of familiarity this group had with attending to speech signals in

lieu of a hearing loss or conversely the ability of the elderly group with

41

the actual hearing loss to utilize their top down processes compensating
for their less accessible bottom up strategies. These results suggest
that the peripheral hearing loss and not effects of cognition per as were

responsible for the increased difficulty.

Wain Further support for this line of thinking
comes from research by van Rooij and Plomp (1991) who worked with speech

reception thresholds (SRT's) as a measure of speech perception perfor-
mance. They found that SRT's of the elderly can be predicted for better
by the audiogram than by measures of cognitive skill. They are proponents
of the "Auditive“ hypothesis, which states that speech perception is a

skill that is relatively impervious to the effects of aging.

 

W13 Finally, in a study at Gall-Bud“
University, utilizing the same experimental procedures as this study,

college age individuals with moderate-to-severe congenital losses were
tested. When those results are compared to the data obtained from
normally hearing college age subjects in the present study, there is a
significant difference in ability to recall digits under the speech
listening condition with the hearing impaired subjects performing much

more poorly. Considering the age match, peripheral hearing loss and not

cognitive factors would seem to account for this discrepancy in
performance.
Hannely and Dormann (1983) caution, however, that not all

sensorineural hearing losses based on audiological configuration alone can
be considered homogeneous. There most likely exist great amounts of
intersubject variability with varying degrees of cochlear pathology and
unspecified cochlear distortions. This may account for subject
variability on performance for the acquired loss group while not negating
the general finding that it is predominantly hearing loss and not

cognitive factors that reduce speech intelligibility.

42

FUTURE STUDIES

Recent studies have looked specifically at the effects of cognition
due to aging and influence of peripheral hearing loss on the ability to
perceive and understand speech as separate factors (CHABA, 1988) . More
studies are needed in this area. Regarding the present study, an
additional experimental group of age matched individuals for the acquired
loss group with normal or near normal hearing abilities would have allowed
us to address the issue of cognitive factors more explicitly. Another
approach would have been to mimic a specific hearing loss in a large group
of subjects with normal hearing and compare performance on digit recall
ability under some speech listening or speech monitoring condition. If
performance proved to be similar among a large population of different age
groups, hearing loss would be shown to be the main contributing factor to
performance. However, if poorer performance tended to be among older

individuals of this group, cognitive decline may be the crucial factor.
METHODOLOGICAL CONSIDERATIONS

If this study was to be extended using larger populations it would
be necessary to streamline the procedure to make data collection more
feasible. Ways in which it would be appropriate to modify present
procedures are to reduce the total number of block pairs, perhaps from ten
down to five. Subject performance after five blocks begins to plateau
(see Figure 4) . Perhaps analysis of the data from the last three block
pairs of 5 would be sufficient, the initial two block pairs being
considered practice blocks. The actual time required for an individual
trial could also be reduced by presenting a 30 second story passage as
opposed to a full minute segment. Finally, digits could be presented
serially for a brief period each. Informal comments by subjects, and
personal experiences of the experimenter suggest that this duration would

be adequate. Also, serial presentation of single digits would not only

43

tax the working memory in a different way - precluding Coding of digit

information - but also for practical concerns require less time.

In regards to addressing the question of cognition more directly, a
simple alteration to the present method would be to employ the use of
computer operated response systems that recorded response time for both
digit recall and answers to content questions. Questions to story

passages would be presented in a multiple choice form.

Finally, the present study required subjects to engage in a
listening task that was familiar and replete with linguistic, syntactic
and semantic cues. Its virtue was that the task tapped the language
processing system as a whole. But perhaps the next study could begin to
tap different language sub-processes. For example, a largely phonetic
listening task would involve detecting the presence of a single phoneme or
making the distinction between two separate phonemes. Lexical decision
tasks involving the use of related and unrelated primes with their
reaction times or a task involving word identification from definitions
would probe higher levels of processing than phoneme detection alone. In
determining the amount of attention required to more fully process a
speech signal, tasks requiring syntactic and semantic decisions would be
useful. This type of decision would not only entail the correct encoding
of the acoustic phonetic information (bottom up processes), but would
also rely on the intervention of knowledge of the linguistic rules and
familiarity with the language (top down processing) by the subject. Being
able to decipher which levels of processing prove to be most difficult is
not only important for hearing impaired listeners but could also provide
more specific insights into the cognitive demands of processing speech

information .

