SEMINARS IN NEUROLOGY—VOLUME 20, NO. 4 2000
Aphasia and the Relationship of Language and Brain
Eleanor M. Saffran, Ph.D.
ABSTRACT
This article provides a brief history of the study of aphasia, along with current data
on aphasic syndromes and their localization in the brain as well as information on
testing procedures. A brief examination of language from a functional perspective is
also provided, as it is difficult to understand the breakdown patterns without an appreciation for the complexities of language processing per se.
Keywords: Aphasia, language disorders, brain damage
A HISTORICAL PERSPECTIVE
Interest in the relationship of language and brain
began with the emergence of a scientific mind set in the
19th century. The question was first posed by phrenologists—practitioners of the pseudoscience that associated
bony protrusions in the skull with the enlargement of
brain tissue below and hence the development of particular capacities. Although most of the traits that interested phrenologists would strike us as fanciful today
(e.g., parental love, conjugality), they did attempt to localize language. They assigned this function to the
frontal poles of both hemispheres and assessed it by
measuring the size of the bony protrusion below the eye.
The young French physician and anthropologist
Paul Broca was skeptical of this approach. He believed
that capacities were associated with particular convolutions of the cerebral hemispheres that were not linked in
any direct way to bumps on the skull. Broca had the opportunity to examine the brain of a language-impaired
patient, M. Leborgne, when it came to autopsy. The patient had been capable of very little speech although his
comprehension appeared well preserved. Broca observed an area of damage in the left cerebral hemisphere,
surrounding the third frontal convolution. After seeing a
number of similar cases, Broca1 postulated that this area
of the brain was the locus for articulate speech. He also
suggested that left hemisphere control of language production was associated with its control of the dominant
right hand, and he proposed mirror image organization—right hemisphere dominance—in left-handers. Although Broca was correct in his first assumption, we
Objectives
On completion of this article the reader will be able to summarize the seminal observations regarding localization of language disturbance in
the brain.
Accreditation
The Indiana University School of Medicine is accredited by the Accreditation Council for Continuing Medical Education to provide
continuing medical education for physicians.
Credit
The Indiana University School of Medicine designates this educational activity for a maximum of 1.0 hours in category one credit toward
the AMA Physicians Recognition Award. Each physician should claim only those hours of credit that he/she actually spent in the
educational activity.
Disclosure
Statements have been obtained regarding the author’s relationships with financial supporters of this activity. There is no apparent conflict
of interest related to the context of participation of the author of this article.
Center for Cognitive Neuroscience, Department of Neurology, Temple University School of Medicine, Philadelphia, Pennsylvania.
Reprint requests: Dr. Saffran, Center for Cognitive Neuroscience, Department of Neurology, Temple University
School of Medicine, 3400 North Broad Street, Philadelphia, PA 19140.
Copyright © 2000 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA.
+1(212) 584-4662. 0271-8235,p;2000,20,04,409,418,ftx,en;sin00095x.
409
SEMINARS IN NEUROLOGY VOLUME 20, NUMBER 4 2000
410
now know that the left hemisphere is dominant for language in most left-handers (approximately 70%) as well
as nearly all right-handers (on the order of 96%).2
Several years later, Carl Wernicke3 made other important observations. He saw a patient whose comprehension was severely impaired; when the patient came to
autopsy, the lesion was discovered in the posterior, superior left temporal lobe. On the basis of additional observations of similar patients, although not always backed
up by anatomical data, Wernicke hypothesized that this
area was the locus of storage of “auditory word images,”
which were necessary for the production as well as the
comprehension of speech. The patients he saw exhibited
anomalies in production along with their comprehension
disorder: they frequently produced words that did not fit
the context and generated combinations of speech
sounds that did not qualify as words in their native language (“neologisms”). Wernicke assumed, further, that
there was a fiber tract connecting this region with the
area identified by Broca and that speech production was
influenced by activity in Wernicke’s area, as this left
hemisphere brain region came to be known. He predicted that damage to the connecting tract would lead to
a speech production disorder with preserved comprehension, as well as a disorder of repetition—a prediction
later verified. That syndrome was termed “conduction
aphasia.” Although it was earlier believed that this disorder arose from interruption of the arcuate fasciculus, the
fiber tract that connects these two areas, it is now generally accepted that the syndrome can be produced by
damage at a variety of locations including the supramarginal gyrus as well as by interruption of fiber tracts lying
deep to sensory cortex in the parietal lobe.2
This approach to the relationship of brain function
and behavior, which posits centers for specific processing operations along with connections between them,
came to be known as “connectionism” (which should be
distinguished from the present usage of the term, which
refers to computational modeling). The connectionist
approach to language was further elaborated by Carl
Lichtheim,4 who posited a concept center, diffusely represented in the brain, that served as the endpoint of language comprehension as well as the source of messages
that were ultimately produced by the speaker. Lichtheim’s diagram for the language system is reproduced
in Figure 1. On the basis of this diagram, he postulated
disorders that had not yet been described, such as
transcortical sensory aphasia, in which the projection
from Wernicke’s area to the concept center is disrupted,
resulting in impaired comprehension and production,
with repetition preserved because of the patent connection between Wernicke’s and Broca’s areas.5
Although the connectionist approach appeared to
be productive and was subsequently extended to reading
disorders,6 it was criticized by later investigators who
adopted a holistic view of the function of the language
area.7,8 The holistic approach prevailed until connectionism was revived by the neurologist Norman Geschwind
in the 1960s.9,10 At around the same time, some speech
pathologists and psychologists promoted the idea that
the language area was not divisible and functioned as a
whole.11 Among neurologists, this view was subse-
Figure 1. Lichtheim’s diagram of the language system.
