Definitions

Aphasia is defined as a disorder of language acquired secondary to brain damage. This definition, adapted from Alexander and Benson (1997), separates aphasia from several related disorders. First, aphasia is distinguished from congenital or developmental language disorders, called dysphasias. (In contrast with British usage, in the United States the term dysphasia applies to developmental language disorders rather than partial or incomplete aphasia.)

Second, aphasia is a disorder of language rather than speech. Speech is the articulation and phonation of language sounds; language is a complex system of communication symbols and rules for their use. Aphasia is distinguished from motor speech disorders (the subject of Part B of this chapter), which include dysarthria, dysphonia (voice disorders), stuttering, and speech apraxia. Dysarthrias are disorders of articulation of single sounds; causes of these disorders may include mechanical disturbance of the tongue or larynx and neurological disorders such as dysfunction of the muscles, neuromuscular junction, cranial nerves, bulbar anterior horn cells, corticobulbar tracts, cerebellar connections, or basal ganglia. Apraxia of speech is a syndrome of misarticulation of phonemes, especially consonant sounds. Unlike dysarthria, in which certain phonemes are consistently distorted, apraxia of speech is characterized by inconsistent distortions and substitutions of phonemes. The disorder is called an apraxia because there is no primary motor deficit in articulation of individual phonemes. Clinically, speech-apraxic patients produce inconsistent articulatory errors, usually worse on the initial phonemes of a word and with polysyllabic utterances. Apraxia of speech, so defined, is commonly involved in speech production difficulty in the aphasias.

Third, aphasia is distinguished from disorders of thought. Thought involves the mental processing of images, memories, and perceptions, usually but not necessarily involving language symbols. Psychiatric disorders derange thought and alter the content of speech without affecting its linguistic structure. Schizophrenic patients, for example, may manifest bizarre and individualistic word choices, with loose associations and a loss of organization in discourse together with vague or unclear references and communication failures (Docherty et al., 1996). Elementary language and articulation, however, are intact. Abnormal language content in psychiatric disorders is therefore not considered to represent aphasia, because the disorder is more one of thought than of language. Language disorders associated with diffuse brain diseases such as encephalopathies and dementias do qualify as aphasias, but the involvement of other cognitive functions distinguishes them from aphasia secondary to focal brain lesions.

An understanding of language disorders requires an elementary review of linguistic components. Phonemes are the smallest meaning-carrying sounds; morphology is the use of appropriate word endings and connector words for tenses, possessives, and singular versus plural; semantics refers to word meanings; the lexicon is the internal dictionary; and syntax is the grammatical construction of phrases and sentences. Discourse refers to the use of these elements to create organized and logical expression of thoughts. Pragmatics refers to the proper use of speech and language in a conversational setting, including pausing while others are speaking, taking turns properly, and responding to questions. Specific language disorders affect one or more of these elements.

Relevant Neuroanatomy

Language processes have a clear neuroanatomical basis. In simplest terms, the reception and processing of spoken language take place in the auditory system, beginning with the cochlea and proceeding through a series of way stations to the auditory cortex, the Heschl gyrus, in each superior temporal gyrus. Decoding sounds into linguistic information involves the posterior part of the left superior temporal gyrus, the Wernicke area or Brodmann area 22, which gives access to a network of cortical associations to assign word meanings. For both repetition and spontaneous speech, auditory information is transmitted to the Broca area in the posterior inferior frontal gyrus. This area of cortex “programs” the neurons in the adjacent motor cortex subserving the mouth and larynx, from which descending axons travel to the brainstem cranial nerve nuclei. The inferior parietal lobule, especially the supramarginal gyrus, also may be involved in phoneme processing in language comprehension and in phoneme production for repetition and speech (Hickok and Poeppel, 2000). These anatomical relationships are shown in Figs. 12A.1 and 12A.2. Reading requires perception of visual language stimuli by the occipital cortex, followed by processing into auditory language information via the heteromodal association cortex of the angular gyrus. Writing involves activation of motor neurons projecting to the arm and hand. A French study that used aphasia testing and magnetic resonance imaging (MRI) scans to evaluate 107 stroke patients confirmed the general themes of nearly 150 years of clinical aphasia research: that frontal lesions caused nonfluent aphasia, whereas posterior temporal lesions affected comprehension (Kreisler et al., 2000).

