Brain and Language 76, 119–129 (2001)
doi:10.1006/brln.2000.2417, available online at http://www.idealibrary.com on

Disorder in Sequential Speech Perception:
A Case Study on Pure Word Deafness
Sachiko Nakakoshi
Tokyo Metropolitan Okubo Hospital

Makio Kashino
NTT Basic Research Laboratories

Atsushi Mizobuchi
Hosei University

Yoshiro Fukada
and
Hiroto Katori
Teikyo University

We described disorders of a patient which were uniquely restricted to speech perception
of syllable sequences after brain damage. The results of series of experiments using syllable
sequences showed ‘‘negative recency effect,’’ in which the subject’s repetition performance
at the latter syllable position was remarkably poor. Experimental analyses suggested that the
‘‘negative recency effect’’ could be due to dual factors: the lower rate of processing of speech
sounds and the memory load of holding processes of preceding syllables imposed on the
succeeding phonological processing. The results also suggested that the holding processes
which imposed the memory load on the succeeding auditory phonological coding processing
were modality nonspecific.  2001 Academic Press

INTRODUCTION

Most neuropsychological studies on cognitive processes of speech sounds have
been mainly focused on impairments of perceiving monosyllables. Most of them
researched patients with pure word deafness (Albert & Bear, 1974; Auerbach, Allard,
Naeser, Allexander, & Albert, 1982; Notoya, Suzuki, Furukawa, & Enokido, 1991;
Tanaka, Yamadori, & Mori, 1987; Saffran, Marin, & Yeni Komshian, 1976; Caramazza, Berndt, & Basili, 1983). Albert and Bear (1974) and Auerbach et al. (1982)
emphasized that the defect in auditory temporal resolution played an important role
for the development of pure word deafness. They found that patients with pure word
deafness were markedly impaired in detecting gaps between acoustic stimuli. Notoya

Address correspondence and reprint requests to Sachiko Nakakoshi, Okubo Hospital, 2-44-1 Kabukicho Shinjuku-ku, Tokyo, Japan 160-8488. Fax: ⫹81-3-5273-7748. E-mail CZQ04340@niftyserve.
or.jp.
119
0093-934X/01 $35.00
Copyright  2001 by Academic Press
All rights of reproduction in any form reserved.

120

NAKAKOSHI ET AL.

et al. (1991) and Tanaka et al. (1987) suggested that besides this deficit, the difficulties
in loudness discrimination and pitch discrimination may also cause word deafness.
These authors have emphasized defects in acoustic level as the underlying defect in
pure word deafness. On the other hand, Denes and Semanza (1975) and Saffran et
al. (1976) emphasized the role of phonemic discrimination. Auerbach et al. (1982)
reviewed previous reports and suggested that there are two distinct types of pure
word deafness: type 1, in which the deficit is prephonemic and related to an auditory
temporal acuity disorder; and type 2, a form that has a deficit in linguistic discrimination, independent of an auditory temporal acuity disorder.
Speech, however, consists of a series of syllable sequences. Therefore, it is essential to analyze this type of disorder experimentally in order to understand the cognitive
processes of speech. Recently we encountered a unique patient whose language deficiencies could not be explained by a deficit in monosyllable perception. Instead, impairment in processing syllable sequences was found to be critical for this case. We
examined this subject’s sequential syllable perception in an effort to clarify the mechanism.
CASE REPORT

The patient KT was a 42-year-old right-handed male who had been working as
an engineer after earning a Master’s degree. He had a left subcortical hemorrhage,
and underwent surgery to remove hematomas twice in 1988. He showed right hemianopsia and right hemiplegia. Although postoperatively he manifested Wernicke’s
aphasia, 1 year later his language ability had recovered well; he scored full marks
in each subtest of the Standard Language Test of Aphasia (Nirayama Conference,
1975) except for several subtests using auditory channel. His main complaint was a
frequent failure to catch syllables in unfamiliar words in his daily life, and this listening disturbance remained even in 1995. An MRI performed in 1993 revealed a
focal lesion in the left putamen (Fig. 1). He scored above average on a WAIS-R
conducted in 1988, showing no overall deficiency in intelligence (VIQ ⫽ 114; PIQ ⫽
104; questions were presented visually). Pure tone test and auditory brain-stem responses were within normal range for both ears. No abnormalities were found in
our screening tests for recognition of environmental sounds and musical melody.
Furthermore KT could detect 3-ms gaps on a gap detection test (Penner, 1977), and
it showed that his temporal acuity was normal. KT could perceive Japanese monosyllables with an accuracy rate of 85%.
EXPERIMENTAL INVESTIGATION

