Neurología y Canto

Singing in the Brain: Insights from Cognitive Neuropsychology
Author(s): ISABELLE PERETZ, LISE GAGNON, SYLVIE HÉBERT and JOËL MACOIR
Source: Music Perception: An Interdisciplinary Journal, Vol. 21, No. 3 (Spring 2004), pp. 373-
390
Published by: University of California Press
Stable URL: http://www.jstor.org/stable/10.1525/mp.2004.21.3.373 .
Accessed: 24/06/2015 23:51
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
University of California Press is collaborating with JSTOR to digitize, preserve and extend access to Music
Perception: An Interdisciplinary Journal.
http://www.jstor.org
This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

All use subject to JSTOR Terms and Conditions
Music Perception © 2004 BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA
Spring 2004, Vol. 21, No. 3, 373–390 ALL RIGHTS RESERVED.
Singing in the Brain: Insights from Cognitive

Neuropsychology
ISABELLE PERETZ
University of Montreal & Institut Universitaire de Gériatrie de Montréal
LISE GAGNON
Sherbrooke Geriatric University Institute
S Y LV I E H É B E R T
University of Montreal & Institut Universitaire de Gériatrie de Montréal
JOËL MACOIR
Laval University and Laval University Geriatric Research Unit
Singing abilities are rarely examined despite the fact that their study repre-
sents one of the richest sources of information regarding how music is
processed in the brain. In particular, the analysis of singing performance in
brain-damaged patients provides key information regarding the autonomy
of music processing relative to language processing. Here, we review the
relevant literature, mostly on the perception and memory of text and tunes
in songs, and we illustrate how lyrics can be distinguished from melody in
singing, in the case of brain damage. We report a new case, G.D., who has
a severe speech disorder, marked by phonemic errors and stuttering, with-
out a concomitant musical production disorder. G.D. was found to pro-
duce as few intelligible words in speaking as in singing familiar songs. Sing-
ing “la, la, la” was intact and hence could not account for the speech deficit
observed in singing. The results indicate that verbal production, be it sung
or spoken, is mediated by the same (impaired) language output system and
that this speech route is distinct from the (spared) melodic route. In sum,
we provide here further evidence that the autonomy of music and language
processing extends to production tasks.
S INGING constitutes the most widespread mode of musical expression. All

individuals across cultures have taken part in singing in some form.
Address correspondence to Isabelle Peretz, Department of Psychology, University of

Montreal, B.O. 6128, succ. centre-ville, Montreal (Que) Canada H3C 3J7. (e-mail:
[email protected])
ISSN: 0730-7829. Send requests for permission to reprint to Rights and Permissions,
University of California Press, 2000 Center St., Ste. 303, Berkeley, CA 94704-1223.
373

374 Isabelle Peretz, Lise Gagnon, Sylvie Hébert, & Joël Macoir
This pleasurable experience is most likely rooted in the early exposure to

maternal singing, which is swiftly imitated by the infant. Infants spontane-
ously sing around the age of 1 year old. At 18 months, the child begins to
generate recognizable, repeatable songs (Ostwald, 1973). These spontane-
ous songs have a systematic form and display two essential features of
adult singing: they use discrete pitches, and they use the repetition of rhyth-
mic and melodic contours. They are unlike adult songs, however, because
they lack a stable pitch framework (Dowling, 1984). It is later, around the
age of five, that children appear to hold a stable tonality and a regular beat
as adults do (Dowling & Harwood, 1986). Thus, by the age of five, chil-
dren have a fairly large repertoire of songs of their own culture and display
singing abilities that will remain qualitatively unchanged in adulthood, unless
the child receives musical tutoring or is regularly practicing in a choir or
ensemble. Thus, even without much practice, the ordinary adult seems to
be endowed with the basic abilities that are necessary to sing simple songs
of their culture.
Despite their ubiquity and early acquisition, the singing abilities of ordi-
nary people are rarely studied. The major reason for this limited attention
is that singing abilities are considered to be unequally distributed in the
general population. This biased point of view is related to the frequent
observation that some people sing out of tune in Western cultures. Never-
theless, the few studies in which song production has been examined in the
general population have yielded a different conclusion. Nonmusicians are
highly consistent in their ability to sing familiar songs. They exhibit precise
memory for both pitch level and tempo (Halpern, 1988, 1989). This preci-
sion in singing seems to hold not only for a given singer, by measuring
individual stability across song renditions (Halpern,1988, 1989; Bergeson
& Trehub, 2002), but also for a group of singers, when measuring consis-
tency across individuals in the sung recall of a popular song (Levitin, 1994;
Levitin & Cook, 1996). Therefore, as far as singing a familiar song is con-
cerned, heterogeneity in singing abilities does not seem to be a serious limi-
tation. Moreover, an easy solution to circumvent the problem of variability
across ordinary singers is to use the singer as his/her own control, as done
in Halpern’s study and the study by Bergeson and Trehub.
This research strategy, which is centered on individual performance, is
typical of cognitive neuropsychology. Typically, the study of a neurologi-
cally impaired individual employs a within-subject design, because the ho-
mogeneity of the disorders experienced by different brain-damaged patients
cannot be assumed a priori (McCloskey, 1993). In fact, any cognitive func-
tion, such as singing, consists of numerous processing components and
pathways of communication, each of which may be damaged as a conse-
quence of a brain injury. More concretely, in our model of music process-
ing derived from the study of neurologically impaired individuals (Peretz

Singing in the Brain: Insights From Cognitive Neuropsychology 375
& Coltheart, 2003), there are 12 individuated processing components and