A, Wernicke’s area; B, concept center; M, Broca’s area;
a → A, auditory input to Wernicke’s area; M → m, motor
output from Broca’s area. A → M, tract connecting Wernicke’s and Broca’s areas; A → B, pathway essential for
understanding spoken input; B → M, pathway essential
for meaningful verbal output. Lesions: at A, Wernicke’s
aphasia; at M, Broca’s aphasia; a → A, pure word deafness; M → m, articulatory disorder (aphemia); A → M,
conduction aphasia; A → B, transcortical sensory aphasia; B → M, transcortical motor aphasia.
quently promoted by Mesulam.12 The controversy between the localizers and those who take a more holistic
approach to language function persists to this day. The
holistic camp takes the co-occurrence of symptoms—for
example, word-finding deficits and grammatical simplification in the production of patients with Alzheimer’s
disease—to reflect common, overlapping storage and retrieval mechanisms for words and syntax.13 An alternative view, favored by those who accept regions of specialization within the language system, is that one deficit
is the cause of the other: in this case, that the word retrieval deficit limits the ability to construct complex
utterances.14
With respect to the localization of functions, the
early researchers were limited by the restriction to autopsy material. The advent of imaging techniques such
as computed tomography (CT) and magnetic resonance
imaging (MRI), as well as the new methods for functional imaging (positron emission tomography [PET]
and functional magnetic resonance imaging [fMRI]),
have vastly increased the scope of these investigations.
More recent studies have demonstrated the variability15
as well as the extent of the lesions that give rise to partic-
APHASIA AND THE RELATIONSHIP OF LANGUAGE AND BRAIN—SAFFRAN
ular language disorders. For example, Broca localized
the center for articulate language at the foot of the third
frontal convolution, even though the lesion in his patient
extended well beyond this area. It has been shown that a
lesion restricted to Broca’s area gives rise to a transient
impairment of language production and that the full
complement of symptoms associated with Broca’s aphasia (articulation problem coupled with simplified sentence structure—the pattern known as “agrammatism”)
is the result of more extensive damage to the frontal cortex.16 There is evidence, moreover, that the articulation
problem present in Broca’s aphasia (“apraxia of
speech”) is associated with damage to a portion of the
insula, a part of the cerebral cortex that is not visible
from the brain’s surface because of the growth of other
parts of the frontal lobe.17 It has also become apparent
that the language area includes the zone now known as
the perisylvian cortex (the cortical area above, below,
and posterior to the sylvian fissure), which includes considerably more brain tissue than the areas identified by
Broca and Wernicke. There is also evidence that certain
types of language impairments result from damage outside the boundaries of the perisylvian cortex, a matter
we will consider below.
THE CLASSICAL SYNDROMES AND
ATTEMPTS AT QUANTIFICATION
The connectionist schema elaborated by Lichtheim4 specified a number of different syndromes, to
which additional ones have been added in recent years
(e.g., global aphasia; anomia). The classical syndrome
patterns are described in Table 1. Important terms to note
include “paraphasia,” which refers to errors in word production; these include “verbal paraphasias,” which are
substitutions of other words (e.g., “doctor” for nurse;
“horse” for house), and “literal (or phonemic) paraphasias,” which are sound substitutions (e.g., “doctin”
for doctor; “ramonica” for harmonica). (Note that the
latter qualify as neologisms, because they are not words
in English.) “Conduite d’approche” refers to the tendency, most evident in conduction aphasics, to make
repeated attempts at a word (e.g., for pretzel, “trep . . .
tretzle . . . trethle . . . tredfles . . . ki”)18 that do not necessarily result in closer approximations to the target.
These attempts indicate that the patient knows what the
word should sound like and is dissatisfied with his or her
efforts. Repeated attempts are less frequent in Wernicke’s aphasia, in which patients appear less successful
in monitoring their output. The term “agrammatism”
refers to a speech pattern in which sentence structures
are simplified and grammatical morphemes—inflections
such as the past tense marker, -ed, as well as freestanding elements such as articles, auxiliary verbs, and prepositions—are frequently omitted. “Semantic reversibility” refers to sentences such as “The boy is kissed by the
girl” or “The dog that the cat chased was black,” in
which either entity could conceivably be performing the
action on the other. Whereas many patients have no difficulty understanding semantically constrained sentences, such as “The bone was eaten by the dog,” they
are impaired in understanding sentences in which the
syntax must be used to determine which entity is the perpetrator and which the recipient of the action. “Paragrammatism” refers to structured production in which
grammatical morphology is not entirely correct, often
because one morpheme is substituted for another. To illustrate these points, Tables 2 and 3 provide examples of
the production of an agrammatic Broca’s aphasic (Table
2) and a fluent Wernicke’s aphasic (Table 3).