Experiment 1 Repetition Tasks
We considered it strange that, compared to his ability of perception of monosyllables, his mistakes
of catching words were very frequent in daily conversation. So to examine KT’s listening disturbances
further, repetition tasks were conducted.
Stimulus materials. Two- to four-syllabic familiar words (nouns) and nonwords 1 were used. Thirty
words were presented for each type.
Subjects. KT and two persons with peripheral hearing disorders who were employed as controls
were the subjects.
Procedure. Each stimulus was presented orally by a female instructor in the free field at normal
speech rate. In advance the information whether the stimulus was a real word or a nonword was given
to the subjects. The subject was asked to repeat the words immediately.
1
Most Japanese nouns or verbs consist of two to seven monosyllables which are similar to vowel
combinations or CV syllables in English, so the stimuli we used here have the same syllable components
as real words.

SEQUENTIAL SPEECH PERCEPTION

121

(a)

(b)
FIG. 1. MR brain scan showing restricted lesion of putamen and dilatation of left lateral ventricle
on axial and coronal slices.

Results and discussion (see Figs. 2a–2c). KT and the control subjects showed difficulty repeating
nonwords compared to real words. This can be explained by the fact that KT and the control subjects
had intact semantic systems, and for real words, they could take advantage of the semantic cues. A
striking feature of KT’s results was the pattern of errors for nonwords. The accuracy of repetition declined
as a function of syllable positions. Errors were mostly substitutions with a syllable quite dissimilar to
the target syllable (for example, /tasajiko/→/tasanina/; /soyayuho/→/sohayabu/), and according to the
results of error analysis we could not find any systematic error such as replacement of the syllable orders.
By contrast this tendency was not shown by the two control subjects. Provisionally we use the term
‘‘negative recency effect’’ to refer to KT’s error pattern in nonwords, in contrast to the recency effect
in a free recall test.
How could we explain the negative recency effect in repetition tasks performed by KT? We think
there are three explanations. The first is that the declination could be due to a faster-than-normal decay
rate in the holding process of phonological information; that is, he forgets ‘‘recent,’’ or latter, syllables
while he reports ‘‘primal,’’ or earlier, syllables. The second is that it could be due to limited capacity
in phonological storage, and the third is that it could be due to disturbance of phonological coding
processes for ‘‘recent syllables’’ or syllables that occur later in a sequence. To explore these possible
explanations, we conducted the following three experiments.

Experiment 2 Backward Repetition Tasks
At first we examined the first explanation using backward repetition. As Japanese words can be easily
decomposed into monosyllables which are similar to CV syllables in English, any words can be pronounced ‘‘backward’’ using such monosyllables as pronunciation units. In these tasks subjects were
asked to report the last syllables in a stimulus nonword at first, so if KT’s deficit were present in the
holding process, the accuracy of the last syllables might improve.
Stimulus materials and subjects. These were the same as for Experiment 1, except that four-syllabic
nonwords were not presented. In four-syllabic nonwords, it was too difficult for subjects to perform this
task and no responses were often seen.

122

NAKAKOSHI ET AL.

(a)

(b)
FIG. 2. (a) Results of repetition tasks in KT. (b, c) Results of repetition tasks in two control subjects
with peripheral hearing disorders. The negative recency effect was not found in these subjects.

Procedure. KT and control subjects were asked to repeat words ‘‘backward.’’
Results and discussion (see Fig. 3). KT showed a negative recency effect again in two- or threesyllabic nonwords in this ‘‘backward’’ repetition task (Fig. 3a). For the preceding syllables in a word
that were heard earlier but were reported later, KT showed no difficulties and generally was successful,
but again had difficulty reporting the syllables in the last syllable position that he was asked to report
first. The control subjects yielded no such pattern of performance (Figs. 3b, 3c). Thus the negative
recency effect found in the forward repetition cannot be attributed to the deficits of the holding process
for phonological information. That is, it is unlikely that he had forgotten syllables in the latter syllable
position while reporting earlier syllables.