44

CLINICAL NOTE

An important clinical finding from this study is the necessity of
those with acquired sensorineural hearing losses to focus their attention
during speech listening to the speech signal itself and avoid additional
or simultaneous inputs of information. This does not necessarily imply
a cognitive deficit but more specifically a hearing loss that is

unforgiving of multiple channels of input in regards to speech listening.

APPENDICES

APPENDIX A

45

APPENDIX A

NAPOLEAN

Napolean Bonaparte was born at Ajaccio in Corsica. His father was poor
but of noble blood. His mother was an energetic and determined woman.
When he was a child his favorite toys were a drum.and a sword. In 1779
he went to school at Brienne in France where he took great interest in
history; He was especially interested.in the lives of the great generals
of antiquity as told by Plutarch. He entered the mdlitary school at
Paris in 1784. A year later he was appointed to the artillery regiment
stationed at valence. Although his notes and essays show that he was
rapidly rastering his profession Napolean spent little time with his
regiment. He was determined to free Corsica which was forcibly annexed

by France.

1. Who was this story about? Napolean

2. What were Napoleans favorite toys? Drum.and Sword

3. Where did he go to military school? France

4. Did his notes and.essays show he was rapidly'mastering

his profession? Yes

5. What was Napolean determined to free? Corsica

46

APPENDIX A

OBCIOLA

One of the most famous chiefs of the Seminoles was Osceola. He was their
leader during their second war against the United States government.
Osceola was the son of an Englishman named William.Powell, and a Creek
woman, the daughter of a chief. He was born in Georgia, near the
Chattahoochee River. When he was four, his mother took him.to Florida
to live among the Seminoles and he became their chief. In 1832 the
Seminoles signed a treaty with the United States by’which they agreed to
move west of the Mississippi. Osceola was angry with the white man
because of unfair treatment he had received at the hands of the Indian
agent, General Wiley Thompson. He urged the Seminoles not to carry out
the treaty. The war began in 1835 when Osceola killed the Indian agent

and his men killed 100 United States soldiers.

1. Who is the story about? Osceola

2. Who did he live among? Semdnoles

3. Was Osceola the son of an Englishman? Yes

4. Where did his people agree to move? West of the Mississippi
5. Why did war begin? Osceola killed the Indian agent and his

men killed 100 United States soldiers.

47

APPENDIX A

NATIONAL ANTHEM

During the war of 1812, British ships under Sir Alexander Cockburn
shelled Fort McHenry near Baltimore. A young United States lawyer,
Francis Scott Key, who had gone to make arrangements with the British for
the release of a friend, was detained on board his ship during the
battle. Key spent the night watching the bombardment. When morning
came, and through the mist he saw the United States flag still flying
over Fort McHenry, his relief and joy were so great that he was inspired
to write the words of “The Star Spangled Banner." Key's lines were
almost immediately set to the music of an old.English song, "Anacreon in
Heaven", by John Stafford Smith. The song with its new words became very
popular and was officially declared the National Anthem of the United

States in March, 1931. ”My Country Tis of Thee”, is not a national

anthem.
1. Who is the story about? Francis Scott Key
2. What did Key spend the night doing? Watch Bombardment
3. What did key write? Star Spangled Banner
4. What song were the lyrics of the National Anthem put to? Old

English Song *Anacreon in Heaven".

5. Is ”My Country Tis of Thee" a national anthem? No

48

APPENDIX A

NAVIGATION

Navigation is the science of finding the position and directing the
course of vessels at sea. When speaking of airplanes, the term, "air
navigation” is used. Piloting is the part of this science in which a
vessel's position is obtained and its course directed by landmarks, such
as headlands, mountains, church steeples, lighthouses and buoys.
Navigation in foggy weather by short range aids such as the radio
detection finder and radar, and by measuring the depth of the water, is
also called piloting. There are several very important instruments used
in piloting. The compass is used for steering and for taking bearings.
The radio detection finder is used to take bearings of radio beacons in
foggy weather. The fathometer measures the depth of the water. The
patent log is a device for measuring the speed of a ship through the

water. Radar with the P.P.I. gives a continuous position of the ship.

1. Which science is this story about? Navigation

2. When speaking of airplanes what navigation term is used? Air
navigation

3. What are used as landmarks for navigation? Headlands,

mountains, church steeples, and buoys.

