Brain and Language 73, 442–455 (2000)
doi:10.1006/brln.2000.2319, available online at http://www.idealibrary.com on

Perception of Dynamic Acoustic Patterns by an Individual
with Unilateral Verbal Auditory Agnosia
Emily Wang,* Richard K. Peach,*,† Yi Xu,‡ Michael Schneck,†
and Charles Manry II*
*Department of Communication Disorders and Sciences, Rush University and RushPresbyterian-St. Luke’s Medical Center, Chicago, Illinois; †Department of
Neurological Sciences, Rush University and Rush-Presbyterian-St. Luke’s
Medical Center, Chicago, Illinois; and ‡Department of Communication
Sciences and Disorders, Northwestern University, Evanston, Illinois
Previous studies have found that subjects diagnosed with verbal auditory agnosia
(VAA) from bilateral brain lesions may experience difficulties at the prephonemic
level of acoustic processing. In this case study, we administered a series of speech
and nonspeech discrimination tests to an individual with unilateral VAA as a result
of left-temporal-lobe damage. The results indicated that the subject’s ability to perceive steady-state acoustic stimuli was relatively intact but his ability to perceive
dynamic stimuli was drastically reduced. We conclude that this particular aspect of
acoustic processing may be a major contributing factor that disables speech perception in subjects with unilateral VAA.  2000 Academic Press
Key Words: verbal auditory agnosia; pure word-deafness; speech acoustics;
speech perception; auditory perception.

INTRODUCTION

An individual with verbal auditory agnosia (VAA), or pure word-deafness,
is unable to perceive spoken language in the absence of a significant hearing
impairment (Goldstein, 1974; Benson, 1996). His reading, spontaneous writing, and speaking abilities are relatively preserved. Since verbal auditory
agnosia is a relatively rare disorder, studies reporting findings on this disorder have been mainly case reports. The pathology of VAA, according to
Benson (1996, p. 304), ‘‘involves the primary auditory cortex (Heschl’s gyrus) or connections between the thalamus and this area.’’ However, reported
We thank all our subjects for their participation, Dr. Patricia McCarthy, and the anonymous
reviewers for their valuable comments on an earlier version of this article.
Address correspondence and reprint requests to Emily Wang, 1653 West Congress Parkway,
203 SENN, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL 60612. Fax: (312) 2438280. E-mail: ewang@rush.edu.
442
0093-934X/00 $35.00
Copyright  2000 by Academic Press
All rights of reproduction in any form reserved.

SPEECH ACOUSTIC PROCESSING IN VAA

443

cases indicate that VAA could have resulted from either cortical or subcortical, unilateral or bilateral brain lesions (e.g., Kanshepolsky, Kelley, & Waggener, 1973; Saffran, Marin, & Yeni-Komshian, 1976; Auerbach, Allard,
Naeser, Alexander, & Albert, 1982; Buchtel & Stewart, 1989; Kazui, Naritomi, Sawada, & Inoue, 1990; Godefroy, Leys, Furby, Reuck, Daems,
Rondepierre, Debachy, Deleume, & Desaulty, 1995). Auerbach et al. (1982)
suggested that there might be two different types of VAA: one is prephonemic in nature due to bilateral temporal lobe lesions and the other phonemic
in nature due to left unilateral lesions because the former could be explained
by deficits in the temporal auditory acuity observed in these patients while
the latter could not. A review of the literature on unilateral VAA, however,
indicates the following: on the one hand, hearing acuity, as defined by the
ability to perceive static pitch, timbre, and loudness and even sound localization, is adequate in these right-handed individuals with left temporal lobe
lesions; therefore, the speech-perception deficits observed cannot be readily
explained by the normal or near-normal hearing acuity (e.g., Gazzaniga,
Glass, Sarno, & Posner, 1973; Albert & Bear, 1974; Denes & Semenza,
1975; Saffran, Marin, & Yeni-Komshian, 1976). On the other hand, the ability to process various other aspects of acoustic signals may be impaired in
these individuals, which may account for in part, if not all, the speech-perception deficits observed (Albert and Bear, 1974; Saffran et al., 1976). For example, Albert and Bear (1974) found that their subject had a significant deficit in auditory temporal processing for both linguistic and nonlinguistic
acoustic signals. On a nonlinguistic click-fusion task, he consistently fused
two clicks at intervals of 15 ms while the normal controls were able to distinguish them as two at 1- to 3-ms separations. On a linguistic task, they found
that if they reduced the speaking rate to 1/3 of the normal rate, the subject’s
comprehension significantly improved. One interesting observation was reported in several studies that investigated VAA involving either left unilateral or bilateral lesions. That is, the subjects were able to perceive simple
vowels but not CV or CVC syllables (e.g., Denes & Semenza, 1975; Auerbach et al., 1982; Kazui et al., 1990; Godefroy et al., 1995). These findings
indicate that (1) prephonemic processing may also be affected in the VAA
associated with left unilateral lesions and (2) the ability for processing
steady-state acoustic information (such as vowels) vs the ability for processing rapidly changing acoustic information (such as formant transitions)
may be differentially affected in these individuals.
The acoustic signal of speech is characterized by rapid temporal changes
in fundamental frequency, intensity, and spectral properties. The ability to
understand spoken language or speech thus requires the ability to process
the temporal variations in the acoustic signal of speech. If one’s ability for
processing such variations is impaired, his or her ability to process spoken
language or speech is limited. The reverse, however, is not necessarily true.
Intact auditory processing ability may not guarantee intact comprehension

