Download Verbal Information-processing Capabilities and

Verbal Information-processing Capabilities and Cochlear
Implants: Implications for Preoperative Predictors of
Speech Understanding
Bjorn Lyxell
Jan Andersson
Stig Arlinger
Goran Bredberg
Henrik Harder
Jerker Ronnberg
Linkoping University, Linkoping, Sweden
In this study, we examined preoperative verbal cognitive capacity in 11 deafened adults who were cochlear implant candidates and reexamined level of speech understanding after
6—8 months' experience with the implant. Verbal cognitive
performance in the implant group was compared in a group
of normal hearing subjects and in a nonimplanted group of
deafened adults. The three groups performed on par with
each other with one exception: The individuals in the cochlear implant group and the nonimplanted group of deafened adults performed significantly worse than those of normal hearing in tasks in which use of internal speech is a key
feature (i.e., rhyme judgement and lexical decision tasks).
Postoperative observations of the implanted individuals' level
of speech understanding suggest that it is possible to predict
the level of speech understanding by means of a preoperative
cognitive assessment. The characteristics of three verbal cognitive abilities proved to be critical indicators of 6—8 months'
postoperative outcome: internal speech functioning, speed of
verbal information processing, and working memory capacity—the first factor proved the most decisive. We discuss the
results with respect to direct versus indirect predictors of
outcomes from cochlear implant operations and the effect of
auditory deprivation on deafened adults' capability to process
auditory information.
A cochlear implantation gives deafened individuals an
opportunity to experience some sense of hearing. How-
Thij research is supported by grants from the Swedish Council for Social
Research (93-0116) awarded to Bjorn LyxeH We thank UUa-Britt Persson for checking the language. Parts of thii research were presented at
the Eighth Conference of the European Society for Cognitive Psychology, Rome, Italy, September, 1995. Correspondence should be sent to
Bjarn Lyxell, Department of Education and Psychology, LinkSping University, S-S81 83 Linkoping, Sweden (e-mail: [email protected]).
Copyright C 19% Oxford University Press. CCC 1081-4959
ever, the outcome of such implantation varies widely
among the recipients. Some individuals can communicate over the telephone and easily follow and understand a speaker who is out of sight, whereas others can
only differentiate between the presence and absence of
sound (Knutson, Hinrichs, Tyler, Gantz, Schartz, &
Woodworth, 1991; Osberger, Todd, Berry, Robbins, &
Miyamoto, 1991). It is important, for a number of reasons, to create means that allow for preoperative predictions of the outcomes. Such predictions could, for
example, give candidates an opportunity to put their
expectations about the outcome in realistic perspective.
It can also help to inform teams working with implant
patients about rehabilitative efforts most fruitful for a
given patient (cf., Knutson, et al., 1991; Summerfield &
Marshall, 1994).
Listening with normal hearing in normal conditions is a quite effortless and highly automatized information-processing task. However, any kind of distortion (e.g., from noise or hearing impairment) of the
spoken signal, places extra demands on the information-processing system. For example, variations in performance on visual speechreading tasks can partly be
explained by the individual's perceptual and cognitive
capability to process incomplete verbal information
(Gailey, 1987; LyxeU, 1994; Ronnberg, 1995). That is,
in situations in which the spoken signal is poorly specified, as in visual speechreading, the individual must
compensate for missing information by explicit use of
alternative verbal information processing components
(Lyxell, 1994; Ronnberg, 1995).
Cochlear Implants and Information Processing
The purpose of this study is to examine possible
predictors for successful speech understanding following a cochlear implant operation. As in speechreading
and other kinds of communicative situations in which
the stimulus signal is poor, it is reasonable to assume
that speech perception with cochlear implants would
tax the individual's information-processing system more than normal listening conditions. Thus, it
should, in principle, be possible to predict some part
of the variation in speech understanding by means of
examining the individual's cognitive capability to process verbal information. The reason for this relatively
straightforward assumption is that a cochlear implant
can never fully simulate the functioning of a normal
ear. This suggestion is supported by clinical and experimental evidence demonstrating that relatively few implanted individuals reach a level of speech understanding comparable to that of normal hearing individuals.
Clinical observations and postoperative follow-up
studies indicate that two factors can predict at least
some part of the variation in speech understanding: (1)
duration of deafness between onset and the implant
operation and (2) preoperative speechreading skill.
These results indicate that a shorter, as opposed to
longer, duration of deafness might result in a greater
number of surviving auditory nerve fibers and more
complete memory for sounds (Gantz, Woodworth, Abbas, Knutson, & Tyler, 1993; Summerfield & Marshall,
1994). Preoperative skill in visual speechreading is assumed to reflect a verbal cognitive capacity necessary
to transfer sensory data into meaningful units (Summerfield & Marshall, 1994). However, the interpretation of these observations must be regarded as rather
indirect since the inferences about the underlying capability derive from another test or observation.
In this study, we examine subjects' performance in
clusters of tasks that more directly assess verbal cognitive functioning and are assumed to be related to postoperative performance following a cochlear implant.
More precisely, we employ tests of working memory
capacity (Baddeley, 1990), speed of verbal information
processing (Hunt, 1985), internal speech functioning
(Baddeley & Wilson, 1985; Conrad, 1979), and visual
speechreading performance. The first three tasks have
previously proved to account for varying proportions
of the variance in visual, audio-visual, and tactually
supported speechreading (Lyxell, 1989).