20 pathways. This amounts to 32 components. Thus, as pointed out by
Coltheart (2003), there are 232 ways that the system may be damaged. All
but one of these ways may correspond to a brain-damaged patient (the one
pattern that does not correspond is where all components are intact). Since
232 is a very large number, the probability of coming across two patients
with exactly the same musical impairment is very unlikely. Each patient
may be unique. Therefore, averaging across groups of patients cannot be
justified in this case. Instead, each patient must be investigated and his/her
data must be reported on an individual basis. This is why research in cog-
nitive neuropsychology usually takes the form of detailed case studies of
individuals with cognitive disorders.
Single case studies are no less valid than group studies. Valid inferences
on the functioning of normal abilities can be drawn from the study of single
cases, provided that careful consideration of the patient’s background and
sufficient data are collected and analyzed under controlled conditions. Data
from just one case are sufficient to falsify a theory. In this article, we illus-
trate this point by showing how a single case study conducted in the tradi-
tion of cognitive neuropsychology can be informative regarding the nor-
mal functional organization of singing abilities. The secondary objective is
to provide further evidence, through the testing of singing abilities, that
music and language are subserved by largely distinct processing systems.
To this aim, we will report the pattern of impaired and spared abilities
obtained in a novel case of expressive “aphasia without amusia.”
Assuming that the reader is not well-versed in neuropsychology, we will
first provide a brief review of what is the condition of “aphasia without
amusia” and what can be learned from studying such neurological deficits
about the processing of language and music in the normal brain. Thereaf-
ter, we will review the relevant literature regarding the possibility that sing-
ing is not the simple addition of the separate outputs of language and mu-
sic processing systems. In fact, classical teaching in neurology as well as
behavioral studies with normal listeners suggest that, in songs, music and
speech are integrated rather than simply aligned with each other.
Aphasia Without Amusia
Aphasia and amusia are generic terms used in neurology to designate

disorders of language and music processing, respectively. The condition of
aphasia without amusia refers to cases where a brain lesion has impaired
the processing of language without interfering with the processing of mu-
sic. One such notorious case is Vissarion Shebalin, who sustained a second
vascular hemorrhage in the left hemisphere of the brain at the age of 57.

This stroke left him without speech and deaf to the spoken world. While
Shebalin could no longer communicate verbally, he continued to teach and
to compose until his death, 4 years later. Shebalin was particularly prolific
musically despite his vast left hemispheric lesion; he wrote 14 chorales, 2
sonatas, 2 quatuors, 11 songs, and 1 symphony. According to Shostakovitch,
one of his peers, Shebalin’s music was undistinguishable from what he had
composed before his illness (Luria, Tsvetkova, & Futer, 1965). Although
Shebalin’s case is spectacular, it is not exceptional, given that similar cases
have been reported in the literature (Assal, 1973; Basso & Capitani, 1985;
Signoret, van Eeckhout, Poncet, & Castaigne, 1987).
The fact that Shebalin was an outstanding musician does not preclude
generalization to the brain of ordinary people. Drastic cases of aphasia
without amusia have been observed in nonmusicians. The latter may lose
their ability to recognize spoken words and yet remain able to recognize
music (Godefroy et al., 1995; Laignel-Lavastine & Alajouanine, 1921;
Mendez, 2001). Similarly, brain-damaged patients who are afflicted with
verbal agnosia (or word deafness), and hence have lost their ability to rec-
ognize spoken words, can maintain normal abilities to recognize nonverbal
sounds, including music (Metz-Lutz & Dahl, 1984; Takahashi et al., 1992;
Yaqub, Gascon, Al-Nosha, & Whitaker, 1988). The existence of such cases
of selective sparing of musical abilities, in both musicians and nonmusicians,
suggests that the processing of music is not mediated by the same system as
the one involved in language.
It could still be argued that music processing is spared because it is
computationally less complex, or more primitive, than language. If this
were the case, then brain damage that is sufficiently severe to interfere with
musical abilities (the simple abilities) should also be detrimental to lan-
guage abilities (the complex abilities). This account predicts that one will
not find individuals in whom brain damage has impaired the ability to
process music while sparing the ability to recognize language. Such cases of
amusia without aphasia, however, do exist. For example, Isabelle R., whom
we have studied in some detail (Peretz, Belleville, & Fontaine, 1997), is an
ordinary woman, devoid of special talents, be it musical or linguistic. She
was a restaurant manager when, at the age of 28, she underwent successive
brain surgeries for the repair of ruptured aneurysms in the left and right
middle cerebral arteries. She survived, but with two vast brain lesions in-
vading the auditory cortex bilaterally and extending to the frontal areas on
the right side. In this context, it is surprising to note that Isabelle R. is fully
functional in language, memory, and intelligence. She even writes poems.
Her persisting and major problem concerns music. Isabelle R. can no longer
recognize the music that was familiar to her before her brain accident; she
cannot relearn the musical corpus because melodies no longer leave a trace
in her memory; finally, she can no longer carry a tune. Isabelle R. regularly

practiced these skills before her brain injury, although she never studied
music. This condition of “amusia without aphasia” has been known for
more than a century (Marin & Perry, 1999). However, solid evidence has
been gathered only recently (Ayotte, Peretz, Rousseau, Bard, & Bojanowski,
2000; Griffiths et al., 1997; Peretz et al., 1994; Piccirilli, Sciarma, & Luzzi,
2000; Steinke, Cuddy, & Jakobson, 2001; Wilson & Pressing, 1999).
The major conclusion to be drawn from these neurological cases of se-
lective aphasia and of amusia is that they point to the existence of at least
two distinct series of processing modules: one for music and one for speech.
Cases such as Shebalin show how the neural circuitries specialized for mu-
sic can be selectively spared. Conversely, cases like Isabelle R. demonstrate
that these music-specific circuitries can be selectively damaged. However,
domain-specificity may arise at different levels in the processing of either
language or music. There is no need for all processing components to be
specialized for their respective domain to account for cases of selective apha-
sia and amusia. Damage to only one or two pivotal processing components
that are specialized for language (or music) may result in a dysfunction of
the entire processing system. This is why it is essential to compare music
and language processing in identical task conditions in order to be able to
specify the processing component that is responsible for the observed dis-
sociation.
Such conditions are ideally met in the processing of text and melody in
songs. The tune and the text of a song are typically heard and learned
together. Accordingly, they are expected to leave memory traces that are
equally familiar and recoverable. Hence, selective disruption of one song
component after brain damage would provide evidence that the processing
of tune and text recruit distinct mechanisms. Such demonstrations do ex-
ist. As alluded to previously, there are several reports of patients who can
no longer recognize the melodies (presented without words) of familiar
songs. Yet, they are normal at recognizing the corresponding spoken lyrics
(Hébert & Peretz, 2001; Peretz, 1996). Similarly, there are individuals who
suffer from lifelong difficulties with music and who can recognize the lyrics
of familiar songs even though unable to recognize the tune that usually
accompanies them (Ayotte, Peretz, & Hyde, 2002). These observations sug-
gest that the recognition of text and tune in songs is mediated by distinct
pathways. This division of labor may occur at the perceptual stage or, at a
later stage, when making contact with the song representation in memory.
The evidence is compatible with both possibilities.
Thus, the available evidence questions, but does not rule out, the possi-
bility that melody and lyrics are integrated in memory for songs (Serafine,
Crowder, & Repp, 1984; Serafine, Davidson, Crowder, & Repp, 1986).
The few researchers who have addressed this question with neurologically
impaired patients have reached mixed results. Samson and Zatorre (1991)