444

WANG ET AL.

of spoken language since language processing also requires the ability to
analyze other aspects of language, such as syntax, semantics, and pragmatics.
The fact that a VAA subject can comprehend written language suggests that
his or her ability to process these aspects of the language is largely retained.
Further, the fact that a VAA subject can perceive simple vowels but not CV
or CVC syllables suggests that it is the ability to process the temporal
changes in the speech signal that is probably responsible for the observed
impairment in auditory comprehension of spoken language or speech.
There are a variety of temporal changes in speech signals. There are clicklike sounds known as transients with durations as short as only a few milliseconds (Fant, 1973); there are rapid spectral changes near the boundaries
of consonants and vowels known as formant transitions whose durations are
typically around tens of milliseconds (Lehiste & Peterson, 1961); there are
also relatively slower spectral changes within a vowel or diphthong with
durations as long as hundreds of milliseconds (Lehiste & Peterson, 1961).
Past speech perception studies have found that all of the temporal changes
discussed above carry important cues for the identification of phonemes—
the smallest sound units in speech. Both the transients and formant transitions
are important for identifying stops (Liberman, Delattre, & Cooper, 1952);
slower formant changes are important for identifying diphthongs and glides
(Gay, 1968; Bladon, 1985). Kraus, McGee, and Koch (1998) reported their
findings on neurophysiologic responses and behavioral discrimination of
speech sound contrasts in normal and learning-disabled children. They presented the children with two pairs of phonetic contrasts (a /da/–/ga/ contrast
and a /ba/–/wa/ contrast) and measured both their neurophysiologic and
behavioral responses (for a detailed description of the study, please see Kraus
et al., 1998). As a group, the learning-disabled children had more difficulty
in discriminating the /da/–/ga/ contrast than the /ba/–/wa/ contrast. If we
look at the spectral analyses, it is clear that /wa/ in the /ba/⫺/wa/ contrast
has a longer F2 transition than those of stops. Thus, it is possible that this
longer transition is within the temporal processing capacity of these children
while the shorter transitions (in syllables with stops) are not. This finding,
although observed in a different clinical population, nonetheless indicates
that the ability to process relatively short dynamic patterns in speech acoustic
signals is a crucial one since the transition differences among these stimuli
are in tens of milliseconds. Albert and Bear (1974) observed that their subject
could comprehend spoken language significantly better when the examiner
reduced his speaking rate considerably to 1/3 of the normal rate. Although
these authors did not test the subject’s ability to process rapid formant transitions or even moderately slow temporal changes such as those in diphthongs,
their finding that VAA subjects had greater perceptual fusion of sequence
of clicks than normals suggests that the subject’s ability to process release
bursts in consonants is probably hampered. The ability of individuals with
VAA to process temporal changes lasting for tens or hundreds of millisec-

SPEECH ACOUSTIC PROCESSING IN VAA

445

onds, however, is still unclear. If this ability is impaired, it is conceivable
that failure in speech perception should also occur.
In the present study, we examined the ability to process dynamic versus
static variations in acoustic signals in a subject who presented with clinical
behaviors typical of unilateral verbal auditory agnosia. We focused in particular on the subject’s ability to process rapid or moderately slow temporal
acoustic changes lasting either tens or hundreds of milliseconds. In light of
existing findings about VAA, and based on what is known about speech
perception in general and in learning-disabled children, we hypothesized that
our subject’s ability to process the rapid dynamic acoustic variations was
impaired, but his ability to process moderately rapid dynamic acoustic variations was preserved. Hence, our subject should be able to process steadystate sounds such as vowels and moderately rapid dynamic acoustic variations such as those in diphthongs, but not able to process rapid dynamic
acoustic variations such as those in stops or CV syllables. To test this hypothesis, we used various speech and nonspeech signals as stimuli in a series of
perception tests administered to the subjects.
CASE DESCRIPTION