Fig. 12A.1 Lateral surface of the left hemisphere, showing a simplified gyral anatomy and the relationships between the Wernicke area and Broca area. Not shown is the arcuate fasciculus, which connects the two cortical speech centers via the deep subcortical white matter.

Fig. 12A.2 Coronal plane diagram of the brain, indicating the inflow of auditory information from the ears to the primary auditory cortex in both superior temporal regions (xxx), and then to the Wernicke area (ooo) in the left superior temporal gyrus. The motor outflow of speech descends from the Broca area (B) to the cranial nerve nuclei of the brainstem via the corticobulbar tract (dashed arrow). In actuality, the Broca area is anterior to the Wernicke area, and the two areas would not appear in the same coronal section.

These pathways, and doubtless others, constitute the cortical circuitry for language comprehension and expression. In addition, other cortical centers involved in cognitive processes project into the primary language cortex, influencing the content of language. Finally, subcortical structures play increasingly recognized roles in language functions. The thalamus, a relay for the reticular activating system, appears to alert the language cortex, and lesions of the dominant thalamus frequently produce fluent aphasia. Nuclei of the basal ganglia involved in motor functions, especially the caudate nucleus and putamen, participate in expressive speech. No wonder, then, that language disorders are seen with a wide variety of brain lesions and are important in practical neurological diagnosis and localization.

In right-handed people, and in a majority of left-handers as well, clinical syndromes of aphasia result from left hemisphere lesions. Rarely, aphasia may result from a right hemisphere lesion in a right-handed patient, a phenomenon called crossed aphasia (Bakar et al., 1996). In left-handed persons, language disorders are usually similar to those of right-handed persons with similar lesions, but occasional cases manifest with atypical syndromes that suggest a right hemisphere capability for at least some language functions. For example, a patient with a large left frontotemporoparietal lesion may have preserved comprehension, suggesting right hemisphere language comprehension. For the same reason, recovery from aphasia may be better in some left-handed than in right-handed patients with left hemisphere strokes.

Diagnostic Features

Muteness, a total loss of speech, may represent severe aphasia (see Aphemia later in the chapter). Muteness also can be a sign of dysarthria, frontal lobe dysfunction with akinetic mutism, severe extrapyramidal system dysfunction (as in Parkinson disease), non-neurological disorders of the larynx and pharynx, or even psychiatric syndromes such as catatonia. Caution must therefore be taken in diagnosing the mute patient as aphasic. A good rule of thumb is that if the patient can write or type and the language form and content appear normal, the disorder is probably not aphasic in origin. If the patient cannot speak or write but makes apparent effort to vocalize, and if there is also evidence of deficient comprehension, aphasic muteness is likely. Associated signs of a left hemisphere injury, such as right hemiparesis, also aid in diagnosis. Finally, if the patient gradually begins to make sounds containing paraphasic errors, aphasia can be identified with confidence.

Hesitant speech is a feature of aphasia but also of motor speech disorders such as dysarthria or stuttering, and it may be a manifestation of a psychogenic disorder. A second rule of thumb is that if the utterances of a hesitant speaker can be transcribed into normal language, the patient is not aphasic. Hesitancy occurs in many aphasia syndromes for various reasons, including difficulty in speech initiation, imprecise articulation of phonemes, deficient syntax, and word-finding difficulty.

Anomia, or inability to produce a specific name, is generally a reliable indicator of language disorder, although it also may reflect memory loss. Anomia is manifested in aphasic speech by word-finding pauses and circumlocutions, or use of a phrase when a single word would suffice (e.g., “the thing you tell time with” for watch).