Figure 2 illustrates typical locations for the lesions
associated with these syndromes. It should be emphasized, however, that there is considerable variability in
these lesion sites with respect to both the cortical tissue
Table 1. Descriptions of Major Aphasic Syndromes
Syndrome
Speech
Global Aphasia
Broca’s Aphasia
Impaired
Nonfluent;
articulation
impaired
Conduction
Aphasia
Fluent,
paraphasic;
conduite
d’approche
Fluent but
paraphasic; may
be excessive
Fluent,
paraphasic
Preserved but
sparse
Fluent but
hesitant due to
poor word
retrieval
Wernicke’s
Aphasia
Transcortical
Sensory
Transcortical
Motor
Anomia
Sentence
Production
Word
Comprehension
Sentence
Comprehension
Absent
Impaired,
agrammatic
Impaired
Impaired to
relatively good
Impaired
Impaired
Well structured,
paraphasic
Good
Impaired
Structured but
empty
Impaired
Impaired
Impaired for
semantically
reversible
sentences
Variable; may be
impaired for
complex
sentences
Impaired
Impaired
Structured but
empty
Variable
Impaired
Impaired
Preserved
Good
Variable
Preserved
Variable
Relatively good
Preserved
Naming
Impaired
Impaired to
relatively
good; nouns
> verbs
Variable, but
paraphasic
Variable
Impaired
Structured but
impaired by
word finding
difficulty
Repetition
Impaired
Impaired
411
SEMINARS IN NEUROLOGY VOLUME 20, NUMBER 4 2000
Table 2. Speech Pattern of an Agrammatic Aphasic
Telling the Story of Cinderella
Long ago Cinderella. One time many years ago two sisters and
one stepmother. Cinderella is washing clothes and mop floor.
One day big party in the castle. Two girls dresses is beautiful.
Cinderella is poor. Two sisters left. In the castle Cinderella is
. . . Godmother. Oh, what’s wrong? No money. A little mouse.
Cinderella hurry. Queen. Magic wand. Mouses. Oh big men
now. Magic wand pumpkin then chariot. Cinderella dresses no
good. Cinderella. On my god beautiful now. Next time, twelve
o’clock, hex. Then Cinderella party. Many women at the party.
Prince is . . . no good. Oh, prince is . . . Cinderella. Dance
dance. Two hours. Dance dance dance. Twelve o’clock, oh my
god. Cinderella. What’s wrong? I leave, I leave, I leave. One
shoe slip. Cinderella is twelve o’clock. Poor again. Tomorrow,
prince slipper. Village many houses. One house, knock, knock.
What’s wrong. Prince is . . . One, two kids, OK, OK. First one
too small. Second was too big. Hey wait a minute, wait a
minute. Prince. You. What’s wrong. Hey, slip it on. Oh. Does fit.
Godmother. Magic wand. New gown again. Then magic prince
and godmother went in the castle.
involved and the damage to white matter underlying the
cortex.15 Isolating a cortical area by interrupting the
fiber tracts that project to it can produce functional consequences that are similar to those produced by damage
to the cortical region itself. It should also be noted that
the listing in Table 1 and the lesion site examples in Figure 2 reflect some simplification of the actual syndrome
412
Figure 2. Lesion locations for the
major aphasia syndromes.
Table 3. Speech Pattern of a Wernicke’s Aphasic
Responding to a Question about What He
Liked to Cook
I I don’t know how there is any single way, there’s so many
thing, you know, that I like. I like meats, I have liked beef, the
Germans, you know, and what, well the French you koot the
whole, I can’t recall the word that I can’t thay. It was the ———
where you make all the food, you make it all up today and keep
it till the next day. With the French, you know, uh, what is the
name of the word, God, public serpinz they talk about, uh but I
have had that, it was ryediss, just before the storage you know,
seven weeks, I had personal friends that, that, I would cook an’
food the food and serve fer four or six mean for an evening.
See Schwartz, Saffran, Bloch, and Dell (1994) for a complete
transcript).62 Words and nonwords in italics represent errors;
___ represents word that was omitted.
complexes. For example, there are classifications that
distinguish among subtypes of syndromes reflecting different lesion sites. In the case of transcortical motor
aphasia,19 two subtypes are proposed, one associated
with the dorsolateral frontal lesion shown in Figure 2
and another that results from damage to the supplementary motor cortex in the left hemisphere, reflecting infarction of the anterior as opposed to the middle cerebral
artery. They also suggest two loci for transcortical sensory aphasia, although the second type, reflecting dam-
CHRONIC BROCA'S APHASIA
WERNICKE'S APHASIA
CONDUCTION APHASIA
GLOBAL APHASIA
TRANSCORTICAL MOTOR APHASIA
TRANSCORTICAL SENSORY APHASIA
APHASIA AND THE RELATIONSHIP OF LANGUAGE AND BRAIN—SAFFRAN
age to the angular gyrus of the parietal lobe, might be described by others as “anomic aphasia.” No lesion sites
for anomic aphasia are given in Figure 2 for the reason
that this disorder can result from damage to a variety of
loci within the left hemisphere15 and is often the residual
symptom following recovery from a number of different
types of aphasia. For additional information on lesion
sites and variability in the symptomatology, see references 2 and 19.