The subject is a 31-year-old, right-handed man, with a known history of
testicular cancer for which he underwent surgery and chemotherapy 10 years
prior to being seen for this study. He also was diagnosed with congenital
hypertrophic cardiomyopathy at 5 years of age. A pacemaker was placed
after syncopal episodes at age 24. He was admitted to Rush-PresbyterianSt. Luke’s Medical Center (RPSLMC) with a diagnosis of cerebrovascular
accident (CVA) following sudden onset of confusion, aphasia, and right
hemiparesis. Upon admission, the subject underwent a carotid angiogram
with urokinase infusion that revealed an occlusion of the left middle cerebral
artery, and the intraclot lysis with urokinase was successful. A noncontrast
computed tomographic (CT) scan of the head was performed that revealed
an infarct in the territory of the left middle cerebral artery. Repeat noncontrast CT scan performed 2 days later revealed continued evolution of focal
infarction in the left cerebral hemisphere involving the cortical and subcortical white matter of the left temporal lobe and extending superiorly into the
frontoparietal region (see Fig. 1).
Neurological Testing
The neurological examination on admission revealed that the subject was
confused, but occasionally uttered a clear spontaneous word. He had dense
comprehensive deficits. He also demonstrated a right facial droop and had
a right-sided weakness. Two days after admission, the neurological examination revealed the following: the subject’s Mini Mental State (Folstein,
Folstein, & McHugh, 1975) was appropriate to written input but not to spo-

446

WANG ET AL.

FIG. 1. A noncontrast CT scan performed 2 days postadmission which revealed continued
evolution of focal infarction in the left cerebral hemisphere involving the cortical and subcortical white matter of the left temporal lobe and extending superiorly into the frontoparietal
region.

ken conversation. The subject achieved a score of 21/30 on the Mini Mental
State Exam when the test items were presented visually rather than auditorily. His spontaneous speech was described as being appropriate with occasional paraphasias. He was unable to comprehend spoken language or follow
verbal commands but could follow written commands and read and write.
He responded to a tuning fork. His cranial nerve response was 20/30 on the
right and 20/20 on the left. Strength was 5/5 bilaterally for the upper extremities. Reflexes were symmetric. His fine motor ability was slightly decreased
in the right hand. He had right hand and right foot sensory loss. He had no
symptoms of ataxia and his gait was intact.
Initial Language Testing
The Western Aphasia Battery (WAB) was administered 4 days after the
admission. At the time of testing, the subject demonstrated appropriate pragmatic skills and fluent grammatical verbal output of normal phrase length
with phonemic paraphasias. Auditory comprehension was severely impaired

SPEECH ACOUSTIC PROCESSING IN VAA

447

TABLE 1
Pure Tone Threshold (in decibels) Results at 10Day Postonset
Frequency (Hz)
250
500
750
1000
2000
4000
8000

Left ear

Right ear

15
20
25
30
30
25
10

25
25
35
45
40
45
25

even for simple yes/no questions. Repetition was severely impaired, while
naming to confrontation was moderately to markedly impaired. Reading
comprehension was relatively preserved. The subject clearly exhibited contrasting abilities in comprehending language materials presented visually
rather than auditorily. His performance was consistent with a diagnosis of
pure word-deafness, i.e., relatively preserved visuoverbal language with severely impaired auditory-verbal comprehension (see Table 2).
Audiological Testing
An audiological evaluation was conducted 10 days postadmission that included pure-tone audiometry, tympanometry, acoustic reflex testing, speech
awareness threshold, and speech reception and recognition testing. The results from pure-tone threshold audiometry (see Table 1) showed that the
subject’s hearing was within normal limits for the left ear at 250–2 and 4–
8 kHz and a mild sensorineural hearing loss was detected at 1–2 kHz. For
the right ear, his hearing was within normal limits at 250–500 kHz and 8
kHz but a mild-to-moderate sensorineural hearing loss was detected at 750–
4 kHz. The hearing loss observed in this subject might have been related
to unknown etiology since he denied family history of hearing loss, history
of noise exposure, and history of ototoxic medication. Tympanogram and
speech awareness were within normal limits, and acoustic reflexes were present at expected levels for both ears. Speech-reception and -recognition testing
could not be completed due to the fact that the subject could not understand
any speech information even when he was provided with a list of the words
for speech-reception testing.
METHODS

Language Testing
The WAB test was readministered in a face-to-face clinical setting without any effort to
conceal the examiner’s face. However, even with such visual aids as lip reading to assist his

448

WANG ET AL.