have tested song recognition in groups of patients having sustained a uni-

lateral lobectomy for the relief of epilepsy. They found that text recogni-
tion was impaired after a left lobectomy. However, they also found that
tune recognition depended on the text with which it was originally paired.
These findings led the authors to propose that melody would be integrated
to lyrics in memory, but that lyrics would benefit from an additional iso-
lated representation. In other words, there would be a dual code for songs
in memory: one song record in which melody and lyrics are integrated and
another text record in which only the text of the song would be stored.
More recently, Steinke et al. (2001) have discovered a patient with amu-
sia, K.B., whose results also suggest the existence of a special store for
songs. They found that the recognition of familiar song melodies (presented
without lyrics) could be maintained, whereas the recognition of familiar
instrumental music is lost as a consequence of brain damage. This remark-
able sparing of song recognition from hearing the melody was attributed to
the existence of intact associated representation of the text. However, the
results are also compatible with a proposal along the line of Samson and
Zatorre. That is, in addition to a song store in which melody and lyrics
would be integrated, there would be another store in which only instru-
mental music is represented. Hence K.B.’s problem would consist in access-
ing both stores from musical input alone; however, songs would be some-
what spared because they also contain embedded text information. Singing
from memory should throw light on this issue. If indeed lyrics and melo-
dies are integrated in memory, K.B., for example, should be able to sing
melodies with their accompanying lyrics but not the melodies alone. Un-
fortunately, K.B.’s singing abilities could not be tested because he was no
longer able to sing. Singing songs is the issue to which we now turn.
Is Singing Special ?
A classical teaching in clinical neurology is that nonfluent aphasics can

sing words that they cannot pronounce otherwise. This notion goes back
to the discovery of aphasia when Broca (1861) first studied his benchmark
case of expressive aphasia. Broca observed that the patient (Tan) could
pronounce only the word “tan” but was able to produce intelligible words
when singing. This description has been reported in several subsequent
cases (Assal, Buttet, & Javet, 1977; Smith, 1966; Yamadori, Osumi,
Masuhara, & Okubo, 1977). The traditional explanation of this fascinat-
ing observation is that there would be two routes to word articulation: a
normal language-based route via the left hemisphere of the brain and a
singing route via the right hemisphere (Cadalbert, Landis, Regard, & Graves,
1994). This account of word articulation is very similar to Samson and

Zatorre’s dual-store account of song memory. The singing route would

involve the store in which words are embedded in melodies and the lan-
guage route would involve the store in which words are represented in
isolation.
Partial support for the existence of this dual route has been recently
obtained with brain imaging techniques (Jeffries, Fritz, & Braun, 2003).
Normal participants have been scanned while speaking or singing the words
to a familiar song. The results showed an increase in activity during speak-
ing relative to singing in left hemisphere structures (including the classical
perisylvian language areas). Relative increases in activity during singing
(versus speaking) were observed in the right hemisphere (being maximal in
the right anterior superior temporal gyrus and insula). What is unclear is to
what extent this latter effect is due to the musical component of the task or
to the use of a special route for singing. The study does not allow a conclu-
sion to be drawn on this point because the melodic control condition (sing-
ing “la, la, la”) was not included in the design.
If there were a special singing route, the examination of singing perfor-
mance in patients such as Tan, who has Broca’s aphasia, should provide the
relevant evidence. Unfortunately, all the early studies (Assal et al., 1977;
Smith, 1966; Yamadori et al., 1977) are merely descriptive; the reports are
not substantiated by quantitative data and do not include adequate control
conditions. Furthermore, the only study in which aphasics’ production has
been quantified does not support the notion of a special singing route for
speech (Cohen & Ford, 1995). Aphasic patients were found to produce
more, not less, intelligible words when pronouncing the text in a speaking
mode than when singing them. However, the evidence is not conclusive.
Patients were not screened for type of aphasia or for the presence of amu-
sia. Thus, patients’ group performance might be the result of variable per-
ceptual and expressive deficits.
Recently, we revisited this issue with a nonfluent aphasic who did not
have any concomitant musical disorder (Hébert, Racette, Gagnon, & Peretz,
2003). We found no support for the idea that singing promotes word intel-
ligibility. The patient, C.C., produced as few intelligible words in speaking
as in singing. He made similar errors whether the text was coming from a
familiar song or from an unfamiliar one. Singing “la, la, la” was intact in
all conditions and hence could not account for the speech deficit observed
in singing. Rather, the results indicate that verbal production, be it sung or
spoken, is mediated by the same (impaired) language output system and
that this speech route is distinct from the (spared) melodic route.
This is not an isolated finding. In this article, we report a new converg-
ing case (G.D.) of nonfluent aphasia who was able to sing accurately but
was unable to articulate words in singing any more intelligibly than in
speaking.