SEQUENTIAL SPEECH PERCEPTION

123

(c)
FIG. 2—Continued

Experiment 3 Reading Tasks
This experiment examined KT’s phonological storage capacity by using visually presented syllable
sequence stimuli.
Stimulus materials. One hundred twenty nonwords composed of four syllables were used.
Subject. KT was the only subject.
Procedure. On each trial, four different kana letters (each represents a Japanese monosyllable) were
presented in succession on a computer-controlled tachistoscope display. Each kana letter was presented
for 150 ms on a tachistoscopic display. Every time kana letter was removed, a mask stimulus, [&], was
presented for 50 ms. After the mask stimulus, there was a 300-ms blank, and the next letter was presented
on the right side of the former letter. Four letters in each trial were located at different places on the
tachistoscopic display. KT was to report orally all the letters after they had been presented and removed.
Results and discussion (see Fig. 4). KT’s scores were very good for all the serial positions. In addition
to intact visual coding processing from letters to speech sounds, his phonological storage capacity was
large enough to retain a four-unit syllables for oral reporting. His negative recency effect in repetition,
therefore, could not be explained by a limited phonological storage capacity. Therefore the second possible explanation was eliminated.

Experiment 4 Repetition Tasks and Syllable Matching Tasks with Varied
Presentation Intervals
The remaining possibility, then, is that KT has a deficit in phonological coding processes. We assumed
that he would not have enough time to build phonological representations for ‘‘recent’’ syllables, because
processing of primal syllables would consume an inordinate amount of time, given the natural speed of
sequential presentation. If so, then his scores for recent syllables would be likely to improve when stimuli
were presented by separating the component syllables with longer intervals. This hypothesis was tested
by varying stimulus onset asynchrony (SOA).
Repetition task: stimulus materials. Thirty nonwords composed of four syllables were used for each
SOA.
Subject. KT was the only subject.
Procedure. SOAs were set at 250, 500, 1000, and 2000 ms. Each syllable length was about 250 ms
under any condition. KT was asked to repeat the auditorily presented stimuli.
Results and discussion (see Fig. 5a). KT’s scores improved slightly but the negative recency effect
was still found even with the longest intervals.
In all the experiments described above, retaining earlier syllables could have interfered with processing

124

NAKAKOSHI ET AL.

(a)

(b)
FIG. 3. (a) Results of backward repetition tasks in KT. The X axis shows syllable position of stimulus. KT reported from the right to the left in X axis. The syllables presented earlier were produced
better while KT reported later. These results do not support that KT’s holding processes of phonological
information are disturbed. (b, c) Results of backward repetition tasks in two control subjects. The control
subjects did not exhibit KT’s pattern of performance.

SEQUENTIAL SPEECH PERCEPTION

125

(c)
FIG. 3—Continued
latter syllables, as KT was asked to report all components presented. To examine this possibility, we
conducted syllable-to-syllable matching tasks.
Syllable matching tasks: stimulus materials. One hundred eighty nonwords composed of four syllables for each SOA were used. Among the stimuli, 120 nonwords had a target syllable embedded in one
of four syllable positions and 60 nonwords had no target syllables.
Subject. KT was the only subject.
Procedure. SOAs were set at 250, 500, 1000, and 2000 ms. After a target syllable was given in
print in advance, a stimuli was auditorily presented. KT was asked to judge whether the target was
contained in the stimulus. Five target syllables were used.
Results and discussion (see Fig. 5b). KT’s scores of no target syllable trials showed over 90% correct
through these tasks. Therefore it could be judged that his responses were reliable. Under conditions
where SOAs were longer than 1000 ms, the negative recency effect disappeared; that is, KT earned over
90% correct for syllables in all positions. This pattern of performance supports the hypothesis posed
above that the rate of phonological coding process should be lower.
Then why in the repetition tasks did his performance for recent syllables not improve to a normal
level even in slow presentations? In the repetition tasks, KT had to retain all syllables and to repeat

FIG. 4.

Results of visually presented tasks in KT.

126

NAKAKOSHI ET AL.

(a)

(b)
FIG. 5. (a) Results of repetition tasks varied in SOA. The negative recency effect was found even
with the longest intervals. (b) Results of matching tasks varied in SOA. In this task the scores of later
syllables improved under the condition that SOAs exceeded 1000 ms.

them. In the matching tasks, on the other hand, KT did not have to retain all syllables, because this
experiment was designed to judge only whether the target was contained. Therefore it could be suggested
that the retention of earlier syllables might affect the processing of recent syllables. Let us call this
‘‘memory load hypothesis.’’ To confirm it further, we conducted a final experiment in which two conditions were set using different combinations of modalities for presenting a stimulus composed of four
syllables.