TABLE 2
Results from WAB Testing
WAB subtest

Initial

1 week
postonset

1 month
postonset

5 months
postonset

Spontaneous Speech (20)
Auditory Comprehension (10)
Repetition (10)
Naming (10)
Aphasia Quotient (93.8)
Reading (100)

15
3.85
1.4
3.1
46.7
68

16
4.45
3.3
6.8
61.1
100

18
9.75
7.2
8.4
86.7
94

19
9.75
7.8
9.0
91.1
100

auditory comprehension, the subject continued to demonstrate severely impaired comprehension for language materials presented auditorily. Treatment was provided by a licensed speechlanguage pathologist on a daily basis after the initial language evaluation. Emphasis was placed
on training the subject to use visual information to aid his auditory comprehension. Consistent
improvement was observed in this area.
As can be seen from Table 2, the subject’s inability to comprehend and repeat information
presented auditorily was in clear contrast with his ability to comprehend complex information
presented visually. While the subject’s ability to comprehend written language reached a ceiling 7 days postonset, his ability to comprehend spoken language remained limited. He continued to improve on the auditory comprehension and repetition subtests at 1 month postonset,
although these scores largely reflected his improved ability to use visual cues following treatment. Evidence for this conclusion was obtained from the subject’s performance on a series of
discrimination/identification tasks that were administered concurrently with language testing.

Perception Testing—Subject
Perception testing was performed 1 week postonset of the subject’s CVA. The stimuli were
delivered via headphones in a quiet room at an intensity level of approximately 70 dB SPL.
Two speech perception tests were administered: identification and discrimination. Twentyfour minimal pairs of unique, single-syllable natural speech stimuli were used for both tests,
with half of them differing in vowels and the other half differing in initial or final consonants
(Appendix A). For the identification test, each stimulus was presented twice in a random order
with an ISI of 5000 ms, resulting in a total of 96 responses. The subject’s task was to circle
on an answer sheet the word he just heard from the pair of written stimuli presented visually.
For the discrimination test, the stimuli were presented in the original minimal pairs. To eliminate the order effect, the stimuli were counterbalanced, i.e., each stimulus was presented as
the first one as well as the last one within each pair. After counterbalancing, each pair of
stimuli was presented twice randomly to the subject, resulting in a total of 192 responses. The
subject’s task was to circle the word ‘‘same’’ or ‘‘different’’ on the answer sheet after he
heard each stimulus pair. Prior to the testing, written instructions for the tasks were provided
to the subject. Practice runs were conducted until he felt comfortable with the tasks and was
able to carry them out as required. The subject performed at chance level for both tests (45.83%
for identification and 46.32% for discrimination), indicating an inability to perceive the phonetic differences in the stimuli.
At 1 month postonset, a series of discrimination tests were administered (see Table 3). For
Tests A and B (see below), the stimuli were delivered via headphones in a sound-treated suite.
For Tests C–F, the stimuli were delivered via headphones in a quiet room at an intensity level
of approximately 70 dB SPL. All natural speech stimuli were produced by a male native

449

SPEECH ACOUSTIC PROCESSING IN VAA

TABLE 3
Result Summary from Perception Testing at 1 Month Postonset
Discrimination test
Amplitude
Frequency
Simple Vowels
Diphthongs
Word Stress
Sine Wave with Dynamic
Changes

No. of
total responses

% Correct—subject
(n ⫽ 1)

% Correct—controls
(n ⫽ 15)