Paraphasic speech refers to the presence of errors in the patient’s speech output. Paraphasic errors are divided into literal or phonemic errors involving substitution of an incorrect sound (e.g., “shoon” for “spoon”) and verbal or semantic errors involving substitution of an incorrect word (e.g., “fork” for “spoon”). A related language symptom is perseveration, the inappropriate repetition of a previous response. Occasionally, aphasic utterances involve nonexistent word forms called neologisms. A pattern of paraphasic errors and neologisms that so contaminate speech that the meaning cannot be discerned is called jargon speech.

Another cardinal symptom of aphasia is failure to comprehend the speech of others. Most aphasic patients also have difficulty with comprehension and production of written language (reading and writing). Fluent paraphasic speech usually makes an aphasic disorder obvious. The chief considerations in the differential diagnosis here include aphasia, psychosis, acute encephalopathy or delirium, and dementia. Aphasic patients usually do not appear confused and do not exhibit inappropriate behavior; they are not agitated and do not misuse objects, with occasional exceptions in acute syndromes of Wernicke or global aphasia. In contrast, most psychotic patients speak in an easily understood, grammatically appropriate manner, but their behavior and speech content are abnormal. Only rarely do schizophrenics speak in “clang association” or “word salad” speech. Sudden onset of fluent paraphasic speech in a middle-aged or elderly patient should always be suspected of representing a left hemisphere lesion with aphasia.

Patients with acute encephalopathy or delirium may manifest paraphasic speech and “higher” language deficits, such as the inability to write, but the grammatical expression of language is less disturbed than its content. These language symptoms, moreover, are less prominent than accompanying behavioral disturbances such as agitation, hallucinations, drowsiness, or excitement, and cognitive difficulties such as disorientation, memory loss, or delusional thinking.

Chronic encephalopathies, or dementias, pose a more difficult diagnostic problem because involvement of the language cortex produces readily detectable language deficits, especially involving naming, reading, and writing. These language disorders (see Language in Dementing Diseases later in this chapter) differ from aphasia secondary to focal lesions mainly by the involvement of other cognitive functions such as memory and visuospatial processes.

Broca Aphasia

In 1861, the French physician Paul Broca described two patients, establishing the aphasia syndrome that now bears his name. The speech pattern is nonfluent; on bedside examination, the patient speaks hesitantly, often producing the principal meaning-containing nouns and verbs but omitting small grammatical words and morphemes. This pattern is called agrammatism or “telegraphic speech.” An example is “wife come hospital.” Patients with acute Broca aphasia may be mute or may produce only single words, often with dysarthria and apraxia of speech. They make many phonemic errors, inconsistent from utterance to utterance, with substitution of phonemes usually differing only slightly from the correct target (e.g., /p/ for /b/). Naming is deficient, but the patient often manifests a “tip-of-the-tongue” phenomenon, getting out the first letter or phoneme of the correct name. Paraphasic errors in naming more frequently are of literal than of verbal type. Auditory comprehension seems intact, but detailed testing usually reveals some deficiency, particularly in the comprehension of complex syntax. For example, for persons with Broca aphasia, sentences with embedded clauses involving prepositional relationships cause difficulty in comprehension as well as in expression (“The rug that Bill gave to Betty tripped the visitor”). A positron emission tomography (PET) study in normal persons (Caplan et al., 1998) showed activation of the Broca area in the frontal cortex during tests of syntactic comprehension; the Broca area thus appears to be involved in syntactical operations, both expressively and receptively. Repetition is hesitant in these patients, resembling their spontaneous speech. Reading often is impaired despite relatively preserved auditory comprehension. Benson termed this reading difficulty of Broca aphasics the “third alexia,” in contradistinction to the two classical types of alexia (see Aphasic Alexia later in the chapter). Patients with Broca aphasia may have difficulty with syntax in reading, just as in auditory comprehension and speech. Writing is virtually always deficient in Broca aphasics. Most patients have a right hemiparesis necessitating use of the nondominant left hand for writing, but this left-handed writing is far more abnormal than the awkward renditions of a normal right-handed person attempting to write left-handed. Many patients can scrawl only a few letters.