There have been a number of attempts to develop
quantitative descriptions of these syndromes based on
data from batteries of aphasia tests. Currently, the most
widely used are the Boston Diagnostic Examination for
Aphasia (BDEA), developed by Goodglass and Kaplan,18 and the Western Aphasia Battery, developed by
Kertesz.20 Two examples of syndrome patterns from the
BDEA are illustrated in Figure 3. Although the tests offer rather strict guidelines for classifying patients, syndrome designations are applied rather broadly by practicing clinicians. Thus, for example, many utilize the
term Broca’s aphasia for nonfluent production patterns
irrespective of whether the patient demonstrates the
agrammatic speech characteristics specified in Figure 3.
As a result, the variability tolerated under this label may
render the syndrome designation too broad for the purpose of grouping patients for study.21 There is also evidence that patients may shift from one syndrome to another as they recover; for example, some patients begin
as Wernicke’s aphasics and later, with some recovery of
comprehension, are reclassified as conduction aphasics
or anomics; others are initially diagnosed as global aphasics and subsequently qualify as Broca’s aphasics. Although it is important to acknowledge shortcomings in
applying these labels, it is nevertheless the case that the
syndrome designations serve as a useful shorthand for
the major deficit patterns exhibited by aphasic patients.
The standard aphasia examination involves tests of
comprehension at the word and sentence level, word
(e.g., picture naming) and sentence production, and repetition. Although the ability to repeat what others say has
little function in adult life, this capacity is critical in differentiating certain of the aphasic syndromes from one
another. For example, the symptoms of Wernicke’s and
transcortical sensory aphasia are similar except for the
preservation of repetition in the latter. Reading and writing capacities are also examined, although they will not
be considered in this article.
LANGUAGE: A FUNCTIONAL PERSPECTIVE
To appreciate the full range of symptoms exhibited
by patients with language disorders, it is necessary to
consider language from a functional perspective—that
is, the set of processes thought to account for the comprehension and production of language. The result of investigations within the subfield of psychology known as
psycholinguistics, these processes are outlined in Figure
4. It is clear, for example, that there are specialized
mechanisms for the perception of speech, which consists of brief stimuli that change rapidly in wavelength
composition. Cutting off input to critical left temporal
areas as a result of a left hemisphere lesion, or in some
cases lesions in both hemispheres (the lesion on the
right deprives the left hemisphere of transcallosal input), results in the disorder known as “pure word deafness,” in which patients can hear but cannot understand
speech; their native tongue, for example, sounds to them
like a foreign language. Moreover, these patients have
difficulty discriminating between speech sounds (between “pa” and “ba,” for example), although they have
little or no difficulty producing speech and understanding written language.22
Hearing speech adequately is not sufficient, however; words must be recognized as such and associated
with their meanings. To examine the capacity for word
recognition, psychologists use a task termed “lexical decision” in which the subject must determine whether a
string of sounds (or letters, if visual presentation is used)
is a real word or nonword. The retrieval of word meaning is often assessed by matching a word to a picture.
The most sensitive way to construct such a test is to embed the correct picture among others that are related to
the target word (e.g., if the word is “pear”, the nonmatch
drawings would depict other fruits). The reason for this
is that patients sometimes retain some aspects of the
meaning of a word (such as its semantic category) and
therefore have difficulty distinguishing it from other category members, although they may perform very well if
the choices come from other classes. But not all words
are picturable; to examine knowledge of abstract words,
subjects may be asked to match a word to one of several
on the basis of meaning (e.g., freedom to liberty, justice,
and knowledge). Sensitivity to verb meaning should also
be tested; this can be done with pictures or by means of
the word-matching paradigm described for abstract
words.
Whereas understanding the meaning of single
words is essential to language comprehension, most
messages are expressed as sentences, where structural
considerations (grammar or syntax) are also important.
To assess syntactic capacities, it is necessary to use sentences that are semantically reversible (e.g., “The girl is
kissed by the boy”)—in other words, sentences in which
either participant could conceivably be acting upon the
other. If semantic constraints are present (as in “The
book is read by the boy”), they may be utilized to construct the correct interpretation. Thus, Broca’s aphasics
appeared to have adequate sentence comprehension
when they were tested with semantically constrained
sentences; their difficulty with sentence-level materials
did not emerge until they were tested with semantically
reversible sentences.23 It should be noted, however, that
currently available aphasia batteries are not always sensitive to the reversibility factor. Other tasks used to examine syntactic capacities include judgments of grammaticality. Patients hear sentences that are either
grammatical or contain some syntactic violation and are
asked to judge whether the sentence represents “good”
or “bad” English.24 Interestingly, patients may perform
well on the judgment task even though they are impaired
in comprehending reversible sentences (this matter is
discussed in more detail later). In addition, the information serially extracted from the sentence must be main-
413
Figure 3. (A) Deficit pattern for Broca’s aphasia. (B) Deficit pattern for Wernicke’s aphasia. (Redrawn from Goodglass H, Kaplan E. The Assessment of Aphasia and Related Disorders. 2nd ed. Philadelphia: Lea & Febiger; 1983.)