Case Study
CASE DESCRIPTION
G.D. is a 74-year-old, left-handed man. He is a native speaker of French,

has a grade 12 level of education, and worked as an office manager. He
had no musical education. He was referred to the Sherbrooke Geriatric
University Institute because of cognitive problems, mainly involving lan-
guage impairments. According to his wife, language problems started about
10 years before his examination. Computed tomography scans obtained
in June 1994 and June 1996 revealed symmetric calcifications at the level
of the basal ganglia and the cerebellar nuclei, as well as cortical atrophy of
the left frontoparietal region. The patient had no history of cerebrovascu-
lar accident and was in good health, without medication. Primary progres-
sive aphasia was diagnosed, a frontotemporal form of dementia (see be-
low).
Neuropsychological Assessment
A summary of G.D.’s cognitive functioning is available in Table 1. At the
time of testing (April 1997), G.D. presented signs of general cognitive de-
cline, especially in reasoning abilities (executive functions). However,
memory and intellectual functions (as assessed by the Raven progressive
matrices) were still intact.
TABLE 1
Intelligence and Memory Assessments of G.D.
Cognitive Functions Score
Memory
Working memory : digit span (forward) 5
Long term memory : visuoverbal recognition
(Signoret, 1991)
Immediate recall 23/24
Delayed recall 24/24
Praxis
Gestures Normal
Copy of Rey figure 23/36 (impaired)
Visual functions
Discrimination 10/10
Benton’s line orientation test 13/15
Spatial attention test 31/35
Executive functions
Trail making test 25th percentile
(impaired)
Raven progressive matrices 50th percentile

Language Assessment
Language expression was nonfluent. Articulation was preserved but
marked by severe stuttering. Stuttering was characterized by repetitions of
initial syllables of words, often distorting them (e.g., “animaux” becomes
“zaminjo”). Expressive language also contained neologisms (e.g., “ban-
dit” becomes “espiette”) and incomplete responses (e.g., “il n’a pas de …
et lui aussi”/”he has no . . . and he too”). Automatic speech was correct for
digits and months but poor for familiar lyrics of songs due to stuttering
and phonemic transformations. G.D.’s speech was both hard to understand
and irritating. Comprehension was largely preserved. A summary of G.D.’s
language functioning is given in Table 2. His comprehension and repetition
difficulties were generally related to sentence length. The more syllables
there were in words or words in sentences, the more impaired G.D. was.
This length effect was apparent in oral repetition, in reading and in com-
prehension, hence suggesting a common phonemic output problem. Over-
all, performance showed a discrepancy between impaired expressive and
spared receptive language abilities.
Musical Assessment
G.D.’s singing abilities are intact. When prompted with 35 song titles,
G.D. eagerly sang the first line of 26 of them. The melody is remarkably in
TABLE 2
G.D.’s Performance on Language Tests
Tests G.D. Normals’ Mean (S.D.)
Expression
Boston Naming Test (Kaplan, Goodglass, & 7/60 47
Weintraub, 1983)
MT-86 ß Aphasia Battery (Béland, Lecours,
Giroux, & Bois, 1993)
Automatic speech 1/3 3
Verbal fluency 3 words 23 (5.4)
Repetition 27/30 29 (1)
Oral reading 26/30 29 (0.9)
Oral picture description 3/18 13 (3)
Discrimination and comprehension
MT-86 ß Aphasia Battery
Word and sentence picture-matching 35/47 45 (2)
Body-part identification under oral instruction 8/8 8
Token test 22.5/36 29-36
Tasks adapted from PALPA (Kay, Lesser, &
Coltheart, 1992)
Same-different discrimination 40/40 39 (1.6)
Auditory lexical (word vs. nonword) decision 50/60 57 (3.5)

tune and recognizable. The text is often unintelligible. Sometimes G.D.

manages to articulate the first words correctly but then stutters or switches
to “la, la, la.” A video example can be seen on our web site (www.fas
.umontreal.ca/psy/iperetz.html).
In comparison to his preserved singing abilities, G.D.’s receptive musical
abilities are slightly impaired. On the Montreal Battery of Evaluation of
Amusia (MBEA; Peretz, Champod, & Hyde, 2003), G.D. was found to
perform at a lower level than seven matched control subjects (mean: 70
years of age; 13 years of education) on three of six tests. G.D. scored low in
the discrimination of pitch contour and interval changes as well as in memory
(Table 3). However, some pitch-based discrimination abilities seem to be
spared, because G.D. performed in the normal range in the discrimination
of melodies differing by a tone that is out of key. Although G.D. has some
perceptual difficulties with music, his residual pitch and temporal discrimi-
nation abilities seem sufficient to allow recognition of familiar music, which
is normal. Note that the recognition of lyrics is normal too, indicating that
G.D.’s difficulty with song lyrics is at the output stage.
Summary and Diagnosis

G.D. presented with nonfluent speech marked by phonemic errors and
word-finding difficulties along with signs of cognitive decline. This clinical
profile, along with the history of progressive worsening and the presence
of particular features such as stuttering, impaired repetition, and reading
and writing disorders is suggestive of a diagnosis of nonfluent primary pro-
TABLE 3
G.D.’s Performance on Musical Tests
Matched Control Subjects
Test G.D. Mean (Range)
Singing from memory
Melody 26/26
Text 1/26
Discrimination and memory (MBEA;
Peretz et al., 2003)
Scale 24/30 26.0 (21-29)
Contour 21/30 25.1 (23-27)
Interval 18/30 24.7 (22-27)
Rhythm 26/30 28.9 (27-30)
Meter 24/30 26.7 (21-30)
Incidental memory 18/30 25.4 (22-27)
Familiarity decision
Instrumental melody 19/20
Spoken lyrics 18/20