191
Tests of working memory are included because
when part of the spoken signal is missing or ambiguous, the individual must temporarily store information
to be able to fill in information (by means of inferences). Exceptional working memory capacity is one
key feature in cases of high visual speechreading skill
(Lyxell, 1994; the case of SJ), audio-visual speech understanding (Ronnberg et al., 1995; the case of MJ),
and in tactually supported speechreading (Ronnberg,
1993; the case of GS). Also, Knutson et al. (1991), using a visual monitoring task in which short-term storage of information was a task demand, demonstrated
a significant relationship between auditory and audiovisual performance with the implant.
Verbal information-processing speed is the speed
with which lexical information can be accessed from
long-term memory (Hunt, 1985; Sternberg, 1985).
This speed accounts for the variability in a number of
tasks such as general verbal ability (Hunt, 1978; 1980;
1985), reading comprehension (Baddeley, Logie,
Nimmo-Smith, & Brereton, 1985), chronological
age (Salthouse, 1982), and, relevant in this context,
sentence-based speechreading (Lyxell & Ronnberg,
1992). The importance of a relatively fast access of verbal information in speechreading is evident, as the spoken stimuli are degraded and short-lived. A relatively
slow access to verbal information would create a "processing bottle-neck" and inhibit the allocation* of
resources to other necessary processes (cf. Lyxell &
Ronnberg, 1992).
Previous studies have demonstrated the involvement of an internal speech code in visual speechreading (Conrad, 1979; Lyxell, Ronnberg, & Samuelsson,
1994). Recent results suggest that postlingually deaf
adults perform worse than normal hearing controls in
tasks demanding usage of internal speech (Lyxell, Andersson, & Ronnberg, 1995; Lyxell et al., 1994). Problems encountered by adults in using an internal speech
code could be tied to (a) distortions in the transformation process (i.e., the internal representation of sounds
is intact), to (b) representational aspects (i.e., the transformation process is intact), or some possible combination of (a) and (b). However, Lyxell and his colleagues
have found that these results hold only for tasks that
require some explicit form of internal speech (e.g.,
rhyme judgment tasks), and not when the requirements on internal speech are less explicit (e.g., name
192
Journal of Deaf Studies and Deaf Education 1:3 Summer 1996
matching, Posner & Mitchell, 1967; or semantic decision making, Shoben, 1982). Thus, the data suggest
that the deterioration of deafened adults' internal
speech functioning is not complete, but occurs under
specific conditions. Further examination of the performance of the deafened adults suggests that the problem
is localized at the representational level, rather than the
transformational. Performance for deafened adults in
rhyme judgment tasks is further correlated with
speechreading performance and also with the number
of years that an individual has been deaf. Wellfunctioning internal speech should, at least intuitively,
be important for individuals receiving cochlear implants. That is, a cochlear implant gives the individual
some sense of hearing, and it is important that auditory
impressions can be matched against existing internal
auditory representations. Clinical observations support
this assumption, as one determinant of postoperative
level of speech understanding is the time elapsed between the onset of deafness and the point when the individual receives an implant (Tyler, in press).
In this study we focus on the type of spoken communication that the individual can manage after six to
eight months of experience with a cochlear implant.
The subjects were classified in one of four categories
with respect to their functional communicative ability:
the lowest level of this scale was awareness of environmental acoustical stimuli; the next was improvement in
speechreading with the implant; next was the ability to
understand speech when the speaker is out of sight;
and, finally, the highest level was the ability to understand speech transmitted by the telephone. Data also
include clinical observations from members of cochlear
implant teams and self-reports from the implant patients themselves.
Method
Subjects. Eleven cochlear implant candidates (mean
age = 51.6 years, standard deviation [SD] = 10,
range = 28-70; see Table 4 for data on each individual)
participated in our study. Seven of the 11 individuals
received their cochlear implantation at the University
hospital in Linkoping, Sweden, and four at Sodersjukhuset, Stockholm. They were given the cognitive tests
when they were candidates for implantation and visited
the clinics for medical examination. Ten of the eleven
Table 1 Hearing threshold levels determined by sound
field testing with cochlear implants for each subject
Individuals
1
2
3
4
5
6
7
8
9
10
11
250 Hz
500 Hz
1KHz
2KHz
4KHz
30
30
35
25
35
45
40
55
30
30
25
30
20
30
40
40
45
35
45
35
40
30
30
30
30
25
30
30
30
35
30
45
35
40
35
30
25
35
30
30
40
30
45
45
55
50
55
50
35
40
35
40
35
35
35
implant candidates received the Nucleus 22-channel
implant and one the Ineraid implant. Their mean duration of deafness was 13.1 years (SD = 14.8, range =
1-51; see Table 4 for data for each individual). Their
postoperative hearing threshold levels 6—8 months
after the implantation were determined by sound field
testing with the cochlear implant, using warble tones
and calibrating in terms of dB HL with data according
to ISO 226 (1987) as reference. The mean hearing
threshold levels were for 250 Hz 39 dB (SD = 10.7,
range = 25-55), for 500 Hz 34 dB (SD = 8.0, range =
20-45), for 1 kHz 35 dB (SD = 3.9, range = 30-40),
for 2 kHz 33 dB (SD = 5.6, range = 25-45) and, finally, for 4 kHz 37 dB (SD = 9.8, range = 25-55).
Using the average sound field hearing threshold levels
as a basis, the application of simple Articulation Index
(Killion, Mueller, Pavlovic, & Humes, 1993) calculations would estimate a perceptual speech intelligibility
on the order of 4O%-50%. The Articulation Index can
be considered a measure of the proportion of speech
cues that are audible and thus theoretically available for
the implantees. Threshold values for each implant patient are displayed in Table 1.