Associated neurological deficits of Broca aphasia include right hemiparesis, hemisensory loss, and apraxia of the oral apparatus and the nonparalyzed left limbs. Apraxia in response to motor commands is important to recognize because it may be mistaken for comprehension disturbance. As mentioned earlier, comprehension should also be tested by responses to yes/no questions or commands to point to an object. The common features of Broca aphasia are listed in Table 12A.1.

Table 12A.1 Bedside Features of Broca Aphasia

An important clinical feature of Broca aphasia is its frequent association with depression (Robinson, 1997). Patients with Broca aphasia typically are aware of and frustrated by their deficits. At times they become withdrawn and refuse help or therapy. Usually the depression lifts with recovery from the deficit, but it may be a limiting factor in rehabilitation.

The lesions responsible for Broca aphasia usually include the traditional Broca area in the posterior part of the inferior frontal gyrus, along with damage to adjacent cortex and subcortical white matter. Most patients with lasting Broca aphasia, including Broca’s original cases, have much larger left frontoparietal lesions, including most of the territory of the upper division of the left middle cerebral artery. In such patients, the deficit typically evolves from global to Broca aphasia over weeks to months. Patients who manifest Broca aphasia immediately after their strokes, by contrast, have smaller lesions of the inferior frontal region, and their deficits generally resolve quickly. In computed tomography (CT) scan analyses at the Boston Veterans Administration Medical Center, lesions restricted to the lower precentral gyrus produced only dysarthria and mild expressive disturbance. Lesions involving the traditional Broca area (Brodmann areas 44 and 45) resulted in difficulty initiating speech, and lesions combining the Broca area, the lower precentral gyrus, and subcortical white matter yielded the full syndrome of Broca aphasia. In other studies at the center, damage to two subcortical white matter sites—the rostral subcallosal fasciculus deep to the Broca area and the periventricular white matter adjacent to the body of the left lateral ventricle—was required to cause permanent nonfluency. These concepts concerning the Broca area and its mainly temporary role in Broca aphasia have been confirmed by a recent MRI study, indicating that MRI lesions in the Broca area correlate with Broca or global aphasia in acute stroke, but not in the chronic period (Ochfeld et al., 2010). Fig. 12A.3 shows an MRI scan of the brain from a patient with Broca aphasia.

Fig. 12A.3 Magnetic resonance imaging study of the brain of a patient with Broca aphasia. In this patient, the cortical Broca area, subcortical white matter, and the insula all were involved in the infarction. The patient made a good recovery.

Aphemia

A rare variant of Broca aphasia is aphemia, a nonfluent syndrome in which the patient initially is mute and then becomes able to speak with phoneme substitutions and pauses. All other language functions are intact, including writing. This rare and usually transitory syndrome results from small lesions of the Broca area or its subcortical white matter or of the inferior precentral gyrus. Because written expression and auditory comprehension are normal, aphemia is not a true language disorder; aphemia may be equivalent to pure apraxia of speech.