4. Name one instrument used for navigation. Compass, radio
detection finder, fathometer, patten log, radar.

5. Is measuring the depth of the water also a form.of piloting?

Yes

APPENDIX B

N DIGITS CORRECT

N QUESTIONS
CORRECT

Group:
Subject
Gender:
Age(yrs):

49

APPENDIX B

NH
AH

21

 

l
65—

50— ..........

55-

50-

 

[j——-D NO Interference
I—I SPEECH Interference
------ Maximum Possible

 

 

a?

 

 

I - l
2 3 4 5 6

TEST BLOCK

 

O

 

Jem-

N DIGITS CORRECT

N QUESTIONS
CORRECT

50

APPENDIX B

Group: NH
Subject: AR
Gender: M

Age(yrs): 26

 

60-

55-

45—

40-

30

 

l l l I I I

CI---D NO Interference
I—I SPEECH Interference
................ MCXImum Possible

 

 

_\\l

 

rerun)
~wm
I1 I l

 

3
"II

 

I I I I - I
I 2 3 4 5 5

TEST BLOCK

‘1:—

IO

 

wllllll

N DIGITS CORRECT

N QUESTIONS
CORRECT

65

60

55

50

51

APPENDIX B

Group: NH

Subject: AT
Gender: F
Age(yrs): 20

 

 

| l l l l l

EI---l'_'l N0 Interference
I—I SPEECH Interference
................ Maximum POSSIbIe

 

 

*1

 

 

L l I I I
1 2 3 4 5 6

TEST BLOCK

 

KIIIIII

N DIGITS CORRECT

N QUESTIONS
CORRECT

52

APPENDIX B

 

 

 

 

 

 

 

Group: NH
Subject: .18
Gender: M
Age(yrs): 23
I r I T I j I I I I
65 — .—
D---D N0 Interference
I—I SPEECH Interference
60 — ......... Maximum Possible _.
55 — q
50 — _
45 — -—
40 — _
35 - .—
30 - _
25 - __
If I I I I l I l I I I K?
I I I
f I J I I I I I / {I

 

 

 

2 3 4 5 6

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

65

60

55

50

45

40

35

30

53

 

APPENDIX B
Group: NH
Subject: NT
Gender: M
Age(yrs): 20
T I I I I l I
EI---CI N0 Interference
I—I SPEECH Interference
— ................ Maximum POSSIbIe —
— —I
P

 

 

 

\VL

 

 

TEST BLOCK

 

Nllllll

N DIGITS CORRECT

N QUESTIONS
CORRECT

Group: NH
Subject: LW
Gender: F
Age(yrs): 27
I I I I I I I I I
65 r- .—
EI---EI N0 Interference
I—I SPEECH Interference
60 — ................ Maximum Possible .—
55 — -
50 - .—
45 — .—
40 - —
35 -‘- _
30 — _
25 F .4
I)? I I I I I I I I I I)?
I I I I I I I I I
25 - ..
23 - ..
21 - ..
19 - ..
l7 - -
15 ,- .I
I I I I I I- I I I I I
I 2 3 4 5 6 7 9 10

54

APPENDIX B

 

 

 

 

 

 

 

 

 

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

65

60

55

50

45

40

55

APPENDIX B

Group: NH
Subject: TW
Gender: M
Age(yrs): 29

 

EI---CI NO Interference
I—I SPEECH Interference
oooooooooooooooo MOXimum POSSib'e

 

 

 

*I

 

 

15-
IL I I I I I I

I 2 3 4 5 6

TEST BLOCK

 

KVLIIIII

N DIGITS CORRECT

N QUESTIONS
CORRECT

56

 

APPENDIX B
Group: AL LPTA(dBHL): 50* HAStotus: MON(R)
Subject: YA RPTAIdBHL): 45 HAYrs: 10
Gender: F SRS(%): 80 HAUse: CONSIST.
Age(yrs): 52
I I I I I I I I I
55 — _
Cl---EI NO Interference
I—I SPEECH Interference
60 - ................ Maximum Possible —
55 — ............................................................. _
50 — -
45 - R p —
40 q I \\ // a
" /
/ \ /

 

 

 

 

 

I I I
I 2 3 4 5 6

TEST BLOCK

 

 

N DIGITS CORRECT

N QUESTIONS
CORRECT

65

60

Group: AL

Subject: CB
Gender: F
Age(yrs): 51

57

APPENDIX B

LPTA(dBHL): 3O
RPTA(dBHL): 40*
SRS(%): 100

HAStotus: N/ A
HAYrs: O
HAUse: NONE

 

 

I I I I I I I I I

D---EI NO Interference

I—I SPEECH Interference

................ MCXImum Possible —
a’r’

 

 

 

 

IIIII

 

a?"