Ten nonimplanted deafened adults (mean age =
57.4 years, SD = 8, range = 22-67) and 10 normal
hearing individuals (mean age = 56.8, SD = 8,
range = 22-70) participated in our study on a voluntary basis. None of the nonimplanted deafened adults
had any functional residual hearing with hearing aids
according to their most recent medical records. A test
of vocabulary size (antonyms; F-testet, Psykologi forlaget, Stockholm) was given to match the subjects in the
Cochlcar Implants and Information Processing
three groups on one index of verbal ability. No significant difference among the groups emerged (/> > .05).
Tests. The tests used in this study have previously been
used to examine speechreading and tactile-supported
speechreading under conditions similar to those examined here. Therefore, a condensed description of the
tests will be presented here, and readers can consult
Lyxell (1994), Ronnberg, Arlinger, Lyxell, & Kinnefors
(1989), or Ronnberg (1993) for more detailed descriptions.
1. Sentence-based speechreading test. To measure
subjects' sentence-based speechreading ability, 24
different sentences-to-be-speechread were presented
(i.e., same test as in Ronnberg, et al., 1989; Lyxell &
Ronnberg, 1989, 1992; Ronnberg, 1990, 1993). The 24
sentences were divided into three blocks, each including eight sentences. All eight sentences in a given block
were presented together before any sentence from another block was presented. Prior to the presentation of
the sentences in each of the three blocks, the subjects
were instructed to read a context-creating script,
which was printed on an answer-sheet and informed
the subjects about a scenario in which a hypothetical
conversation took place. Three different scenarios were
used (i.e., a "train scenario"; a "restaurant scenario";
and a "shop scenario"), each with a unique framehistory. In half of the sentences, the subjects were
offered an extra contextual cue, which informed them
about the semantic content of the sentence-to-bespeechread.
A JVC VHS videotape recorder was used to present the material on a 66-cm Tandberg TV-monitor in
the following way. First, the videoscreen was red for a
period of five seconds; following this, a male actor (JA)
appeared on the videoscreen. The actor was silent for
a period of three seconds, before and after presentation
of a sentence. After the presentation, the response interval started (25 seconds), during which the TVscreen was gray. This procedure,was repeated for all
sentences in the test. The subjects were placed at a distance of three meters from the screen during the test
session and were instructed to write all words that they
had been able to speechread and to guess the rest of the
message (cf, Lyxell & Ronnberg, 1991, 1992).
Scoring of the responses was carried out with respect to both correctness of the word and their correct
193
position in the sentence. The number of correctly recalled words per sentence was expressed as a proportion for each sentence and subject for the purpose of
data analysis.
Short-term /working memory tests.
1. Reading span test. The subjects' task was to
comprehend sentences and to recall either the first or
the final words of a presented sequence of sentences in
correct serial order (cf., Baddeley et al., 1985). The
words were presented in a word-by-word fashion, at a
rate of one word per .80 sec with an interstimulus interval of .07 sec Half of the sentences were absurd (e.g.,
"The train sang a song"), and half were normal sentences (e.g., "The girl brushed her teeth"). The subjects'
task was to respond "yes" (for a normal sentence) or
"no" (for an absurd sentence) during a 1.75 sec interval. After a sequence of sentences (3-6 sentences), the
experimenter pointed at the words "First" or "Final"
written on a piece of paper, indicating that the subjects
should start to recall either the first or the final words
of each presented sentence in the sequence in their correct serial order. The order ("First" or "Final") was
randomized.
2. Word span. The subjects' task was to recall in
correct serial order a string of words, which were presented one at a time on the computer screen. The first
span size employed (after a practice session) was three
words, the next was four and so on up to eight words.
Tests ofverbal information processing speed.
1. Physical matching. The subjects' task was to
judge whether two simultaneously presented letters
had the same physical shape (e.g., A-A) or not (A-a).
The subjects were to respond by pressing predefined
buttons for "yes" and "no" answers. Half of the pairs
were identical and half were not.
2. Name matching. Instead of matching letterpairs for their physical identity, the subjects were asked
to match letter-pairs for their name. Sixty-four responses were collected in four trials of 16 responses.
The same letter-pairs were used as in the test of physical matching.
3. Lexical decision-making. The subjects' task was
to judge whether a string of three letters constituted a
real word or not. One hundred items were used in the
194 Journal of Deaf Studies and Deaf Education 1:3 Summer 1996
test, half being real words and half not. Twenty-five
items that were not a real word sounded like a real word
when pronounced (i.e., homophone), whereas the remaining 25 did not sound like real words (i.e., nonhomophones). The real words used in the present test
were all, according to Allen (1970), familiar Swedish
words.
4. Semantic decision-making. The subjects' task
was to decide whether a word belonged to a certain
predefined semantic category or not (cf., Shoben,
1982). Four trials were used with 24 to-be-categorized
items, of which 12 items belonged to a semantic category and 12 items were foils. The four trials were "colors," "occupations," "diseases," and "parts of the body."
For all long-term memory access tests, latency data
were based on the average of each subject's yes/no responses.
Rhyme judgment test of internal speech. The subjects were
given a rhyme judgment test in which they had to decide whether two simultaneously presented words
rhymed. The subjects were to respond by pressing predefined buttons for "yes" and "no" answers. Half of the
pairs rhymed and half did not.