Experiment 5 Visual and Auditory Tasks
Stimulus materials. Thirty nonwords composed of four syllables were used for each trial.
Subject. KT was the only subject.
Procedure. Under the first condition, the first and second two syllables were presented visually in
print and the third and fourth two syllables were presented auditorily (V-A condition). Under the second
condition the first and second two syllables were presented auditorily and the third and fourth two syllables were presented visually (A-V condition). KT was asked to report all the syllables orally immediately
after a complete sequence was presented. The SOAs were set at 500, 1000, and 2000 ms.
Results and discussion (see Fig. 6). Under the V-A condition performance for the third and fourth
two syllables was poor; thus the negative recency effect was found again. In contrast, under the A-V
condition, the third and fourth syllables were reported very accurately. The percentage correct of the
third and fourth syllables under the V-A condition was lower than that of the first and second syllables

SEQUENTIAL SPEECH PERCEPTION

127

FIG. 6. Results of auditory–visual tasks in KT. Under the first condition, the first two syllables
were presented visually in print and the last two syllables were presented auditorily (V-A condition).
Under the second condition, the first two syllables were presented auditorily and the last two syllables
were presented visually (A-V condition). KT was asked to report all the syllables verbally immediately
after a complete sequence was presented.

under the A-V condition. This pattern of performance suggested that the retention of earlier syllables
interfered with the processing of latter syllables presented auditorily, regardless of whether the primal
syllables were presented visually or auditorily. The memory load hypothesis, therefore, is supported.
But memory load should not affect phonological processing of latter syllables presented visually.

GENERAL DISCUSSION

The main points of our findings can be summarized as follows. First, KT had no
deficit in acoustic temporal resolution, which have been considered to play an important role for patients with pure word deafness as a result of bilateral lesions. Although
KT preserved fairly good ability to perceive monosyllables, he had remarkable difficulties in perceiving nonsense syllable sequences, showing the ‘‘negative recency
effect’’ in repetition tasks. KT’s repetition performance at the latter syllable position
in a nonword was remarkably poor. The effect disappeared only when stimuli were
presented slowly and retention of earlier syllables was not required, as shown in the
matching tasks. These findings suggested that KT’s disorders could be explained by
dual factors: a lowered rate of speech processing, and some interferences made by
retained earlier syllables. That is, KT could not make phonological code for latter
syllables because he took longer to make phonological codes for earlier syllables.
Moreover, retention of earlier syllables somehow interfered with the phonological
coding process of latter syllables (Fig. 7).
This case is classified as ‘‘pure word deafness’’ or ‘‘phonemic deafness’’ because
his disorder was restricted to speech perception without severe hearing loss, although
his disorder was extremely mild. As we related in the Introduction, recent neuro-

128

NAKAKOSHI ET AL.

FIG. 7. Model of speech sound processing in KT. S1, first syllable; S2, second syllable; S3, third
syllable; S4, fourth syllable. Dashed arrow shows the interference.

psychological research on word deafness has focused mainly on the analysis of monosyllable perception. Although a few earlier studies suggested that comprehension of
patients with word deafness improved to some degree when the speech rate was slow
(Albert & Bear, 1974; Okada, Hanada, Hattori, & Shoyama, 1963; Metz-Lutz & Dahl,
1984), little attention has been paid to the deficits in sequential speech perception. We
believe that KT’s case may shed some additional light on understanding of pure word
deafness.
According to Bub, Daniel, and Arguin (1995), in patients with pure alexia, or letterby-letter reading, word length greatly affects response time. Such patients needed 3
or 4 s to read even common three-letter words and response time increased by 2 or
3 s for one additional letter. Although there are differences in information processing
between visual and auditory tasks, KT’s negative recency effect is very similar to
the effects in pure alexia, because both types of patients are characterized by slow
performance. Generally speaking, in visual tasks, all materials that compose words
are presented simultaneously, while in auditory tasks, each syllable is presented one
by one. Therefore, it could be that in visual tasks the lower rate of processing presents
as increased response time, whereas in auditory tasks it presents as errors in latter
syllables. Moreover, there is the discrepancy that KT could have utilized cues for
word meaning, but a patient with pure alexia could not. We think further studies
comparing with both visual and auditory disorders will be beneficial for understanding mechanisms of information processing in brain-damaged patients.
Furthermore, memory-to-perception interference might offer a new point of view
about the function of working memory (Baddeley & Wilson, 1993). The relationship
of speech perception and short-term memory has been recently discussed by many
researchers. Allport (1984) and Friedrich et al. (Friedrich, Glenn, & Marin, 1984)
argued that short-term deficits were inseparable from speech perception deficits. On
the other hand, Vallar and Baddeley (1984a,b), and Randi and Breedin (1992) reported patients who showed a phonological short-term memory deficit, yet did not
show speech perception difficulties. Our data showed the opposite: the possibility
that retaining information might interfere with perceptual processing.
Last, we mention KT’s lesion. It was restricted to the left putamen, as we described,
but the MR brain scan also showed dilatation of the left lateral ventricle. It is very