20
20
72
72
64
64

95
95
100
100
60.93
60.93

—
—
—
—
99.69
98.13

speaker of American English from the same region as the subject. They then were digitized
and sequenced using the SoundEdit program (Macromedia).
(A) Intensity discrimination. The subject’s ability to detect differences in sound intensity
was measured. The discrimination test for intensity consisted of four unique pairs of steadystate pure-tone stimuli. Two intensity levels were used: 60 and 75 dB HL at 2 kHz. These
levels were minimally 20 dB HL above the subject’s pure-tone thresholds (see Table 1). Each
stimulus was presented at a duration of 1000 ms with an ISI of 500 ms. Each pair was presented
five times in a random order resulting in a total of 20 responses. The subject’s task was to
circle the word ‘‘same’’ or ‘‘different’’ on the answer sheet after he heard each pair of stimuli.
He performed with 95% accuracy.
(B) Frequency discrimination. The subject’s ability to detect differences in sound frequency
was measured. The Pitch Pattern Sequences (Pinheiro, 1977) were used for this task. The test
consists of four pairs of steady-state pure-tone stimuli. Two frequency levels are used: 880
and 1430 Hz. The stimuli were delivered at 85 dB HL, which is 50 dB HL above the subject’s
pure-tone-average (PTA) threshold. Each pair was presented five times with an ISI of 500 ms
in a random order, resulting in a total of 20 responses. Again, the subject performed with
95% accuracy.
(C) Simple-vowel discrimination. The discrimination test for simple vowels consisted of
three naturally produced tense vowels [a, i, u]. After counterbalancing, there were nine pairs
of unique stimuli. Each pair was presented eight times with an ISI of 500 ms in a random
order, resulting in a total of 72 responses. The subject performed with 100% accuracy.
(D) Diphthong discrimination. The discrimination test for diphthongs consisted of three
naturally produced diphthongs ([a, ou, e] at 590, 570, and 530 ms respectively). After
counterbalancing, there were nine pairs of unique stimuli. Each pair was presented eight times
with an ISI of 500 ms in a random order, resulting in a total of 72 responses. The subject
performed with 100% accuracy.
(E) Word-stress discrimination. The discrimination test for word stress consisted of two
pairs of naturally produced words differing in the primary word stress, e.g., ‘‘Whitehouse’’
vs ‘‘white house,’’ ‘‘hotdog’’ vs ‘‘hot dog.’’ After counterbalancing, there were eight pairs
of unique stimuli. Each pair was presented eight times with an ISI of 500 ms in a random
order, resulting in a total of 64 responses. The subject performed with 60.9% accuracy. A
one-sample t test with a hypothesized mean of 0.5 was conducted to determine whether the
performance was different from chance. The result was not statistically significant (t ⫽
⫺1.779, p ⫽ .080). A one-factor ANOVA was conducted to determine if the subject’s performance was any different for the two stimuli. The result was not statistically significant (F ⫽
.064, p ⫽ .802).
(F) Discrimination of sine waves with dynamic changes. The discrimination test for sine
waves with dynamic changes consisted of two pairs of synthesized stimuli. One pair had no
steady state (rising–falling), i.e., their frequencies changed continuously (800–1250 and 1250–

450

WANG ET AL.

800 Hz). The duration of the stimuli was 302 ms. The other pair had a short onset ramp (650–
1000 Hz signaling ‘‘up’’ and 1350–1000 Hz signaling ‘‘down’’) followed by a steady-state
tone of 1000 Hz. The duration of the onset ramp was 50 ms for both ‘‘up’’ and ‘‘down.’’
The total durations of the stimuli (including the duration of steady-state tone) were 292 ms
for ‘‘up’’ and 312 ms for ‘‘down.’’ After counterbalancing, there were eight pairs of unique
stimuli, i.e., four pairs for the ‘‘rising–falling’’ type and four for the ‘‘up–down’’ type. Each
pair was presented eight times with an ISI of 500 ms in random order, resulting in a total of
64 responses. The subject performed with 60.9% accuracy. A one-sample t test with a hypothesized mean of .5 was conducted to determine whether the performance was different from
chance. The result was not statistically significant (t ⫽ ⫺1.779, p ⫽ .08), indicating that the
subject’s performance was no better than chance. A one-factor ANOVA was conducted to
determine if there was any difference in the subject’s performance for the two types of stimuli
(i.e., continuous changing vs with an onset ramp). The result was not statistically significant
(F ⫽ .578, p ⫽ .450).