gressive aphasia, one of the prototypical clinical syndromes of frontotem-

poral lobar degeneration (Neary et al., 1998).
EXPERIMENTAL INVESTIGATION OF SINGING AND SPEAKING
The first song line of the 30 most familiar songs in Quebec (Peretz, Babaï,
Lussier, Hébert, & Gagnon, 1995) were either sung with words, sung as
“la, la, la,” or spoken by the experimenter. G.D. was instructed to repeat
the line immediately after hearing it, in the same expressive mode. Upon
request, the trial was repeated. G.D. was allowed to make as many at-
tempts as he wished. Therefore, G.D. was tested across several sessions,
each time in a single expressive mode. That is, spoken and singing modes
of expression were not mixed in a given session so as to avoid mode-switch-
ing errors. Indeed, G.D.’s tendency was always to sing, not to speak. Each
session was recorded with a Sony tape recorder.
Song lines comprised 10 notes/syllables on average (range, 6-15). As is
normally the case, song lyrics were presented at a faster rate when spoken
(mean duration: 2257 ms; 3.8 syllables/s) than when sung (mean duration:
4351 ms; 2.4 syllables/s).
All productions were saved in a digital format. G.D.’s performance was
highly variable from one repetition attempt to the next, the first attempt
frequently being the best one. Therefore, only G.D.’s best performance for
each song line in each version was submitted to analyses. Two speech thera-
pists independently scored the text. Two musically trained judges indepen-
dently scored the melody. For text, the percentage of correctly articulated
syllables was calculated instead of words because G.D.’s errors are syllabic.
Criteria for considering a syllable as correct were lenient: if one syllable
matched the syllable of the original word in any order, it was considered
correct. One point was withdrawn for each additional syllable produced.
That is, if the total number of syllables exceeded the number of possible
syllables by, say, 5 syllables (as in the spoken rendition represented in Fig-
ure 1), then this number was subtracted from the number of correctly re-
produced syllables (thus, in the example, G.D. produced 5 correct syllables
– 5 additional syllables, hence obtaining a score of 0). For melodies, the
percentage of correctly repeated notes was computed. Out-of-tune notes
were considered as mistakes. One point was withdrawn for each additional
note and for rhythmic mistakes.
Results
Interrater agreements were calculated separately for the syllables and
notes for each song. For the syllables in the spoken and sung versions, the
interrater correlation was r (58) = .95, p < .01. The 20 syllables that were

Fig. 1. Notation and scoring of G.D.’s repetition of the first line of the song “Ce n’est qu’un
au revoir.” The model represents the sung production of the experimenter; Below is the best
sung reproduction with words of G.D. and the best spoken reproduction of G.D. The scor-
ing in terms of syllables and notes are presented in percentages. The auditory files can be
downloaded from our web site (www.fas.umontreal.ca/psy/iperetz.html).
not rated similarly were discarded from the analysis. For the notes in the
sung versions with and without words, the interrater correlation was r (58)
= .96, p < .01. The 28 notes that were not rated similarly were also dis-
carded from the analysis. The precise numbers of syllables and notes con-
sidered in the analysis are presented in Table 4.
The percentage of correctly repeated syllables and notes, for each ex-
cerpt in each condition, served as dependent variables. The data (see Table
4) were submitted to an ANOVA taking songs as the random variable, and
Component (text vs. melody) and Condition (isolated vs. combined) as
within-songs variables. The only effect to reach significance was the song
Component, with F (1, 29) = 54.6, p < .001, indicating that performance
TABLE 4
G.D.’s Percentage of Correct Note and Syllable Repetition of
30 Beginnings of Familiar Songs Presented in Combination (By
Singing) or in Isolation (Spoken or Sung La, La, La)*
Song Presentation Text Melody
Combined (sung) 65 93
(301) (293)
Isolated 58 96
(305) (284)
* Raw numbers in parentheses.

on melody was much higher than performance on text. The slightly supe-
rior performance observed in text singing over speaking was far from sig-
nificant, t(29) = 1.01. Thus, singing did not help G.D. to articulate syl-
lables in any systematic fashion, although G.D. tended to produce less
supernumerary syllables while singing (with 5%) than speaking (16%; χ2 =
21.4, with 1 d.f., p < .001; see Figure 1). Singing tends to constrain G.D.’s
speech. However, the nature of the errors was similar in both expression
modes (see Table 5)
In general, G.D.’s speech errors are typical of acquired neurological stut-
tering disorder. G.D. involuntarily repeated uttered syllables in both sing-
ing and speaking, a behavior similar to what was observed in spontaneous
speech. G.D. produced similar phonemic errors (e.g., “Au clair de la lune”
becomes “Au kair de la lune”) and neologisms (e.g., “D’où viens-tu bergère”
becomes “D’où viens-tu vieiarchère”), as illustrated in Figure 1 and quan-
tified in Table 5.
CASE DISCUSSION
G.D. has dementia with primary progressive aphasia and preserved mu-
sical expression. In the present study, we show that his loss of speech af-
fected speaking and singing in a similar fashion. This observation argues
against the notion that singing enhances speech fluency. It also questions
the claim that stuttering can be alleviated by singing (Bloodstein, 1995).
G.D.’s spoken and sung production was marked by stuttering, a condi-
tion that can supposedly be improved by singing. Singing is usually de-
scribed as one of the conditions that enhance fluency and word intelligibil-
ity in patients with developmental stuttering (Healey, Mallard, & Adams,
1976). However, cases of acquired stuttering, which occurs as a conse-
quence of brain damage, as in G.D., are less common and the evidence is
less clear. Some have reported fluent singing (Horner & Massey, 1983)
while some have not (Helm, Butler, & Benson, 1978). The evidence is weak-
ened by the fact that these observations are anecdotal and obtained in poorly
controlled conditions. This is not the case with G.D., who suffers from
acquired stuttering with intact singing. Therefore, G.D. can be considered
as the first documented patient with acquired stuttering whose speech is
TABLE 5
Speech Error Types in Singing and Speaking Expressed as
Percentages
Repetition Phonemic Neologisms
Sung 37 43 20
Spoken 25 71 4

examined while singing. As stated previously, this case study does not sup-
port the notion that singing improves speech fluency. Rather, it argues for
the complete autonomy of speech and music in production.
The idea that musical abilities are autonomous from language abilities,
and generally isolable from other cognitive functions, is also given as a
prime characteristic of dementia. Demented people are typically described
as musically apt while being severely compromised by the disease in their
general cognitive functioning (Brontons, 2000). Some demented patients
can even learn new songs while being unable to learn how to find their
own room in the geriatric care unit (Beatty et al., 1988). G.D.’s general
behavior and neurological condition certainly fit this broad characteriza-
tion. One unexpected outcome of the present investigation was that G.D.
stopped his constant mumbling and started humming instead, to the relief
of his wife.
General Discussion
The pattern of performance observed in G.D. supports the model by