3 4 5 6
TEST BLOCK

 

K\I_IIIII

N DIGITS CORRECT

N QUESTIONS
CORRECT

58

APPENDIX B

Group: AL LPTA(dBHL): 50* HAStotus: BIN
Subject: DD RPTA(dBHL): 50 HAYrs: I .5
Gender: M SR8“): 84 HAUse: OCCAS.

Age(yrs): 65

I I I I I I I I I I

65h _

EI---C| NO Interference
H SPEECH Interference
60 - ................ Maximum Possible

 

50—

45-

4o—

35

30—

 

 

 

 

 

 

 

KIIIIII

L I I I I I
I 2 3 4 5 5 7 8 9 10

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

65

60

55

50

45

59

 

 

 

 

 

 

 

APPENDIX B
Group: AL LPTA(dBHL): 31 HAStotus: MON(L)
Subject: EE RPTAIdBHL): 34* HAYrs: 3
Gender: F SRSIZ): 92 HAUse: OCCAIS.
Age(yrs): 66
I I I I I I I I I I
D---EJ NO Interference
I—I SPEECH Interference
— ................ Maximum Possible ‘
i
”r’
I
I
f

 

 

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

Group: AL LPTA(dBHL): 33* HAStotus: BIN
Subject: TG RPTA(dBHL): 33 HAYrs: NEW(3 WKS)
Gender: M SRS(%): 92 HAUse: CONSIST.
Age(yrs): 52
I I I I I I I I I
65 — —
EI---EI N0 Interference
I—I SPEECH Interference
60 _ ................ Maximum Possib'e _'
55 — ............................................................. ._I
50 — —
45 - 4
40 - A
35 '— -!
30 r —
25 '- l 9
1 II I I I I I I I I I
I I I I I I I I r
25 - -I
23 - ..
21 - -
I9 - -
I7 - -
IS ,- -
1 I\ I I I I I I I j
I 2 3 4 5 6 7 8 IO

60

APPENDIX B

 

 

 

 

 

 

 

 

 

 

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

Group: AL

Subject: DS
Gender: F
Age(yrs): 42

61

APPENDIX B

LPTA(dBHL): 42 HAStotus: BIN

RPTA(dBHL): 42* HAYrs: 0
SRS(%): 100 HAUse: CONSIST.

 

65-

60—

50-

45—

40—

w
U'I
I

 

I I I

EI---EI NO Interference
I—I SPEECH Interference
oooooooooooooooo MOXimum POSSEbIe

 

 

_\\l

 

 

I
Ej

 

TEST BLOCK

N DIGITS CORRECT

N QUESTIONS
CORRECT

62

APPENDIX B

 

 

 
 
    

 

 

 

 

 

 

GroUp: AL LPTA(dBHL): 53 HAStotus: BIN
Subject: GS RPTA(dBHL): 40*
Gender: M SRS(%): 84 CONSIST.
Age(yrs): 60
I I I I I I I
65 —' —
EI---EI N0 Interference
I—I SPEECH Interference
60 — ................ Maximum Possible —
55 _ .......................... _
50 — —
45 F- —
4o — —-
35 -- _
30 — .—
25 — _)
I I I I I I I I 1
I I I I I I I
25 - ..
23 ..
19 - ..
17 - -
15 ..
i I I I I I I I j
I 2 3 4 5 6 10

TEST BLOCK

APPENDIX C

63

APPENDIX C

DIGIT RECALL

************************

SID: Date:

 

 

Block: Condition:

 

 

Comments:

 

ITEM

ITEM

ITEM

ITEM

ITEM

#1

#2

#3

*4

#5

(1)
(2)
(3)
(4)
(5)
(1)
(2)
(3)
(4)
(5)
(1)
(2)
(3)
I4)
(5)
(1)
(2)
(3)
(4)
(5)
I1)
(2)
(3)
(4)
I5)

64

APPENDIX C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BIBLIOGRAPHY

BIBLIOGRAPHY

American National Standards Institute. (1969). American National
Standard Specifications for Audiometers. ANSI 53.6-1969, New'York.