Perception Testing—Normal Controls
In order to determine if the stimuli were too difficult to discriminate even for normal subjects, 15 control subjects participated in discrimination tests E and F (see below). The subjects
were 1st- and 2nd-year graduate students in the speech-language pathology program at Rush
University with ages ranging from 22.5 to 34.4 years old. All control subjects were free from
speech and hearing impairment. Two of the 15 control subjects were older than the subject.
Both discrimination tests were conducted under the same experimental conditions used for
the subject. The results are presented below.
(E) Word-stress discrimination. The mean percentage correct for the discrimination of word
stress was 99.69% (ranging from 95.31 to 100% correct). A one-sample t test with a hypothesized mean of 0.5 was performed. The result was statistically significant (t ⫽ 279.874, p ⬍
.0001), indicating that the control subjects’ performance as a group was well above chance.
A one-factor ANOVA was conducted to compare the performance of the subject with the
group means of the control subjects. The result was statistically significant (F ⫽ 6.324, p ⬍
.0001), indicating that as a group, the control subjects performed better than the subject.
(F) Discrimination of sine waves with dynamic changes. The mean percentage correct for
discrimination of sine waves with dynamic changes was 99.97% (ranging from 99.90 to 100%
correct) A one-sample t test with a hypothesized mean of .5 was performed. The result was
statistically significant (t ⫽ 91.790, p ⬍ .0001), indicating that the control subjects’ performance as a group was well above chance. A one-factor ANOVA was conducted to compare
the performance of the subject with the group means of the control subjects. The result was
statistically significant (F ⫽ 6.080, p ⬍ .0001), indicating that the control subjects performed
better than the subject.
In order to determine if the control subjects performed differently for the two types of
stimuli, the percentage correct scores were examined. Of the 15 control subjects, 14 scored
100% correct and one 99.9% correct for the rising–falling type, while seven scored 100%
correct and eight ranged from 99.8 to 99.9% correct for the up–down type. A one-factor
ANOVA was conducted to determine if the observed difference between the two types of
stimuli was statistically significant. The result was indeed statistically significant (F ⫽ 7.314,
p ⬍ .0115). This seems to indicate that for the normal subjects, the up–down type of stimuli
was perceptually more challenging than the rising–falling type, although they all performed
well above chance.

DISCUSSION

Although verbal auditory agnosia has been described as an inability to
interpret the meanings of verbal sounds (a phonemic decoding disorder, see

SPEECH ACOUSTIC PROCESSING IN VAA

451

Goldstein, 1974), in the absence of hearing difficulties, recent studies have
shown that certain aspects of processing speech acoustic signals may be affected (Kanshepolsky et al., 1973; Albert & Bear, 1974; Saffran et al., 1976;
Buchtel & Stewart, 1989). Even though such changes were observed more
frequently in cases with bilateral temporal cerebral lesions (e.g., Kanshepolsky et al., 1973; Glass, Sarno, & Posner, 1973; Auerbach et al., 1982), they
were observed in cases with left unilateral temporal cerebral lesions as well.
Albert & Bear (1974) reported decreased ability to handle temporal resolution in their subject, and Saffran, Marin, & Yeni-Komshian (1976) reported
the need for increased intensity level for sound processing in their subject.
These findings suggest that it is possible that in certain cases of unilateral
VAA, similar to VAA involving bilateral temporal cerebral lesions, acoustic
processing of speech signals may be affected at the prephonemic level rather
than solely at the phonemic level, as suggested by Auerbach et al. (1982).
In this study, we have found that a subject with verbal auditory agnosia
from focal infarction of the left cerebral hemisphere, had no difficulty detecting differences in the intensity and frequency of nonspeech stimuli nor
among simple vowels or even diphthongs. However, the same subject
showed great difficulty discriminating sine waves with dynamic changes,
minimal-pair words, and word stress. His difficulty could not be explained
by changes in his hearing since all stimuli were presented well above his
pure-tone thresholds. Inspection of those stimuli that the subject could discriminate and those he could not revealed that it was his ability to handle
dynamic changes in acoustic signals that was really impaired differentially.
The subject’s inability to perceive rapid dynamic variations lasting 50 ms
is consistent with his inability to discriminate different consonant places of
articulation (such as /ba/ vs /da/). It has been well established that formant
transitions, whose duration is usually around 50 ms, are critical for identifying consonant place of articulation (Liberman et al., 1952). VAA subjects
who are unable to process either click-like sounds (Albert & Bear, 1974),
which are similar to the release bursts in stop consonants, or pure-tone glides
with durations similar to that of formant transitions, as found in the present
study, have lost the ability to detect two of the most important acoustic cues
for the perception of stop consonants. It follows therefore that such subjects
cannot properly perceive stop consonants in speech. Phillips and Farmer
(1990) analyzed existing evidence and concluded that the temporal processing deficit seen in VAA subjects was largely restricted to sound elements
that are very brief in duration (i.e., from milliseconds to tens of milliseconds).
In this study, we have taken a step further to show that the deficits in VAA
subjects are not only in their inability to process sound elements that are
short in duration, but also their inability to process dynamic acoustic changes
that occur in a short period of time.
What is surprising is the finding that the auditory impairment of this VAA
subject was so severe that he could not even discriminate pure-tone glides
(rising–falling) as long as 300 ms. For the normal controls in this study, the