Four experimental conditions were created in the
rhyme judgment test (cf., Lyxell et al., 1994). The first
condition involved word-pairs which rhymed and were
orthographically similar in spelling (e.g., KATTHATT), the second by words that rhymed, but were
orthographically dissimilar in spelling (e.g., DAGSLAX). In the third condition the words did not rhyme,
but were orthographically similar in spelling (e.g.,
STAL-STEL); in the final condition the words did not
rhyme and were orthographically dissimilar in spelling
(e.g., CYKEL-SPADE).
Results
The results will be presented in two parts. The first
part examines the performance of the implanted group
of deafened adults, the nonimplanted deafened adults,
and normal hearing adults on the verbal cognitive
tasks. The second examines the pattern of verbal cognitive performance of the 11 implanted patients and relates performance to the individuals' levels of speech
understanding with their implants.
Preoperative level of verbal cognitive performance. Tables 2
and 3 give descriptive data for the three groups in this
study. Performance on tasks assessing working memory/short-term memory is comparable for all three
groups and replicates previous results, both with respect to mean performance and to performance level
on tests, where groups of deafened adults and normal
hearings have been compared (Lyxell & Ronnberg,
1989; Ronnberg, 1990). A similar pattern of results
emerges when verbal information processing speed and
accuracy level are examined: no differences (with one
exception) between the groups in any of the tests (or
test conditions) occurred. The performance levels
closely resemble previously reported levels when similar (cf, Hunt, 1980) or the same tests have been used
(Lyxell, 1994; Lyxell & Ronnberg, 1992; Ronnberg,
1990). In one condition of the lexical decision test, the
deafened adults (i.e., both the implanted and the nonimplanted individuals) proved to have a significantly
lower level of accuracy (t(18) = 2.95, p < .01, and
t(19) = 2.16 , p < .05) than the normal hearing subjects: when the string of letters did not constitute a
word but when pronounced sounded like a real word
(i.e., pseudohomophones).
In the rhyme judgment test, no differences between
the groups occurred when speed of judgment was examined, which is in line with previous results (although
tested here with slightly different rhyme judgment
tasks; cf. Lyxell et al., 1994). The accuracy level was
significantly lower for the two groups of deafened
adults in two conditions: when the words in the wordpair rhymed and were orthographically dissimilar
(t(18) = 2.35,/) < .05, t(19) = 2.33,/> < .05), and when
the words did not rhyme but were orthographically similar (t( 18) = 2.15,/) < .05, t(19) = 2.15,/) < .05). No
differences occurred in the two conditions wherein
words rhymed and were orthographically similar or did
not rhyme and were orthographically dissimilar (ps >
.05).
Taken together, the results at the group level can
be summarized by two points: First, the two groups of
deafened adults performed comparably in all cognitive
tasks used in this study, suggesting that the group of
implant patients does not differ from other populations
of deafened adults. Second, similar to other studies
(Cowie & Douglas-Cowie, 1993; Lyxell & Ronnberg,
Cochlear Implants and Information Processing
Table 2 Mean performance for the three groups on the tests used in the study
Deafened adults,
Cochlear implants
Speechreading (proportion
correct)
Tests of short-term/working
memory (proportion
correct)
Reading span
Word span
Deafened adults,
nonimplanted
Normal hearing
.48
.38
.34
.38
.53
.38
.46
.39
.48
Tests of informationprocessing speed (in
seconds)
Physical matching
Name matching
Semantic decision making
Lexical decision making
Yes
No: homophones
No: nonhomophones
.85
.87
.73
.82
.89
.75
.73
.83
.69
.78
1.03
.73
.80
1.07
.72
.74
.92
.69
Rhyme-judgement: wordpairs (in seconds)
Yes: visually similar
Yes: visually dissimilar
No: visually similar
No: visually dissimilar
1.15
1.71
1.62
1.36
1.17
1.84
1.86
1.28
'
1.13
1.65
1.59
1.20
Table 3 Accuracy level for the tests of verbal information-processing speed and rhymejudgement for the three groups
Deafened adults,
Cochlear implants
Deafened adults,
nonimplanted
Normal hearing
Test of informationprocessing speed
Physical matching
Name matching
Semantic decision making
Lexical decision making
Yes
No: homophones
No: nonhomophones
.89
.92
.96
.96
.96
.98
.98
.97
.98
.93
.80
.93
.94
.82
.96
.94
.93*
.96
Rhyme-judgement wordpairs
Yes: visually similar
Yes: visually dissimilar
No: visually similar
No: visually dissimilar
.98
.65
.63
.89
.98
.67
.59
.96
.97
.92*
.87*
1.00
*p < .05.