414
APHASIA AND THE RELATIONSHIP OF LANGUAGE AND BRAIN—SAFFRAN
Figure 4. Language processing
from a functional perspective.
tained in a short-term or working memory. Virtually all
aphasic patients suffer from verbal short-term limitations (as measured, for example, by asking them to repeat digit strings; their performance tends to be well below the normal span of about seven.). One particular
group with left posterior parietal lesions25 suffers from
short-term memory limitations but little else, and many
of these patients have sentence comprehension deficits
similar to those described in Broca’s aphasics.26
On the output side, production begins with an intention, initially encoded as a nonverbal message. The next
task is to retrieve words that correspond to the message,
which are first represented abstractly, as “lemmas”—that
is, not yet spelled out phonologically as sequences of
sounds. The syntactic specifications of these words are
also retrieved (i.e., whether they are nouns, verbs, adjectives, etc.; if verbs, the kinds of syntactic structures they
require, and so forth). Then a sentence structure is specified (for example, whether the voice is active or passive,
whether a relative clause is required). The sentence structure specifies the order in which the words are to appear.
Then the words are retrieved, with their sound specifications, in the order dictated by the structure of the sentence.
From this string of phonologically specified elements, a
motor speech program is assembled and ultimately sent to
the articulators. The manner in which this is accomplished
is not yet clear. One possibility is that the speaker retrieves successive syllables from a syllabary that is acquired in learning to speak the language.27 One virtue of
this theory is that it constrains the speaker to the production of sound sequences that are present in the language;
these sequences are seldom violated, even in the speech
errors of normal speakers and aphasics. It should be noted
that the motor speech program is specified not only for
content words, such as nouns, verbs, and adjectives, but
also for grammatical elements such as inflections on
verbs, auxiliary verbs, relative pronouns, and prepositions. At least some of these elements must be inserted at a
late stage of sentence production, as their form depends
on context (e.g., use of “a” or “an,” depending on whether
the following word starts with a vowel or consonant). The
grammatical features are an essential part of the prosodic
(rhythmic) structure of the utterance, as they are pronounced together with the content words that follow
or precede them (e.g., theboy istalking . . .). In English,
these elements are often omitted by agrammatic speakers
(see Table 2); in other languages, where they cannot be
deleted (as in Hebrew, where the vowels of a verb denote
tense), they are prone to error because of the occurrence
of substitutions.28
415
SEMINARS IN NEUROLOGY VOLUME 20, NUMBER 4 2000
Note, then, that both language comprehension and
production entail the retrieval of several different types
of information that must also be integrated—meanings,
word forms, syntactic specifications for words, and all of
these with syntactic structure that is derived from the
perceived sentence or sequenced for the sentence to be
uttered. One can imagine that these complex tasks require a good deal of integration across knowledge
sources, which may explain why they frequently break
down in patients with aphasia.
OPEN QUESTIONS: WHERE IS SYNTAX?
416
The finding that agrammatic Broca’s aphasics also
had a problem in sentence comprehension gave rise to
the hypothesis that they suffered from a “central syntactic deficit”—in other words, that their knowledge of
grammatical structure was compromised, affecting comprehension as well as production.29 It was subsequently
found, however, that not all of these aphasics suffered
from a comprehension problem;30 further, they performed relatively well in judging the grammaticality of
sentences spoken to them.24 The ability to perform the
grammaticality judgment task clearly depends on the
preservation of syntactic knowledge. There is also evidence that although the spontaneous production of these
patients is restricted to simple sentence forms, there are
ways of getting them to produce more complex utterances. One such task involves alternating sentence repetition with picture description, where the pictures have
no overlap with the content of the sentences provided for
repetition.31 But the production of more complex sentences does not guarantee that these patients will use
correct morphology. For example, after repeating a passive sentence (e.g., “The dog was pulled by the man”)
and being presented with a picture of a boy stung by a
bee, a patient might say “The boy is stinging for the
bee.” Although this sentence has the same underlying
syntactic structure as the passive, the verb is improperly
inflected and the preposition is incorrect.32
The fact that these patients can judge grammaticality with reasonable accuracy suggests that the critical
area for syntactic processing lies outside the region of
damage that results in Broca’s aphasia. Indeed, unpublished data from our laboratory indicate that Wernicke
patients perform worse than Broca’s aphasics on grammaticality judgment tasks. But if syntax is preserved,
why do these patients have a comprehension problem, at
least with respect to semantically reversible sentences?
This difficulty is often exacerbated when the noun that
precedes the verb is not the doer of the action (as, for example, in the case of the passive), probably because the
agent generally precedes the verb in English sentences.33
One possibility is that the patients extract syntactic
structure from the sentences they hear but fail to integrate structural with other critical information—such as
the identities of the lexical items, their meanings, and
knowledge of the thematic structure of the verb (i.e., the
roles of the nouns in relation to the verb—e.g., agent or
recipient of the action—and their relation to syntactic
structure). Instead, they rely on semantic constraints and
on strategies such as taking the preverbal noun to be the
agent.