which sung text is governed by the language processing system that medi-
ates normal speech. In G.D., the disruption of speech planning procedures
impaired both singing and speaking in a similar manner. Moreover, speech
disruption left melodic expression intact. Hence, G.D.’s singing performance
provides further evidence for the independence between language and mu-
sic processing. With this new case, the current evidence is quite consistent
in arguing for the complete separation of music and language processors,
including in songs where text and melody are so tightly combined.
To our knowledge, all sources of evidence, from behavioral studies to
functional neuroimaging studies, point to the autonomy of language and
music processors despite the fact that the two functions seem to engage
similar processing (Patel, 2003). Even the team of researchers (Serafine et
al., 1984, 1986) who were the first to challenge this modular view con-
cluded that text and melody are separable in song memory (Crowder,
Serafine, & Repp, 1990). In the final experiment of their series, Crowder
and collaborators (1990, Expt. 3) presented spoken texts accompanied by
hummed melodies. In these “divided” songs, lyrics and melodies were con-
nected (in time) but retained their physical independence. Nevertheless,
participants recognized the melody better when it was paired with the
matched studied text (true old pair) than when it was paired with another
old text (mismatched pair). The presence of superior recognition for the
old pairing over the mismatched one cannot be attributed to the retention
of some integrated representation of melody and text because the presenta-
tion of each was physically separate. Therefore, the superior retention of

the original combination of text and melody is best understood as the re-
sult of the memorization of contingent links between separate memory
traces.
This conclusion is also consistent with recent research showing the inde-
pendence of melody and lyrics in divided attention tasks. For instance, lis-
teners are able to monitor operatic songs for the presence of semantic and
tonal incongruities, independently. Supportive evidence has been obtained
in the on-line electrophysiological recording of brain responses (Besson,
Faïta, Peretz, Bonnel, & Requin, 1998) and in the analysis of behavioral
responses with measures derived from the signal detection theory (Bonnel,
Faïta, Peretz, & Besson, 2001).
However, as suggested in the introduction of this article, isolation of a
cognitive process of interest is not trivial, and much of the current func-
tional neuroimaging research is plagued by this problem. The method of
cognitive neuropsychology can be more powerful, as illustrated here. The
reason why neurological disorders are so informative regarding the func-
tioning of the brain is that the ensuing deficits often expose the inner work-
ings of a complex system more clearly than with any other method
(McCloskey, 2001). Most human activities rely on the operation of highly
complex systems. The apparent ease and speed with which these abilities
are normally carried out often make it difficult to uncover the underlying
mechanisms. By breaking down this smooth functioning, brain damage
renders the complex system more amenable to investigation.
Studying abnormalities or malfunctions to shed light on normal struc-
ture and processes is a well-established research strategy in health sciences.
For example, research on AIDS has contributed to knowledge of the func-
tioning of the normal immune system. Similarly, research on singing per-
formance in aphasics contributes to the understanding of normal singing
and speaking abilities. Unlike most diseases, however, homogeneity of the
cognitive deficits after brain damage cannot be assumed a priori. This is
why cognitive neuropsychology researchers have developed a methodologi-
cal expertise in the study of single cases. This does not mean that single
case data are less valuable or more subject to artifacts than group data. The
data of a single case, such as G.D.’s, are as valid as any other method used
in cognitive psychology to falsify a theory, because all brain-damaged pa-
tients with the same amount of musical experience are assumed to share
the same architecture of the system before brain damage. Uniformity of
functioning across people is assumed to be as true in cognitive neuropsy-
chology as it is in music cognition and in the cognitive sciences in general.
Thus, to challenge the present findings supporting the existence of two
distinct pathways for singing the tune and the text of songs, one would
need to either argue that the uniformity assumption has been violated in
G.D. or to show that there was a methodological problem in our study.

To conclude, a wealth of information can be obtained by studying sing-

ing performance in ordinary people. Work on aphasics and amusics is ex-
emplary in this regard because neurologically impaired individuals can serve
as their own control. Nevertheless, the study of singing abilities should not
be limited to special populations. Singing abilities are basic skills and not
the privilege of a select few individuals. Unless a given individual is born
with a musical deficiency (e.g., Peretz & Hyde, 2003), humans are equipped
with the necessary mechanisms to enable singing.1
References
Assal, G. (1973). Wernicke’s aphasia without amusia in a pianist. Revue Neurologique
(Paris), 129, 251–255.
Assal, G., Buttet, J., & Javet, R. C. (1977). Aptitudes musicales chez les aphasiques. Revue
Medicale de la Suisse Romande, 97, 5–12.
Ayotte, J., Peretz, I., & Hyde, K. (2002). Congenital amusia : A group study of adults
afflicted with a music-specific disorder. Brain, 125, 238–251.
Ayotte, J., Peretz, I., Rousseau, I., Bard, C., & Bojanowski, M. (2000). Patterns of music
agnosia associated with middle cerebral artery infarcts. Brain, 123, 1926–1938.
Basso, A., & Capitani, E. (1985). Spared musical abilities in a conductor with global apha-
sia and ideomotor apraxia. Journal of Neurology, Neurosurgery & Psychiatry, 48, 407–
412.
Beatty, W. W., Zavadil, K. D., Bailly, R. C., Rixen, G. J., Zavadil, L. E., Farnham, N., et al.
(1988). Preserved musical skill in a severely demented patient. International Journal of
Clinical Neuropsychology, 10, 158–164.
Béland, R., Lecours, A. R., Giroux, F., & Bois, M. (1993). The MT-86 ß Aphasia Battery: a
subset of normative data in relation to age and level of school education. Aphasiology, 7,
359–382.
Bergeson, T. R., & Trehub, S. E. (2002). Absolute pitch and tempo in mother’s songs to
infants. Psychological Science, 13, 72–75.
Besson, M., Faïta, F., Peretz, I., Bonnel, A.-M., & Requin, J. (1998). Singing in the brain:
Independence of lyrics and tunes. Psychological Science, 9, 494–498.
Bloodstein, O. (1995). A handbook on stuttering. New York: Chapman & Hall.
Bonnel, A.-M., Faïta, F., Peretz, I., & Besson, M. (2001). Divided attention between lyrics
and tunes in operatic songs: Evidence for independent processing. Perception & Psycho-
physics, 67, 1201–1213.
Broca, P. (1861). Remarques sur le siège de la faculté du langage articulé, suivies d’une
observation d’aphémie. Bulletin et Mémoires de la Société Anatomique de Paris, 2, 330–
357.
Brontons, M. (2000). An overview of the music therapy literature relating to elderly people.
In D. Aldridge (Ed.), Music therapy in dementia care: More new voices. London: J.
Kingsley.
Cadalbert, A., Landis, T., Regard, M., & Graves, R. E. (1994). Singing with and without
words: hemispheric asymmetries in motor control. Journal of Clinical and Experimental
Neuropsychology, 16, 664–670.
1. We thank Aniruddh Patel for insightful comments, and G.D and his wife for their
collaboration. At the time of testing, G.D. was autonomous. Today, he lives in a nursing
home. The work was supported by a grant from the International Human Frontier Science
Program (I.P.) and from the Natural Science and Engineering Research Council of Canada
(S.H.). L.G., J.M. and S.H. are all supported by fellowships from the Fonds de la Recherche
en Santé du Québec.