Wernicke Aphasia

Wernicke aphasia may be considered the opposite of Broca aphasia in that expressive speech is fluent, but comprehension is impaired. The speech pattern is effortless and sometimes even excessively fluent (“logorrhea”). A speaker of a foreign language would notice nothing amiss, but a listener who shares the patient’s language detects speech empty of meaning, containing verbal paraphasias, neologisms, and jargon productions. Neurolinguists refer to this pattern as paragrammatism. In milder cases, the intended meaning of an utterance may be discerned, but the sentence goes awry with paraphasic substitutions. Naming in Wernicke aphasia is deficient, often with bizarre, paraphasic substitutions for the correct name. Auditory comprehension is impaired, sometimes even for simple nonsense questions. Deficient semantics is the major cause of the comprehension disturbance in Wernicke aphasia, along with disturbed access to the internal lexicon. Repetition is impaired; whispering a phrase in the patient’s ear, as in a hearing test, may help cue the patient to attempt repetition. Reading comprehension is usually affected in a fashion similar to that observed for auditory comprehension, but occasionally patients show greater deficit in one modality than in the other. The discovery of spared reading ability in Wernicke aphasics is important in allowing these patients to communicate. In addition, neurolinguistic theories of reading must account for the access of visual language images to semantic interpretation, even in the absence of auditory comprehension. Writing also is impaired, but in a manner quite different from that of Broca aphasia. The patient usually has no hemiparesis and can grasp the pen and write easily. Written productions are even more abnormal than oral ones, however, in that spelling errors are also evident. Writing samples are especially useful in the detection of mild Wernicke aphasia.

Associated signs are limited in Wernicke aphasia; most patients have no elementary motor or sensory deficits, although a partial or complete right homonymous hemianopia may be present. The characteristic bedside examination findings in Wernicke aphasia are summarized in Table 12A.2.

Table 12A.2 Bedside Features of Wernicke Aphasia

The psychiatric manifestations of Wernicke aphasia are quite different from those of Broca aphasia. Depression is less common; many Wernicke aphasics seem unaware of or unconcerned about their communicative deficits. With time, some patients become angry or paranoid about the inability of family members and medical staff to understand them. This behavior, like depression, may hinder rehabilitative efforts.

The lesions of patients with Wernicke aphasia usually involve the posterior portion of the superior temporal gyrus, sometimes extending into the inferior parietal lobule. Fig. 12A.4 shows a typical example. The exact confines of the Wernicke area have been much debated. Damage to this area (Brodmann area 22) has been reported to correlate most closely with persistent loss of comprehension of single words, although only larger temporoparietal lesions have been found in patients with lasting Wernicke aphasia. In the acute phase, the ability to match a spoken word to a picture is quantitatively related to decreased perfusion of the Wernicke area on perfusion-weighted MRI, indicating less variability during the acute phase than after recovery has taken place (Hillis et al., 2001). Electrical stimulation of the Wernicke area produces consistent interruption of auditory comprehension, supporting the importance of this region for decoding auditory language. A receptive speech area in the left inferior temporal gyrus has also been suggested by electrical stimulation studies and by a few descriptions of patients with seizures involving this area (Kirshner et al., 1995), but aphasia has not been recognized with destructive lesions of this area. Extension of the lesion of Wernicke aphasia into the inferior parietal region may predict greater involvement of reading comprehension. In terms of vascular anatomy, the Wernicke area lies within the territory of the inferior division of the left middle cerebral artery.

Fig. 12A.4 Brain imaging in Wernicke aphasia. The patient was an elderly woman. A and B, Axial and coronal magnetic resonance images. C, Axial positron emission tomography (PET) scan. There is a large left superior temporal lobe lesion. The onset of the deficit was not clear, and the PET scan was useful in showing that the lesion had reduced metabolism, favoring a stroke over a tumor.

Pure Word Deafness

Pure word deafness is a rare but striking syndrome of isolated loss of auditory comprehension and repetition, without any abnormality of speech, naming, reading, or writing. Hearing for pure tones and nonverbal noises (e.g., animal cries) is intact. Most cases have mild aphasic deficits, especially paraphasic speech. Classically, the anatomical substrate is a bilateral lesion isolating the Wernicke area from input from both the Heschl gyri. Pure word deafness is thus an example of a “disconnection syndrome” in which the deficit results from loss of white matter connections rather than of gray matter language centers. In some cases of pure word deafness, however, the underlying cause is a unilateral left temporal lesion. These cases closely resemble Wernicke aphasia, with greater impairment of auditory comprehension than of reading.