A possible account of the effect of damage to
frontal language cortex is that this area serves to activate
posterior language processing regions—as is the case for
other capacities, such as spatial processing carried out
by parietal cortex under the influence of frontal activation.34 Functional imaging studies in normal subjects
have shown that activation of the frontal language area
increases with syntactic complexity;35,36 in addition,
electrophysiological investigations reveal that left frontal activation is increased by the presence of grammatical violations.37 One might expect demands on the syntactic processor to increase as a function of increasing
complexity or grammatical error. It is interesting to note
that Broca patients also have difficulty with certain
grammaticality judgment tasks, in particular, those that
entail linking particular word identities (and/or their
meanings) to particular locations in the sentence. Thus,
they proved to be insensitive to violations involving reflexive sentences (e.g., “The woman looked at himself in
the mirror”) in which the gender of the pronoun conflicts
with that of the noun to which it refers.24,38 This further
suggests that the capacity to integrate the various types
of information required for the understanding of sentences is limited in these patients. It may also be the case
that frontal activation is critical for sentence production.
One view of the sentence production deficit in Broca’s
aphasics is that it reflects a timing problem in which lexical items are retrieved too slowly to integrate with sentence structure.39 Decreased activation of lexical capacities because of the loss of frontal activation might
account for such a deficit. It should be noted, however,
that many other explanations of these impairments have
been proposed, ranging from short-term memory limitations40 to specific syntactic functions attributed to
frontal language cortex.41 The literature on this question
is still mired in controversy.42
OPEN QUESTIONS: THE NATURE OF
SEMANTIC ORGANIZATION
Although most of the classical syndromes reflect
damage to the perisylvian language area, the lesions that
produce the transcortical aphasias—sensory and motor—generally lie outside this region. Other disorders
also result from damage to cortical structures outside the
perisylvian cortex. Lesions that affect anterior inferior
regions of the left temporal lobe (or both temporal lobes)
are known to result in semantic deficits in which patients
have difficulty finding words and understanding them
and often exhibit impairments with pictured materials as
well.43 In most cases, these deficits do not result from
the vascular etiologies that are the primary causes of the
major aphasia syndromes. One source of semantic impairment is herpes simplex encephalitis, which has a
predilection for anterior (and medial) temporal tissue;
another is a form of progressive aphasia, which leaves
the patient fluent but with severe word-finding and comprehension difficulty because of deterioration of structures in the left anterior temporal lobe (sometimes both
APHASIA AND THE RELATIONSHIP OF LANGUAGE AND BRAIN—SAFFRAN
lobes). The latter disorder has been termed semantic dementia,44,45 reflecting the fact that other cognitive functions (e.g., recent memory, spatial processes) remain intact for some period of time. This area is also affected in
Alzheimer’s disease, where the semantic deficit is accompanied by other cognitive disturbances.46
Although the semantic disturbance is in some cases
quite general, in other instances there is some selectivity
in the impairment. For example, some cases (of both
herpes encephalitis and semantic dementia) have been
described in which abstract words are better preserved
than concrete words.47–49 This is a surprising finding, as
normal subjects typically show a superiority for concrete
words,50 which is often exacerbated by left hemisphere
brain damage.51 One possible account of this pattern is
that the anterior temporal lobe is the locus of convergence of information from sensory association areas on
its way to hippocampus and surrounding structures; consider, for example, that this area lies anterior to brain regions responsible for visual object recognition.52 Indeed,
knowledge of the perceptual properties of objects—critical aspects of meaning for concrete words—appears deficient in patients who demonstrate the abstract superiority effect.47 Other patients, including most of those who
show abstract superiority,49 demonstrate better preserved knowledge of man-made objects (e.g., tools and
utensils) than living things (animals, fruits and vegetables as well as other types of food). Again, perceptual
properties may be particularly important in distinguishing among objects in these categories. Consider, for example, the primary distinctions among animals such as
tigers, lions and leopards, who share many features of
behavior and habitat; we differentiate them primarily on
the basis of the way they look. Although perceptual
properties may be critical here, loss of this information
is not the only source of specificity in semantic impairment. Other patients with left frontoparietal lesions are
more impaired on artifacts than living things; they are
also typically impaired on body parts.53 Evidence suggests that this deficit reflects loss of information as to the
manner in which objects are manipulated, which may reflect impairment of motor and sensory areas in the left
hemisphere.54
These findings suggest that semantic information is
broadly distributed in the brain and that meaning resides
in other regions in addition to the classical language
areas. This is to be expected, given continuity with nonverbal creatures, who also acquire knowledge of the
world they inhabit. It is likely, of course, that the representations of an object across different regions (visual,
auditory, tactile, and linguistic) are interconnected and
that all parts of the representation are normally activated
when a word or object is encountered.55 In fact, functional imaging studies have shown that words and objects activate the same brain regions in semantic tasks56
but that there is also some selectivity, depending on the
type of object that is tested. For example, left frontal cortex is activated when the materials are tools.57–59 It also
seems reasonable that in cases of brain damage, subregions may be disconnected from one another. Although
it is reasonable to propose that semantic information is
distributed in the brain, there is still considerable debate
as to whether there is a unitary semantic system in humans60 or one that includes subregions specialized for
particular types of information.49,61
CONCLUDING REMARKS
Studied for well over a century, the clinical manifestations of aphasia have been well described, and reasonably good data that localize the major syndrome
types are available. Nevertheless, our knowledge of the
relationships between language and brain remains incomplete. We lack an adequate understanding of how
sentences are comprehended as well as how they are
generated. The path from some sort of fairly abstract
phonological representation of an utterance to its motor
implementation is also poorly understood. And, there is
still considerable debate about how the meanings of
words and objects are represented. Although the study of
brain-damaged individuals has contributed a great deal
to our understanding of these and other questions, much
remains to be done. Future research will no doubt entail
collaboration between psycholinguists and clinical scientists on the various forms of language breakdown, and
the implementation of functional imaging techniques in
studies of brain-damaged as well as normal individuals.