Cohen, N. S., & Ford, J. (1995). The effects of musical cues on the nonpurposive speech of
persons with aphasia. Journal of Music Therapy, 32, 46–57.
Coltheart, M. (2003). Cognitive neuropsychology. In J. Wixted (Ed.), Stevens’ handbook of
experimental psychology (Vol. 4). New York: John Wiley & Sons.
Crowder, R. G., Serafine, M. L., & Repp, B. (1990). Physical interaction and association by
contiguity in memory for the words and melodies of songs. Memory & Cognition, 18,
469–476.
Dowling, W. J. (1984). Development of musical schemata in children’s spontaneous sing-
ing. In W. R. Crozier & A. J. Chapman (Eds.), Cognitive processes in the perception of
art (pp. 145–163). Amsterdam: North-Holland.
Dowling, W. J., & Harwood, D. (1986). Music cognition. New York: Academic Press.
Godefroy, O., Leys, D., Furby, A., De Reuck, J., Daems, C., Rondepierre, P., et al. (1995).
Psychoacoustical deficits related to bilateral subcortical hemorrhages: A case with ap-
perceptive auditory agnosia. Cortex, 31, 149–159.
Griffiths, T. D., Rees, A., Witton, C., Cross, P. M., Shakir, R. A., & Green, G. G. (1997).
Spatial and temporal auditory processing deficits following right hemisphere infarction:
A psychophysical study. Brain, 120, 785–794.
Halpern, A. R. (1988). Perceived and imagined tempos of familiar songs. Music Perception,
6, 193–202.
Halpern, A. R. (1989). Memory for the absolute pitch of familiar songs. Memory & Cogni-
tion, 17, 572–581.
Healey, E. C., Mallard, A. R. III, & Adams, M. R. (1976). Factors contributing to the reduc-
tion of stuttering during singing. Journal of Speech & Hearing Research, 19, 475–480.
Hébert, S., & Peretz, I. (2001). Are text and tune of familiar songs separable by brain
damage? Brain and Cognition, 46(1-2), 169–175.
Hébert, S., Racette, A., Gagnon, L., & Peretz, I. (2003). Revisiting the dissociation between
singing and speaking in expressive aphasia. Brain, 126, 1838–1850.
Helm, N. A., Butler, R. B., & Benson, D. F. (1978). Acquired stuttering. Neurology, 28,
1159–1165.
Horner, J., & Massey, E. W. (1983). Progressive dysfluency associated with right hemi-
sphere disease. Brain & Language, 18, 71–85.
Jeffries, K. J., Fritz, J. B., & Braun, A. R. (2003). Words in melody: an H2O15 PET study of
brain activation during singing and speaking. NeuroReport, 14, 749–754.
Kaplan, E., Goodglass, H., & Weintraub, S. (1983). Boston Naming Test. Malvern, PA: Lea
& Febiger.
Kay, J., Lesser, J., & Coltheart, M. (1992). Psycholinguistic assessments of language pro-
cessing in aphasia. Hove, England: Psychology Press.
Laignel-Lavastine, M., & Alajouanine, T. (1921). Un cas d’agnosie auditive. Société de
Neurologie, 37, 194–198.
Levitin, D. J. (1994). Absolute memory for musical pitch: evidence from the production of
learned melodies. Perception & Psychophysics, 56, 414–423.
Levitin, D. J., & Cook, P. R. (1996). Memory for musical tempo: additional evidence that
auditory memory is absolute. Perception & Psychophysics, 58, 927–935.
Luria, A. R., Tsvetkova, L. S., & Futer, D. S. (1965). Aphasia in a composer (V. G. Shebalin).
Journal of Neurology Sciences, 2, 288–292.
Marin, O. S. M., & Perry, D. W. (1999). Neurological aspects of music perception and
performance. In D. Deutsch (Ed.), The psychology of music (2nd ed., pp. 653–724).
New York: Academic Press.
McCloskey, M. (1993). Theory and evidence in cognitive neuropsychology : A “radical”
response to Robertson, Knight, Rafal, and Shimamura (1993). Journal of Experimental
Psychology : Learning, Memory, and Cognition, 19, 718–734.
McCloskey, M. (2001). The future of cognitive neuropsychology. In B. Repp (Ed.), The
handbook of cognitive neuropsychology: What deficits reveal about the human mind
(pp. 593–610). Philadelphia: Psychology Press.
Mendez, M. (2001). Generalized auditory agnosia with spared music recognition in a left-
hander: Analysis of a case with a right temporal stroke. Cortex, 37, 139–150.