195
196
Journal of Deaf Studies and Deaf Education 1:3 Summer 1996
Table 4
Age, duration of deafness, and mean performance for the cochlear implant patients on the tests used in the study
Speech reception
performance
Telephone
conversation
Individuals
2
3
Age
Duration of deafness
28
4
39
4
Speechreading (proportion
correct)
Tests of short-term working
memory (proportion
correct)
Reading span
Word span
Test of informationprocessing speed (in
seconds)
Physical matching
Name matching
Semantic decision making
Lexical decision making
Yes
No: homophones
No: nonhomophones
Rhyme-judgement: wordpairs (in seconds)
Yes: visually similar
Yes: visually dissimilar
No: visually similar
No: visually dissimilar
Listening
Speechreading improvement
Sound awareness
8
1
10
4
5
6
7
9
11
58
10
70
10
61
12
29
1
70
10
58
5
57
51
42
1
59
27
.65
.68
.68
.27
.69
.36
.04
.22
.77
.80
.25
.52
.58
.56
.52
.57
.74
.28
.81
.44
.52
.41
.54
.19
.63
.06
.19
.33
.47
.39
.60
.17
.31
.51
.57
.53
.70
.70
.50
.83
.89
.83
.87
.88
.74
.86
.87
.74
.65
.74
.59
1.65
1.12
.76
1.26
1.38
1.19
.71
.78
.81
.76
.87
.68
.68
.78
.72
.59
.68
.55
.61
.78
.55
.95
1.03
.76
.70
1.16
.81
.78
1.0
.77
.64
.99
.76
.90
.65
1.17
1.45
1.15
.85
1.27
.85
.80
1.13
.73
.78
1.03
.73
.88
1.22
1.45
.99
.96
1.36
1.42
1.21
1.17
1.65
1.19
1.22
1.01
1.56
1.27
1.02
1.40
2.58
2.03
1.40
1.18
1.13
1.26
1.85
1.94
1.30
1.98
1.62
1.60
1.46
2.32
1.82
2.04
1.06
1.20
1.23
1.10
1.16
2.00
2.26
1.11
1991; Mogford, 1987), the subjects in all three groups
performed comparably in the speechreading task, although the range of performance was larger in both
groups of deafened adults than in the group of normal
hearing controls. This indicates that these two groups
include both proficient and less proficient speechreaders (see Table 4, for the group of implanted deaf
adults). A similar pattern of results with respect to the
accuracy level in the deafened adults' performance on
the rhyme judgment task has recently been observed in
our own laboratory (Lyxell & Ronnberg, 1994; Lyxell
et al., 1995; Lyxell et al., 1994; all three studies with
somewhat larger number of subjects). These results are
also in line with results from other studies in which internal phonological abilities of other groups of deaf
(i.e., prelingually deaf children and adults) have been
examined (although with a different methodology than
.55
1.71
1.76
1.34
in this study; Alegria, Leybaert, Charlier, & Hage,
1992; Hanson & McCarr, 1989). The novel finding
from this study is that we can extend our observation
of a deterioration in internal speech functioning to
tasks other than rhyme judgment tasks (i.e., lexical decision tasks).
Relation between preoperative measures and postoperative
speech understanding. Postoperative clinical observations
by the members of the implant teams and self-reports
reveal individual differences in type and quality of gain
in speech understanding with the implant. This is by
no means a new finding; rather, it is observed each time
a population of implanted individuals has been examined (Knutson et al., 1991; Osberger et al., 1991). The
subjects were classified into one of four categories on
an ordinal scale with respect to their functional com-
Cochlear Implants and Information Processing
197
implant in a functional communicative sense. Finally,
one subject (i.e., 11) has only sound awareness and
some ability to identify enviromental sounds.
Further examination of the individuals' verbal cognitive capacity reveals a correspondence between their
preoperative verbal cognitive performance and postoperative level of speech understanding. The three subjects who can communicate by means of telephone can
be characterized as skilled speechreaders (i.e., they
outperform the controls by approximately a factor of
two); they have a relatively large working memory capacity (i.e., they outperform the normal hearing controls by approximately a factor of 1.5). Furthermore,
when verbal information-processing speed and speed
of performance in the rhyme judgment tasks are considered, their performance is faster (subjects 2 and 3)
or comparable to the controls (subject 8). All three subjects demonstrate an accuracy level somewhat higher
than that for the control group on the verbal information-processing tasks. They perform on par with the
normal hearing control for the rhyme judgment tasks.
municative ability. The lowest level on this scale was
awareness of environmental acoustical stimuli; the next
highest was improvement in speechreading with the
implant; next was the ability to understand speech
when the speaker is out of sight; and the highest was
the ability to understand speech transmitted by the
telephone.
The subjects were classified in the four categories
as follows: Three individuals (i.e., the individuals 2, 3,
and 8) can communicate quite effortlessly with other
individuals by means of telephone. More specifically,
two of the three report no problem in telephone use
even if the speaker and the topic of the conversation
are unfamiliar to them; one can quite easily follow a
conversation if the speaker and the topic are familiar.
Two individuals (i.e., 1, 10) can without effort follow a
conversation when an unfamiliar speaker is out of sight
and when the topic is unfamiliar. These individuals can
sometimes understand a few words spoken over the
telephone, but they cannot manage a telephone conversation in a functional sense. Three (i.e., individuals 4,
5, 6) report that their speechreading is substantially
improved with the implant or less effortful (i.e., 7 and
9). Similar to the previous group, this group of subjects
can hear some words when the speaker is out of sight,
but the speaker must be visible before they can use the
The two subjects who can follow an oral communication when the speaker is out of sight are in one case
skilled in speechreading (i.e., 10) and in the other
slightly below the average (i.e., 1). As for the memory
tests, one subject is considerably below the average for
Table 5 Accuracy level for the tests of verbal information-processing speed and rhyme-judgement for each cochlear
implant patient
Telephone
Speech reception
performance
conversation
Listening
Speechreading improvement
Sound awareness
Individuals
Test of informationprocessing speed
Physical matching
Name matching
Semantic decision making
Lexical decision making
Yes
No: homophones
No: non-homophones
Rhyme-judgement: wordpairs
Yes: visually similar
Yes: visually dissimilar
No: visually similar
No: visually dissimilar
2
3
8
1
10
4
5
6
7
9
11
1.0
.94
1.0
.91
.86
.98
.94
1.0
.99
.75
.95
.96
.94
.97
1.0
.97
.97
1.0
.50
.88
.90
.94
.67
.91
.94
.91
.89
1.0
.94
.99
.94
.98
.95
.90
1.0
.96
.94
1.0
.92
.98
.96
.96
.96
.71
.92
.90
.96
.89
.92
.75
.96
.90
.65
.96
.98
.54
.92
.92
.54
.92
.92
.88
.96
.96
.83
.92
1.0
.92
1.0
1.0
1.0
.92
.38
1.0
1.0
1.0
1.0
.92
1.0
1.0
.92
1.0
1.0
.50
.77
1.0
1.0
.58
.23
1.0
.83
.42
.31
.15
1.0
.17
.63
1.0
1.0
.42
.15
.92
1.0
.92
1.0
1.0
1.0
.33
.54
1.0
198 Journal of Deaf Studies and Deaf Education 1:3 Summer 1996
the reading span task and above on the word-span task.