Anderson, C.M.B. and Craik, F.I.M. (1974). "The effect of concurrent
task on recall from primary memory," J. Verbal Learning and Verbal
Behav., 13, 107-113.

Baddeley, A. (1992). "WOrking Memory," Science. 255, 556-559.

Boothroyd, A. (1984). ”Auditory perception of speech contrasts by
subjects with sensorineural hearing loss, " J. Speech Hearing Res.
27, 134-144.

Broadbent, D.E. (1958). Perception and Communication. Pergamon Press,
New York.

Committee on Hearing, Bioacoustics, and Biomechanics (CHABA).(1988).
"Speech understanding and aging," J. Acoust. Soc Am, 83, 859-893.

Committee on Hearing, Bioacoustics, and Biomechanics (CHABA).(1991).
”Speech perception aids for hearing-impaired people: Current status
and needed research," J. Acoust. Soc. Am. 90(2), 637-683.

Gagne, J.P. (1988). "Excess masking among listeners with.a sensorineural
hearing loss,” J. Acoust. Soc. Am: 83(6), 2311-2321.

Hanneley, M., and Dormann, M.F. (1983). "Susceptibility to intraspeech
spread of masking in listeners with sensorineural hearing loss,“
J. Acoust. Soc. Am. 74(1), 40-51.

Humes, L.E., Nelson, K.J., and Pisoni, D.B. (1991). ”Recognition of
synthetic speech by hearing impaired elderly listeners," J. Speech
Hearing Res. 34, 1180-1184.

Logan, 6.6. (1979). "On the use of concurrent memory load to measure
attention and automaticity,” J. Exper. Psych.: Human Percep. and
Perform. 5(2), 189-207.

Logan, 6.6. (1985). Skill and.automaticity: Relations, implications, and
future directions," Canandian J. Psych. 39(2), 367-386.

Luce, P.A., Feustel, T.C., and Pisoni, D.B. (1983). "Capacity demands in
short-term memory for synthetic and natural speech, " Human Factors
25(1), 17-32.

Meyers, J.L. (1972). Fundamentals of Experimental Design. Allyn and
Bacon, Inc., Boston.

65

66

Moore, B. (1982). Introduction to the Psychology of Hearing. Academic
Press Inc., London.

Otto, W.C., and McCandless, G.A. (1982). "Aging and auditory site of
lesion," Bar and Hearing: 3(3), 110-117.

Pisoni, D.B. and.Greene, 3.6. (1990). "The role of cognitive factors in
the perception of synthetic speech, " Research on Speech Perception
Progress Report. (Nb. 16), 193-216.

Scharf, B. and Florentine, M. (1982). 'Psychoacoustics of elementary
sounds,” in The Vanderbilt Hearing Aid Report, edited by G.A.
Studebaker and F.H. Bess (Monographs in Contemporary Audiology,
Upper Darby, PA), pp. 203-208.

Schneider, W. and Shiffrin, R. M. (1977). ”Controlled and automatic
human information processing: 1. Detection, search, and
attention," Pysch. Rev. 84(1), 1-66.

Shiffrin, R.M., and Schneider, W. (1977). “Controlled and automatic
human information processing: II. Perceptual learning, automatic
attending, and a general theory,“ Psych.Rev. 84, 127-190.

Smith, A.D. (1980). ”Age differences in encoding, storage, and
retrieval,” in New Directions in Memory and Aging, edited by L.W.
Poon, J.L. Fozard, L.S. Cermak, D. Arenberg, and L.W. Thompson
(Erlbaum, Hillsdale, NJ).

Stroop, J.R. (1935). “Studies of interface in serial verbal reactions,”
J. Exper. Psych. 18, 643-62.

Summers, W.V., and Leek, M.R. (1992). "The role of spectral and
temporal cues in vowel identification by listeners with impaired
hearing,” J. Speech Hearing Res. 35, 1189-1199.

van Rooij, J.C., and Plomp, R. (1991). "Auditive and cognitive
factors in speech perception by elderly listeners. III. Additional
data and final discussion,” J. Acoust. Soc. Am” 91(2), 1028-1033.

Ill H II

HICHIG N STATE UNIV.
IWII IIIIIII‘
3129300910