REFERENCES
1. Broca P. Sur le siège de la faculté de langage articulé. Bull Anthropol 1865;5:377–393
2. Alexander MP. Aphasia: clinical and anatomic aspects. In: Feinberg TE, Farah MJ, eds. Behavioral Neurology and Neuropsychology. New York: McGraw-Hill; 1997:133–150
3. Wernicke C. The aphasic symptom complex: a psychological
study on a neurological basis. Breslau: Cohn and Weigert; 1874.
Reprinted in Eggert GH, ed. and trans. Wernicke’s Works on
Aphasia. The Hague: Mouton; 1977
4. Lichtheim C. On aphasia. Brain 1885;7:433–484
5. Caplan D. Neurolinguistics and Linguistic Aphasiology. Cambridge: Cambridge University Press; 1987
6. Dejerine JJ. Contribution a l’étude anatomo-pathologique et clinique des différentes variétés de cecute verbale. Mem Soc Biol
1892;4:61–90
7. Marie P. Révision de la question de l’aphasie: la troisieme convolution frontale gauche nejoue aucun rôle spéciale dans la fonction du langage. Sem Med (Paris) 1906;21;241–247
8. Head H. Aphasia and Kindred Disorders of Speech. Cambridge:
Cambridge University Press; 1926
9. Geschwind N. Disconnection syndromes in animals and man: part
I. Brain 1965;88:237–229
10. Geschwind N. Disconnection syndromes in animals and man: part
II. Brain 1965;88:585–644
11. Schuell H, Jenkins JJ, Carroll JM. Factor analysis of the Minnesota Test for differential diagnosis of aphasia. J Speech Hear
Res 1962;5:439–369
12. Mesulam MM. Large scale neurocognitive networks and distributed processing for attention, language and memory. Ann Neurol 1990;28;597–613
13. Bates E, Harris C, Marchman V, Wulfeck B, Kritchevsky M. Production of complex syntax in normal aging and Alzheimer’s disease. Lang Cognitive Processes 1995;10:487–544
14. Bird H, Franklin S. Cinderella revisited: a comparison of fluent
and non-fluent aphasic speech. J Neurolinguist 1996;9:187–206
15. Kertesz A. Aphasia and Associated Disorders. New York: Grune &
Stratton; 1979
16. Mohr JP, Pessin MS, Finkelstein S, et al. Broca’s aphasia: pathological and clinical. Neurology 1978;28:311–324
417
SEMINARS IN NEUROLOGY VOLUME 20, NUMBER 4 2000
17. Dronkers NF. A new brain region for coordinating speech articulation. Nature 1996;384:159–161
18. Goodglass H, Kaplan E. The Assessment of Aphasia and Related
Disorders. 2nd ed. Philadelphia: Lea & Febiger; 1983
19. Benson DF, Ardila A. Aphasia: A Clinical Perspective. New York:
Oxford University Press; 1996
20. Kertesz A. The Western Aphasia Battery. New York: Grune &
Stratton; 1982
21. Caramazza A. The logic of neuropsychological research and the
problem of patient classifi-cation in aphasia. Brain Lang 1984;
21:9–20
22. Saffran EM, Marin OSM, Yeni-Komshian G. An analysis of speech
perception in word deafness. Brain Lang 1976;3:209–228
23. Caramazza A, Zurif EB. Dissociation of algorithmic and heuristic
processes in language comprehension: evidence from aphasia.
Brain Lang 1976;3:572–582
24. Linebarger M., Schwartz MF, Saffran EM. Sensitivity to grammatical structure in so-called agrammatic aphasics. Cognition
1983;13:641–662
25. Shallice T, Vallar G. The impairment of auditory-verbal short-term
storage. In: Vallar G, Shallice T, eds. Neuropsychological Impairments of Short-Term Memory. Cambridge: Cambridge University Press; 1990:11–53
26. Saffran EM, Martin N. Short-term memory impairment and sentence processing: a case study. In Vallar G, Shallice T, eds. Neuropsychological Impairments of Short-Term Memory. Cambridge: Cambridge University Press; 1990:428–447
27. Levelt WJM, Roelofs A, Meyer AS. A theory of lexical access in
speech production. Behav Brain Sci 1999;22:1–38
28. Grodzinsky Y. The syntactic characterization of agrammatism.
Cognition 1984;16:99–120
29. Berndt RS, Caramazza A. A redefinition of the syndrome of
Broca’s aphasia: implications for a neuropsychological model
of language. Appl Linguist 1980;1:225–278
30. Miceli G, Mazzucchi A, Menn L, Goodglass H. Contrasting cases
of Italian agrammatic aphasics without comprehension disorder.