Metz-Lutz, M. N., & Dahl, E. (1984). Analysis of word comprehension in a case of pure
word deafness. Brain and Language, 23, 13–25.
Neary, D., Snowdon, J. S., Gustafson, L., Passant, U., Stuss, D., Black, S., et al. (1998).
Frontotemporal lobar degeneration: A consensus on clinical diagnostic criteria. Neurol-
ogy, 51, 1546–1554.
Ostwald, P. F. (1973). Musical behavior in early childhood. Developmental Medicine &
Child Neurology, 15, 367–375.
Patel, A. (2003). Language, music, syntax and the brain. Nature Neuroscience, 6, 674–681.
Peretz, I. (1996). Can we lose memories for music? The case of music agnosia in a
nonmusician. Journal of Cognitive Neurosciences, 8, 481–496.
Peretz, I., Babaï, M., Lussier, I., Hébert, S., & Gagnon, L. (1995). Corpus d’extraits musicaux:
indices relatifs à la familiarité, à l’âge d’acquisition et aux évocations verbales. Canadian
Journal of Experimental Psychology, 49, 211–239.
Peretz, I., Belleville, S., & Fontaine, S. (1997). Dissociations entre musique et langage après
atteinte cérébrale: un nouveau cas d’amusie sans aphasie. Canadian Journal of Experi-
mental Psychology, 51, 354–368.
Peretz, I., Champod, A.-S., & Hyde, K. L. (2003). Varieties of musical disorders: The Montreal
Battery of Evaluation of Amusia. Annals of the New York Academy of Sciences, in press.
Peretz, I., & Coltheart, M. (2003). Modularity of music processing. Nature Neuroscience,
6, 688–691.
Peretz, I., & Hyde, K. (2003). What is specific to music processing? Insights from congeni-
tal amusia. Trends in Cognitive Sciences, 7, 362–367.
Peretz, I., Kolinsky, R., Tramo, M., Labrecque, R., Hublet, C., Demeurisse, G., et al. (1994).
Functional dissociations following bilateral lesions of auditory cortex. Brain, 117, 1283–
1301.
Piccirilli, M., Sciarma, T., & Luzzi, S. (2000). Modularity of music : Evidence from a case of
pure amusia. Journal of Neurology, Neurosurgery & Psychiatry, 69, 541–545.
Samson, S., & Zatorre, R. J. (1991). Recognition memory for text and melody of songs
after unilateral temporal lobe lesion: Evidence for dual encoding. Journal of Experimen-
tal Psychology : Learning, Memory and Cognition, 17, 793–804.
Serafine, M. L., Crowder, R. G., & Repp, B. H. (1984). Integration of melody and text in
memory for songs. Cognition, 16, 285–303.
Serafine, M. L., Davidson, R. J., Crowder, R. G., & Repp, B. H. (1986). On the nature of
melody: Text integration in memory for songs. Journal of Memory and Language, 25,
123–135.
Signoret, J. L. (1991). Batterie d’efficience mnésique: BEM144. Paris: Elsevier.
Signoret, J. L., van Eeckhout, P., Poncet, M., & Castaigne, P. (1987). [Aphasia without
amusia in a blind organist: Verbal alexia-agraphia without musical alexia-agraphia in
braille]. Revue Neurologique (Paris), 143, 172–181.
Smith, A. (1966). Speech and other functions after left (dominant) hemispherectomy. Jour-
nal of Neurology, Neurosurgery and Psychiatry, 29, 467–471.
Steinke, W. R., Cuddy, L. L., & Jakobson, L. S. (2001). Dissociations among functional
subsystems governing melody recognition after right-hemisphere damage. Cognitive
Neuropsychology, 18, 411–437.
Takahashi, N., Kawamura, M., Shinotou, H., Hirayama, K., Kaga, K., & Shindo, M. (1992).
Pure word deafness due to left hemisphere damage. Cortex, 28, 295–303.
Wilson, S. J., & Pressing, J. (1999). Neuropsychological assessment and the modeling of
musical deficits. In R. R. Pratt & D. Erdonmez Grocke (Eds.), Music medicine and music
therapy: Expanding horizons (pp. 47–74). Melbourne: The University of Melbourne.
Yamadori, A., Osumi, Y., Masuhara, S., & Okubo, M. (1977). Preservation of singing in
Broca’s aphasia. Journal of Neurology Neurosurgery & Psychiatry, 40, 221–224.
Yaqub, B. A., Gascon, G. G., Al-Nosha, M., & Whitaker, H. (1988). Pure word deafness
(acquired verbal auditory agnosia) in an Arabic speaking patient. Brain, 111, 457–466.


Neurología y Canto

Uploaded by

Copyright:

Available Formats

Neurología y Canto

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Neurología y Canto

Uploaded by

Copyright:

Available Formats

Singing in the Brain: Insights from Cognitive Neuropsychology

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

Singing in the Brain: Insights from Cognitive

S INGING constitutes the most widespread mode of musical expression. All

Address correspondence to Isabelle Peretz, Department of Psychology, University of

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This pleasurable experience is most likely rooted in the early exposure to

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

& Coltheart, 2003), there are 12 individuated processing components and

Aphasia Without Amusia

Aphasia and amusia are generic terms used in neurology to designate

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

have tested song recognition in groups of patients having sustained a uni-

A classical teaching in clinical neurology is that nonfluent aphasics can

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

Zatorre’s dual-store account of song memory. The singing route would

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

G.D. is a 74-year-old, left-handed man. He is a native speaker of French,

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

tune and recognizable. The text is often unintelligible. Sometimes G.D.

Summary and Diagnosis

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

gressive aphasia, one of the prototypical clinical syndromes of frontotem-

EXPERIMENTAL INVESTIGATION OF SINGING AND SPEAKING

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

* Raw numbers in parentheses.

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

The pattern of performance observed in G.D. supports the model by

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

To conclude, a wealth of information can be obtained by studying sing-

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

This content downloaded from 130.179.16.201 on Wed, 24 Jun 2015 23:51:54 PM

You might also like