Thus, this individual possesses an excellent working
memory capacity when the task demand in the memory
task only emphasizes storage (i.e., word-span), whereas
his capacity to deal with tasks emphasizing both processing and storage (i.e., reading span) is poor. The
second individual within this category possesses a normal working memory capacity. Both perform on par
with the controls on the tasks assessing verbal information-processing speed. For speed of rhyme judgment,
subject 1 proves to be faster than the control group,
whereas subject 10 is considerably slower. The accuracy level for subject 1 is on par with the controls for
the tests of verbal information-processing speed and
rhyme judgement, with two exceptions: the physical
matching and one condition in lexical decision task.
Subject 10 has a similar rate of accuracy for the verbal
information-processing tasks as the control group, but
performs substantially lower on the rhyme judgment
task.
Three individuals (i.e., 4, 5, and 6) improved in
their speechreading substantially and two (i.e., 7 and 9)
did not improve in their speechreading but report that
their speechreading is less effortful with the implant.
The subjects who improved in their speechreading
performance (two cases) on average are somewhat below the control group or substantially below in one
case. As for the working memory span tasks and the
tests of verbal information-processing speed, one subject performs comparably to the controls, and two are
inferior to them. In the rhyme judgment task, all three
perform slower than the normal hearing controls. On
verbal information-processing tasks, the three subjects'
accuracy level is in one case on par with the controls
and in two somewhat below both controls. However,
when rhyme judgment is considered, the accuracy level
is clearly below that of normal hearing subjects.
Two subjects cannot follow a conversation when
the speaker is out of sight and did not increase their
speechreading performance with the implant. However, both are skilled speechreaders and, compared to
the other three within this category, there might be less
room for improvement. Subject 7 performs on par on
some of the verbal information-processing tasks and
less well on the rhyme judgment task. Parenthetically,
subject 7 is the only deafened adult that we have ob-
served who is a skilled speechreader without wellfunctioning internal speech. Subject 9 performs similarly to the normal hearing controls on most of the
tasks and should, at least in comparison with the other
subjects in this study, reach a higher level of speech understanding. However, further medical examination revealed that the absence of improvement in this case
may be due to a combination of technical and medical
problems. Finally, one subject can discriminate only
between conditions of sound and nonsound in the
acoustic environment. An examination of his verbal
cognitive profile reveals that his performance is below
or substantially below compared to normal hearing
subjects on most tasks in the study.
The results from this study can be summarized
based on the observed data-pattern: Implanted individuals who are able to communicate with others over
the telephone perform better or on par with the normal
hearing controls on verbal information-processing
measures. Individuals who can follow a conversation
when the speaker is out of sight perform at the same
level as the normal hearing subjects on most tasks with
one exception (in each of the two cases). Individuals
who improve only their speechreading or receive
acoustical orientation are inferior to the controls on
most tasks. Similar to previous studies, we can observe
that skill in speechreading is a feature among those implant patients who are most proficient in interpreting
spoken acoustic stimuli, but we can also observe skilled
speechreaders among those individuals who are less
proficient. This finding further suggests that in order
to preoperatively predict success with a cochlear implant, verbal cognitive capacity of the individual must
be properly understood as well.
Discussion
Our study was designed to examine the possibility that
the preoperative characteristics of some verbal cognitive components could be related to the level of speech
understanding six to eight months after operation.
Here, we can identify three verbal cognitive factors
that can serve as predictors of the level of speech
understanding following a cochlear implantation:
working memory capacity, speed of verbal informationprocessing, and internal speech. Our results suggest
Cochlear Implants and Information Processing
that the individual must possess a capacity that is at
least similar to or better than that of normal hearing
individuals in each of these three tasks in order to
understand speech without visual contact with the
speaker. If it is the case that the performance level in
one of the three components does not resemble that of
the normal hearing subjects, understanding of speech
without seeing the speaker is not possible after six to
eight months' experience with the implant.
The results are partly in line with results presented
by Summerfield and Marshall (1994), who demonstrated that better performance in speech understanding could be explained by (a) duration of deafness
(where shorter periods lead to better results), (b) possession of some usable residual hearing, and (c) preoperative skill in lipreading. Implications for the first two
factors above are reflected in this study by the fact that
relatively intact internal speech is found only among
individuals with the most successful outcome (i.e., capable of following a conversation when the speaker is
out of sight). Performance on tasks measuring phonological capabilities or internal speech is further negatively correlated with the number of years that the individual has been deaf (see also, Lyxell et al., 1994;
Conrad, 1979). Our data suggest that the role of preserved internal speech in cochlear implantees is that
the individual must possess the ability to match the experience of external sound with an existing internal
representation of sounds. Given that this possibility
does not exist, the individual will inevitably suffer from
difficulties in interpreting the sounds from the external
world. Other independent support for such a hypothesis comes from studies using PET-scan methodology,
which has demonstrated that the metabolic activity in
the primary auditory cortex declines progressively with
time as a function of auditory deprivation (Ito, Iwasaki,
Sakakibara, & Yonekura, 1993; Soderfeldt, 1994). A
question for further research is whether this deterioration is irreversible or whether a cochlear implantation
possibly can alter this state of affairs. An implication
for the latter possibility was demonstrated in a recent
study with six subjects (with varying experience of the
implant) in which their metabolic rate returned to an
almost normal level after receiving an implant (Naito et
al., 1995).