Brain Lang 1983;19:56–97
31. Hartsuiker RJ, Kolk HHJ. Syntactic facilitation in agrammatic
sentence production. Brain Lang 1998;62:221–254
32. Saffran EM, Martin N. Effects of structural priming on sentence
production in aphasics. Lang Cognitive Processes 1997;12:
877–882
33. Caplan D, Futter C. Assignment of thematic roles to nouns in sentence comprehension by an agrammatic patient. Brain Lang
1986;27:117–134
34. Smith EE, Jonides J, Koeppe RA, et al. Spatial vs. object working
memory: PET investigations. J Cognit Neurosci 1995;7:
337–356
35. Just MA, Carpenter PA, Keller TA, Eddy WF, Thulborn KR. Brain
activation modulated by sentence comprehension. Science
1996;274:114–116
36. Stromswold K, Caplan D, Alpert N, Rauch S. Localization of syntactic comprehension by positron emission tomography. Brain
Lang 1996;52:452–473
37. Hahne A, Friederici AD. Electrophysiological evidence for two
steps in syntactic analysis: early automatic and late controlled
processes. J Cognit Neurosci 1999;11:194–205
38. Linebarger M. Agrammatism as evidence about grammar. Brain
Lang 1995;50:52–91
39. Kolk HHJ. A time-based approach to agrammatic production.
Brain Lang 1995;50:282–303
40. Miyake A, Carpenter PA, Just MA. A capacity approach to syntactic comprehension disorders: making normal adults per-
418
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
form like aphasic patients. Cognit Neuropsychol 1994;11:671–
717
Grodzinsky Y. Theoretical Perspectives on Language Deficits.
Cambridge, MA: MIT Press; 1990
Berndt RS. Sentence processing in aphasia. In: Sarno M. Acquired
Aphasia. 3rd ed. New York: Academic Press; 1998:229–267
Howard D, Patterson K. Pyramids and palm trees: a test of semantic access from pictures and words. Bury St. Edmonds, UK:
Thames Valley Test Company; 1992
Snowden JS, Goulding PJ, Neary D. Semantic dementia: a form of
circumscribed cerebral atrophy. Behav Neurol 1989;2:167–182
Hodges JR, Patterson K, Oxbury S, Funnell E. Semantic dementia:
progressive fluent aphasia with temporal lobe atrophy. Brain
1992;115:1783–1806
Chertkow H, Bub D, Caplan D. Constraining theories of semantic
memory processing: evidence from dementia. Cognit Neuropsychol 1992;9:327–365
Breedin SD, Saffran EM, Coslett HB. Reversal of the concreteness
effect in a patient with semantic dementia. Cognit Neuropsychol 1994;11:617–660
Sirigu A, Duhamel JR, Poncet M. The role of sensorimotor experience in object recognition: a case of multimodal agnosia. Brain
1991;114:2555–2573
Saffran EM, Sholl A. Clues to the functional and neural architecture
of word meaning. In: Hagoort P, Brown C, eds. Neurocognition
of Language. Oxford: Oxford University Press; 1999: 241–272
Paivio A. Dual coding theory: retrospect and current status. Can J
Psycho 1991;45:255–287
Goodglass H, Hyde MR, Blumstein SE. Frequency, picturability
and availability of nouns in aphasia. Cortex 1969;5:104–119
Ungerleider LG, Mishkin M. Two visual systems. In: Ingle DJ,
Goodale MA, Mansfield RJW, eds. Analysis of Visual Behavior.
Cambridge, MA: MIT Press; 1982:549–586
Saffran EM, Schwartz MF. Of cabbages and things: semantic
memory from a neuropsychological perspective: a tutorial review. In: Umilta C, Moscovitch M, eds. Attention and Performance XV: Conscious and Nonconscious Information Processing. Cambridge, MA: MIT Press, 1994:507–536
Buxbaum L, Saffran EM. Knowing “how” vs. “what for”: a new
dissociation. Brain Lang 1998;65:73–76
Allport DA. Distributed memory, modular subsystems and dysphasia. In: Newman SK, Epstein R, eds. Current Perspectives in
Dysphasia. Edinburgh: Churchill Livingstone; 1985:32–60
Vandenberghe R, Price C, Wise R, Josephs O, Frackowiak RSJ.
Functional anatomy of a common semantic system for words
and pictures. Nature 1996;383:254–256
Grafton ST, Fadiga L, Arbib M, Rizzolatti G. Premotor cortex activation during observation and naming of familiar tools. Neuroimage 1997;6:231–236
Perani D, Cappa SF, Bettinardi V, et al. Different neural systems
for the recognition of animals and man-made tools. NeuroReport 1995;6:1637–1641
Martin A, Wiggs CL, Ungerleider LG, Haxby JV. Neural correlates of category-specific knowledge. Nature 1996;379:649–
652
Hillis AE, Rapp B, Caramazza A. Constraining claims about theories of semantic memory: more on unitary vs. multiple semantics. Cognit Neuropsychol 1995;12:175–186
Shallice T. From Neuropsychology to Mental Structure. Cambridge: Cambridge University Press; 1988
Schwartz MF, Saffran EM, Bloch DE, Dell GS. Disordered speech
production in aphasic and normal speakers. Brain Lang 1994;
47:52–88