452

WANG ET AL.

continuous tone glides (rising–falling type) were slightly easier to discriminate than the transitionlike tone glides (up–down type). However, the subject’s performance suggests that the two types of dynamic stimuli, rising–
falling and up–down, were equally challenging to him and that he could not
perform better than chance for either type of stimuli. Many dynamic patterns
in speech last only as long as several hundred milliseconds. The average
syllable duration in spontaneous speech is 256–333 ms (summarized by
Ohde & Sharf, 1992, p. 271), and the speaking rate ranges from 150 to 250
words/min or four to seven syllables/s (Klatt, 1976). The dynamic spectral
changes of vowels, diphthongs, and glides all occur within the duration of a
syllable. If VAA subjects cannot detect dynamic changes in acoustic signals
lasting as long as 300 ms, it is unlikely that they can process speech utterances in any practical way.
It is interesting that this subject was able to discriminate the diphthong pairs
used in this study. There may be two reasons for this result. First, the duration
of these stimuli might have been sufficiently long for him. In a study examining
the patterns of discrimination between tone glides that were consonant–vowel(CV) and vowel–consonant- (VC) like formant transitions, Van Wieringen and
Pols (1995) found that as the transition duration increased, the discriminability
improved even when the total duration of the stimulus remained constant. The
improved discriminability was due to the fact that as the transition duration
increased, all difference limens decreased, the spectral change decreased, and
the processing time increased. The longest transition duration these authors
used was 50 ms. In the current study, the subject was unable to discriminate
stimuli that were as long as 310 ms in duration, indicating a serious problem
in his ability to process dynamic changes. However, the shortest duration of
our diphthong stimuli was 530 ms (since they were all produced in isolation).
This may have exceeded his difference limen for detecting dynamic changes.
The second reason for his successful discrimination of the diphthong stimuli
may be the fact that they were not really minimal pairs in the sense that the
only difference in each pair is in the dynamics rather than anything else. The
three diphthongs used in this study, /a/, /ou/ and /e/, differed not only in
their dynamic patterns, but also in their starting and ending formant values. In
other words, to the subject, these diphthongs may have sounded just like different steady-state vowels.
Speech perception involves a long chain of events beginning with hearing.
The acoustic signal picked up by the ears has to be processed further to
reveal patterns critical for the identification of different speech sounds. As
argued by Liberman and Mattingly (1985, p. 28), speech perception ‘‘uses
all the information in the stimulus that is relevant to phonetic structures:
every potential cue proves to be an actual cue.’’ If, however, somewhere
along the way in the auditory system, the mechanism for detecting dynamic
changes in the acoustic signal is damaged, as found in VAA, conversion of
dynamic patterns into the phonetic forms needed for higher processing would
be hindered, making it impossible for the listener to recognize spoken words

SPEECH ACOUSTIC PROCESSING IN VAA

453

or sentences. In humans, the ability to process dynamic information such as
formant transitions and transients in speech signals is highly remarkable. In
normals, speech signals can be processed at a speed of seven syllables/s
(Klatt, 1976). However, when the auditory system is damaged, this ability
may be greatly affected, ranging from greatly reduced to nonfunctional, depending on the amount and the location of the damage. Evidence supporting
this argument can also be found in various case studies showing that VAA
subjects could discriminate simple vowels but not CV or CVC syllables (e.g.,
Denes & Semenza, 1975; Auerbach et al., 1982; Kazui et al., 1990; Godefroy
et al., 1995) and in the phenomenon of rate-dependent comprehension as
observed by Albert and Bear (1974).
In conclusion, based on the findings of the present study, it seems that
what is critically affected in this type of unilateral verbal auditory agnosia,
or at least in this subject, is the ability to process dynamic patterns in acoustic
signal, which in turn impairs the subject’s ability to perceive speech.

APPENDIX A
Perception Testing Stimuli-Minimal Pairs
Item No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

Phonetic
transcription

Gloss

Phonetic
transcription

Gloss

b{t
f{t
led
l{b
bi:t
ni:t
hi:l
seθ
get
led
di:l
fi:t
i:t
hi:t
pi:tʃ
li:k
fi:d
mi:t
bi:d
si:t
pet
set
led
ʃed

bat
fat
led
lab
beat
neat
heal
Seth
get
led
deal
feet
eat
heat
peach
leak
feed
meat
bead
seat
pet
set
led
shed

p{t
k{t
let
l{d
mi:t
si:t
hi:t
set
met
ted
di:d
fi:d
{t
h{t
p{tʃ
l{k
fed
met
bed
set
p{t
s{t
l{d
ʃ{d

pat
cat
let
lad
meat
seat
heat
set
met
Ted
deed
feed
at
hat
patch
lack
fed
met
bed
set
pat
sat
lad
shad

454

WANG ET AL.