In clinical practice, duration of deafness is often
199
used as a predictor of the outcome. The rationale for
using duration of deafness as a predictor is that it is "a
reflection of number of surviving hair cells and neurones" (Tyler, in press) or the ability "to interpret auditory sensations" (Summerfield & Marshall, 1994).
These are indirect indicators of the functions of internal speech. Even if there is a correlation between duration of deafness and outcomes, however, the correlation
is far from perfect, and a large variability exists among
the individuals (Tyler, in press). Although the population in this study is relatively small, this variability is
to some extent present. The clinical suggestion from
this study is, thus, that a more appropriate approximation of the functioning of the individuals' internal
speech is achieved by testing the ability directly (e.g.,
by means of rhyme judgment tests).
Similar to previously reported results (Summerfield & Marshall, 1994; Tyler, in press), as well as
clinical observations, we found that better skill in visual
speechreading is characteristic of those subjects who
reach the highest levels of speech understanding. However, proficiency above the level in the control groups
is also observed in individuals who do not reach higher
levels in speech understanding. Visual speechreading
may constitute an ability that is necessary but not
sufficient for success with the implant. The reason for
this relation is that some of the verbal cognitive abilities
that are critical determinants for skill in speechreading
are also critical for the outcome of an implant operation, whereas other verbal cognitive abilities, also critical for speechreading performance, might be less critical for performance with the implant (e.g., visual
decoding skill; Gailey, 1987; Lyxell & Ronnberg, 1989).
Speed of verbal information processing and working
memory capacity (Lyxell, 1994; Ronnberg, 1990) are
two abilities critical for performance in both visual
speechreading and with implants. Common to individuals who can communicate with a speaker who is out
of sight is their performance above or on a par with that
of the control groups. This level of performance serves
as a performance threshold, indicating that the threshold must at least be reached before gains in speech understanding other than mere acoustic orientation are
achieved. However, even if the individuals possess a capacious working memory and are fast in accessing information from their long-term memory, speech un-
200 Journal of Deaf Studies and Deaf Education 1:3 Summer 1996
derstanding at levels above gain in speechreading is
still critically dependent on the quality of the individual's internal speech.
In sum, the results from this study indicate that it
might be possible to predict what level of speech understanding the individual will reach six to eight
months after cochlear implantation, by means of preoperative testing of verbal cognitive abilities. Measures
of duration of deafness and preoperative skill in lipreading are indirect predictors as they are assumed to
tap abilities such as memory for sounds and general
cognitive abilities (Summerfield & Marshall, 1994;
Tyler, in press). In clinical practice we can observe individuals with a relatively long duration of deafness
and with intact internal speech functions and, similarly, individuals with a high level of preoperative lipreading skill whose level of speech understanding is below the expected. The results from our study suggest
that higher prediction-accuracy is obtained if these
abilities are tested directly. However, any prediction
must take into account that, given the individual possesses a perfect verbal cognitive architecture, there
must also be optimal anatomical and technical conditions for a successful outcome of the implantation. In
this study, this is illustrated by subject 9, who possesses
almost the same verbal cognitive abilities as the individuals with the most successful outcomes. Further examination of her case suggests that her lack of improvement in speech understanding may be due to technical
problems with the implant, not anatomical problems
or, as indicated by her results, verbal cognitive difficulties.
Collecting preoperative data of cognitive functioning thus serves two purposes: to allow predictions
about the outcome with the implant and to provide
preoperative information that can allow determination
of possible problems with the implant due to the individual's capacity to process verbal spoken information
or other factors (i.e., medical or technical).
References
Alegria, J., Leybaert, J., Charlier, B., & Hage, C (1992). On the
origin of phonological representations in the deaf: Hearing
lips and hands. In J. Alegria, D. Holender, J. Morais, & M.
Radeau (Eds.), Analytic approaches to human cognition. Amsterdam: North-Holland.
Allen, S. (1970). Frequency dictionary ofpresent-day Swedish. (In
Swedish: Nusvensk frekvensordbok.). Stockholm: Almquist & Wiksell.
Baddeley, A. (1990). Human memory: Theory and practice. Hillsdale: LEA.
Baddeley, A., Logic, R., Nimmo-Smith, I., & Brereton, N.
(1985). Components of fluent reading. Journal of Memory
ami Language, 24, 119-131.
Baddeley, A., & Wilson, B. (1985). Phonological coding and
short-term memory in patients without speech. Journal of
Memory and Language, 24, 490-502.
Conrad R. (1979). The deafschoolchild. London: Harper & Row.
Cowie, R., & Douglas-Cowie, E. (1993). Postlingually acquired
deafness: Speech deterioration and the wider consequences. New
York: Mouton de Gruyter.
Gailey L. (1987). Psychological parameters of lipreading skill. In
B. Dodd, & R_ Campbell (Eds.), Hearing by eye. London:
LEA, pp. 115-141.
Gantz, B., Woodworth, G., Abbas, P., Knutson, J. & Tyler, R.