Global Aphasia

Global aphasia may be thought of as a summation of the deficits of Broca aphasia and Wernicke aphasia. Speech is nonfluent or mute, but comprehension also is poor, as are naming, repetition, reading, and writing. Most patients have dense right hemiparesis, hemisensory loss, and often hemianopia, although occasional patients have mild or no hemiparesis. Milder aphasic syndromes in which all modalities of language are affected often are called mixed aphasias. The lesions of patients with global aphasia are usually large, involving both the inferior frontal and superior temporal regions, and often much of the parietal lobe in between. This lesion represents most of the territory of the left middle cerebral artery. Patients in whom the superior temporal gyrus is spared tend to recover their auditory comprehension, with evolution of the deficit toward the syndrome of Broca aphasia. Recovery in global aphasia may be prolonged; in patients with global aphasia, more clinical improvement may occur during the second 6 months than in the first 6 months after a stroke. Characteristics of global aphasia are presented in Table 12A.3.

Table 12A.3 Bedside Features of Global Aphasia

Conduction Aphasia

Conduction aphasia is an uncommon but theoretically important syndrome that can be recognized by its striking deficit of repetition. Most patients have relatively normal spontaneous speech, although some make literal paraphasic errors and hesitate frequently for self-correction. Naming is impaired to varying degrees, but auditory comprehension is preserved. Repetition may be disturbed to seemingly ridiculous extremes such that a patient who is capable of self-expression at a sentence level and can comprehend conversation may be unable to repeat even single words. One such patient could not repeat the word “boy” but said “I like girls better.” Reading and writing are somewhat variable, but reading aloud may share some of the same difficulty as repeating. Associated deficits include hemianopia in some patients; right-sided sensory loss may be present, but right hemiparesis usually is mild or absent. Some patients have limb apraxia, creating a misimpression that comprehension is impaired. Bedside examination findings in conduction aphasia are summarized in Table 12A.4.

Table 12A.4 Bedside Features of Conduction Aphasia

The lesions of conduction aphasia usually involve either the superior temporal or inferior parietal region. Benson and associates suggested that patients with limb apraxia have parietal lesions, whereas those without apraxia have temporal lesions. Conduction aphasia may represent a stage of recovery in patients with Wernicke aphasia in whom the damage to the superior temporal gyrus is not complete.

Conduction aphasia has been advanced as a classical disconnection syndrome. Wernicke originally postulated that a lesion disconnecting the Wernicke and Broca areas would produce this syndrome; Geschwind later pointed to the arcuate fasciculus, a white-matter tract traveling from the deep temporal lobe, around the sylvian fissure, to the frontal lobe, as the site of disconnection. Anatomical involvement of the arcuate fasciculus is present in most if not all cases of conduction aphasia, but doubt has arisen as to the importance of the arcuate fasciculus to this syndrome. Bernal and Ardila (2009) cite evidence that the arcuate fasciculus terminates in the premotor/motor areas, and not in the Broca area. In addition, there is usually also cortical involvement of the supramarginal gyrus or temporal lobe. The supramarginal gyrus appears to be involved in auditory immediate memory and in phoneme perception related to word meaning as well as phoneme generation (Hickok and Poeppel, 2000). Lesions in this area are associated with conduction aphasia and phonemic paraphasic errors. Other investigators have pointed out that lesions of the arcuate fasciculus do not always produce conduction aphasia. Another theory of conduction aphasia suggests a defect in auditory verbal short-term memory.