REFERENCES
Albert, M., & Bear, D. (1974). Time to understand. Brain, 97, 373–384.
Auerbach, A. H., Allard, T., Naeser, M., Alexander, M. P., and Albert, M. L. (1982). Pure
word deafness-analysis of a case with bilateral lesions and a defect at the prephonemic
level. Brain, 105, 271–300.
Benson, D. F., & Ardila, A. (1996). Aphasia. New York: Oxford Univ. Press.
Bladon, A. (1985). Diphthongs: A case study of dynamic auditory processing. Speech Communication, 4, 145–154.
Buchtel, H. A., & Stewart, J. D. (1989). Auditory agnosia: Apperceptive or associative disorder? Brain and Language, 37, 12–25.
Denes, G., & Semenza, C. (1975). Auditory modality specific anomia: Evidence from a case
of pure word deafness. Cortex, 11, 401–411.
Fant, G. (1973). Speech sounds and features. Cambridge, MA: M.I.T. Press.
Folstein, M. D., Folstein, S. E., & McHugh, P. (1975). Mini-Mental State: A practical method
for grading the cognitive state of patients for the clinician. Journal of Psychiatry Research, 12, 189–198.
Gay, T. J. (1968). Effect of speaking rate on diphthong formant movements. Journal of the
Acoustical Society of America, 44, 1570–1573.
Gazzaniga, M. S., Glass, A. V., Sarno, M. T., & Posner, J. B. (1973). Pure word deafness
and hemispheric dynamics: A case history. Cortex, 9, 136–143.
Godefroy, O., Leys, D., Furby, A., Reuck, J. D., Daems, C., Rondepierre, P., Debachy, B.,
Deleume, J.-F., & Desaulty, A. (1995). Psychoacoustical deficits related to bilateral subcortical hemorrhages—a case with apperceptive auditory agnosia. Cortex, 31, 149–159.
Goldstein, M. N. 1974. Auditory agnosia for speech (‘‘Pure word-deafness’’). Brain and Language, 1, 195–204.
Kanshepolsky, J., Kelley, J. J., & Waggener, J. D. (1973). A cortical auditory disorder. Clinical
audiologic and pathologic aspects. Neurology, 23, 699–705.
Kazui, S., Naritomi, H., Sawada, T., & Inoue, N. (1990). Subcortical auditory agnosia. Brain
and Language, 38, 476–487.
Klatt, D. H. (1976). Linguistic uses of segmental duration in English: Acoustic and perceptual
evidence. Journal of Acoustic Society of America, 59(5), 1208–1221.
Kraus, N., McGee, T. J., & Koch, D. B. (1998). Speech sound representation, perception, and
plasticity: A neurophysiologic perspective. Audiology and Neuro-Otology, 3, 168–182.
Lehiste, I., & Peterson, G. E. (1961). Transitions, glides and diphthongs. Journal of the Acoustical Society of America, 33, 268–277.
Liberman, A. M., Delattre, P. C., & Cooper, F. S. (1952). The role of stimulus variables in
the perception of stop consonants. American Journal of Psychology, 65, 497–516.
Liberman, A. M., & Mattingly, I. G. (1985). The motor theory of speech perception revised.
Cognition, 21, 1–36.
Ohde, R. N., & Sharf, D. J. (1992). Phonetic analysis of normal and abnormal speech. New
York: Merrill.
Phillips, D. P., & Farmer, M. E. (1990). Acquired word deafness, and the temporal grain of
sound representation in the primary auditory cortex. Behavioral Brain Research, 40, 85–
94.
Pinheiro, M. L. (1977). Auditory pattern reversal in auditory perception in patients with left
and right hemisphere lesions. Ohio Journal of Speech and Hearing, 12, 9–20.

SPEECH ACOUSTIC PROCESSING IN VAA

455

Saffran, E. M., Marin, O. S. M., & Yeni-Komshian, G. H. (1976). An analysis of speech
perception in word deafness. Brain and Language, 3, 209–228.
Van Wieringen, A., & Pols, L. C. W. (1995). Discrimination of single and complex consonantvowel- and vowel-consonant-like formant transitions. Journal of Acoustical Society,
98(3), 1304–1312.