SEQUENTIAL SPEECH PERCEPTION

129

rare for a subcortical lesion to cause pure word deafness; however, Hamanaka et al.
(Hamanaka, Asano, Morimune, & Seko, 1980) reported a patient with pure word
deafness caused by a unilateral left subcortical lesion. Nadeau and Croson have discussed controversies about whether damage to the putamen can directly cause language disorders (Nadeau & Croson, 1997). It is not clear whether the language disorders discussed by Nadeau and Croson include speech perception disorders, and the
lateralization for KT’s disorders was not clear because the dilatation of the left lateral
ventricle could have affected cortical function.
REFERENCES
Albert, M. L., & Bear, D. (1974). Time to understand. A case study of word deafness with reference
to the role of time in auditory comprehension. Brain, 97, 373–384.
Allport, A. (1984). Auditory–verbal short-term memory and conduction aphasia. In H. Bouman & D.
Bouwhis (Eds.), Attention & performance (Vol. 10, pp. 313–326). Hillsdale, NJ: Erlbaum.
Auerbach, S. H., Allard, T., Naeser, M., Allexander, M. P., & 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.
Baddeley, A., & Wilson, B. (1993). A case of word deafness with preserved span: Implications for the
structure and function of short-term memory. Cortex, 29, 741–748.
Bub, Daniel, N., & Arguin, M. (1995). Visual word activation in pure alexia. Brain and Language, 49,
77–133.
Caramazza, A., Berndt, R. S., & Basili, A. G. (1983). The selective impairment of phonological processing: A case study. Brain and Language, 18, 128–174.
Denes, G., & Semanza, C. (1975). Auditory modality-specific anomia: Evidence from a case of pure
word deafness. Cortex, 11, 401–411.
Friedrich, F. J., Glenn, C. G., & Marin, O. S. M. (1984). Interruption of phonological coding in conduction aphasia. Brain and Language, 22, 266–291.
Hamanaka, H., Asano, K., Morimune, S., & Seko, K. (1980). Ein Fall von reiner Worttaubheit ohne
akustische Agnosie. Studia Phonologica, 14, 16–24.
Metz-Lutz, M., & Dahl, E. (1984). Analysis of word comprehension in a case of pure word deafness.
Brain and Language, 12, 13–25.
Nadeau, S. E., & Croson, B. (1997). Subcortical aphasia. Brain and Language, 58, 355–402.
Nirayama Conference (1975). Standard Language Test of Aphasia. Homeido Press.
Notoya, M., Suzuki, S., Furukawa, M., & Enokido, H. (1991). A case of pure word deafness associated
with Landau–Kleffner syndrome: A long-term study of auditory disturbance. Auris Nasus Larynx,
18, 297–305.
Okada, S., Hanada, M., Hattori, H., & Shoyama, T. (1963). A case of pure word-deafness. Studia Phonologica, 3, 58–65.
Penner, M. J. (1977). Detection of temporal gaps in noise as a measure of the decay of auditory sensation.
Journal of Acoustic Society America, 61, 552–557.
Randi, C. M., & Breedin, S. D. (1992). Dissociations between speech perception and phonological shortterm memory deficits. Cognitive Neuropsychology, 9, 509–534.
Saffran, E. M., Marin, O. S., & Yeni Komshian, G. H. (1976). An analysis of speech perception in word
deafness. Brain and Language, 3, 209–228.
Tanaka, Y., Yamadori, A., & Mori, E. (1987). Pure word deafness following bilateral lesions. A psychophysical analysis. Brain, 110, 381–403.
Vallar, G., & Baddeley, A. D. (1984a). Phonological short-term store, phonological processing, and
sentence comprehension: A neuropsychological case study. Cognitive Neuropsychology, 1, 121–
141.
Vallar, G., & Baddeley, A. D. (1984b). Fractionation of working memory: Neuropsychological evidence
for a phonological short-term store. Journal of Verbal Learning and Verbal Behavior, 23, 151–
161.