(1993). Multivariate predictors of audiological success with
multichannel cochlear implants. Annals of Otology, Rhinology and Laryngology, 102, 909-916.
Hanson, V, & McCarr, N. (1989). Rhyme generation by deaf
adults. Journal ofSpeech and Hearing Research, 32, 2-11.
Hunt, E. B. (1978). The mechanics of verbal ability. Psychological
Review, 8S, 109-130.
Hunt, E. B. (1980). Intelligence as an information-processing
concept. British Journal ofPsychology, 71, 449-474.
Hunt, E. B. (1985). Verbal ability. In R. J. Steinberg (Ed.), Human abilities: an information processing approach (pp. 31—58).
New York: Freeman and Company.
ISO 226. (1987). Acoustics—Normal equal loudness level contours. International Organization for Standardization. Geneva, Switzerland.
Ito, J., Iwasaki, Y, Sakakibara, J., & Yonekura, Y. (1993). Positron
Emission Topography of auditory sensations in deaf patients
and patients with cochlear implants. Annals of Otology, Rhinology and Laryngology, 102, 797-801.
KiUion, M. C, Mueller, H. G., Pavlovic, C V., & Humes, L. E.
(1993). A is for audibility. The Hearing Journal, 46(4), 29.
Knutson, J. E, Hinrichs, J. V, Tyler, R. S , Gantz, B. J., Shartz,
H. A., & Woodworth, G. (1991). Psychological predictors of
audiological outcomes of multichannel cochlear implants.
Annals of Otology, Rhmology and Laryngology, 100, 817-822.
Lyxell, B. (1989). Beyond lips: Components ofspeechreading skills.
PhD. thesis, University of Umea, Umei, Sweden.
Lyxell, B. (1994). Skilled speechreading: A single-case study.
Scandinavian Journal ofPsychology, 35, 212—219.
Lyxell, B., & Ronnberg, J. (1989). Information-processing skills
and speechreading, British Journal of Audiology, 23, 339347.
Lyxell, B., & Ronnberg, J. (1991). Visual speech processing: Relations to sentence-based speechreading and hearingimpairment. Scandinavian Journal ofPsychology, 32, 9-17.
Lyxell, B., and Ronnberg, J. (1992). Verbal ability and sentencebased speechreading. Scandinavian Audiology, 21, 67-72.
Cochlear Implants and Information Processing
Lyxell, B., Andersson, J., & Ronnberg, J. (1995). Internal speech
functioning and speechreading of a connected discourse.
Manuscript, Department of Education and Psychology,
Linkoping University, Linkoping, Sweden.
Lyxell B., Ronnberg, J., & Samuelsson, S. (1994). Internal
speech functioning and speechreading in deafened and normal hearing adults. Scandinavian Audiology, 23, 179-185.
Mogford, K. (1987). Lipreading in the prelingually deaf. In B.
Dodd, & R. CampbeU (Eds.), Hearing by eye (pp. 191-211).
London: LEA.
Naito, Y., Okazawa, H., Honjo, I., Hirano, S., Takahashi, H.,
Shiomi, Y., Hoji, W., Kawano, M., Ishizu, K., & Yonekura,
Y. (1995). Cortical activation with sound stimulation in cochlear implant users demonstrated by Positron Emission
Topography. Cognitive Brain Research, 2, 207—214.
Osberger, M. J., Todd, S. L., Berry, S. W., Robbins, A. M., &
Miyamoto, R. T. (1991). Effect of age at onset of deafness on
children's speech perception abilities with a cochlear implant. Annals of Otology, Rhinology and Laryngology, 100,
883-S88.
Posner, M. I., & Mitchell, R. F. (1967). Chronometric analysis of
classification. Psychological Review, 74, 392-409.
Ronnberg, J. (1990). Cognitive and communicative function:
The effects of chronological age and "handicap-age." European Journal of Cognitive Psychology, 2, 253—274.
Ronnberg, J. (1993). Cognitive characteristics of skilled tactiling:
The case of GS. European Journal of Cognitive Psychology,
5, 19-33.
201
Ronnberg, J. (1995). What makes a skilled speechreader? In G.
Plant, & K-E Spens (Eds.), Profound deafness and speech communication. London: Whurr.
Ronnberg, J., Andersson, J., Samuelsson, S., Soderfeldt, B.,
Lyxell, B., & Risberg, J. (1995). The bilingual enrichment
hypothesis: The case of MJ. Manuscript. Department of
Education and Psychology, Linkoping University, Link6ping, Sweden.
Ronnberg, J., Arlinger, S., Lyxell, B., & Kinnefors, C (1989).
Visual evoked potentials: Relation to adult speechreading
and cognitive function. Journal of Speech and Hearing Research, 32, 725-735.
Salthouse, T. (1982). Adult cognition. New York: Springer Verlag.
Shoben, E. (1982). Semantic and lexical decisions. In C R. Puff
(Ed.), Handbook of research methods in human memory and cognition. New York: Academic Press.
Soderfeldt, B. (1994). Signing in the brain. Ph.D. thesis, University of Uppsala, Uppsala, Sweden.
Sternberg, R. J. (1985). Human abilities: An informationprocessing approach. New York: Freeman & Co.
Summerfield, Q., & Marshall, D. H. (1994). Pre-operative predictors of outcomes from cochlear implantation in adults:
Performance and quality of life. Paper presented at the International Cochlear Implant, Speech and Hearing Symposium; Melbourne, Australia.
Tyler, R. S. (in press). Deprivation, maturation, and acclimatization with cochlear implants. Ear and Hearing, supplement.