Anomic Aphasia

Anomic aphasia refers to aphasic syndromes in which naming, or access to the internal lexicon, is the principal deficit. Spontaneous speech is normal except for the pauses and circumlocutions produced by the inability to name. Comprehension, repetition, reading, and writing are intact except for the same word-finding difficulty in written productions. Anomic aphasia is common but less specific in localization than the other aphasic syndromes. Isolated severe anomia may indicate focal left hemisphere pathology. Alexander and Benson (1997) refer to the angular gyrus as the site of lesions producing anomic aphasia, but lesions there usually produce other deficits as well, including alexia and the four elements of the Gerstmann syndrome: agraphia, right-left disorientation, acalculia, and finger agnosia, or inability to identify fingers. Isolated lesions of the temporal lobe can produce pure anomia, and PET studies of naming in normal persons also have shown consistent activation of the superior temporal lobe. Inability to produce nouns is characteristic of temporal lobe lesions, whereas inability to produce verbs occurs more often with frontal lesions. Even specific classes of nouns may be selectively affected in some cases of anomic aphasia. Anomia also is seen with mass lesions elsewhere in the brain, and in neurodegenerative disorders such as Alzheimer disease. Anomic aphasia also is a common stage in the recovery of many aphasic syndromes. Anomic aphasia thus serves as an indicator of left hemisphere or diffuse brain disease, but it has only limited localizing value. The typical features of anomic aphasia are presented in Table 12A.5.

Table 12A.5 Bedside Features of Anomic Aphasia

Transcortical Aphasias

The transcortical aphasias are syndromes in which repetition is normal, presumably because the causative lesions do not disrupt the perisylvian language circuit from the Wernicke area through the arcuate fasciculus to the Broca area. Instead, these lesions disrupt connections from other cortical centers into the language circuit—hence the name transcortical. The transcortical syndromes are easiest to think of as analogs of the syndromes of global, Broca, and Wernicke aphasias, with intact repetition.

Mixed transcortical aphasia, or the “syndrome of the isolation of the speech area,” is a global aphasia in which the patient repeats, often echolalically, but has no propositional speech or comprehension. This syndrome is rare, occurring predominantly with large watershed infarctions of the left hemisphere or both hemispheres that spare the perisylvian cortex, or in advanced dementias.

Transcortical motor aphasia is an analog of Broca aphasia in which speech is hesitant or telegraphic, comprehension is relatively spared, but repetition is fluent. This syndrome occurs with lesions in the frontal lobe anterior to the Broca area, in the deep frontal white matter, or in the medial frontal region in the vicinity of the supplementary motor area. All of these lesion sites are within the territory of the anterior cerebral artery, separating this syndrome from the aphasia syndromes of the middle cerebral artery (Broca, Wernicke, global, and conduction aphasias). This syndrome has also been reported with watershed infarctions, reflecting carotid artery stenosis.

The third transcortical syndrome, transcortical sensory aphasia, is an analog of Wernicke aphasia in which fluent paraphasic speech, paraphasic naming, impaired auditory and reading comprehension, and abnormal writing coexist with normal repetition. This syndrome is relatively uncommon, occurring with strokes of the left temporo-occipital area and in dementias. “Watershed” infarctions between the left middle and posterior cerebral artery territories may produce this syndrome. Bedside examination findings in the transcortical aphasias are summarized in Table 12A.6.

Subcortical Aphasias

A current area of interest in aphasia research involves the “subcortical” aphasias. Although all the syndromes discussed so far are defined by behavioral characteristics that can be diagnosed at bedside examination, the subcortical aphasias are defined by lesion localization in the basal ganglia or deep cerebral white matter; in other words, diagnosis of this aphasia syndrome is based on brain imaging. As knowledge about subcortical aphasia has accumulated, two major groups of aphasic symptomatology have been described: aphasia with thalamic lesions and aphasia with lesions of the subcortical white matter and basal ganglia.

Left thalamic hemorrhages frequently produce a Wernicke-like fluent aphasia with better comprehension than in cortical Wernicke aphasia. A fluctuating or “dichotomous” state has been described, alternating between an alert state with nearly normal language and a drowsy state in which the patient mumbles paraphasically and comprehends poorly. Luria has referred to this state as a “quasi-aphasic abnormality of vigilance.” One way of thinking of thalamic aphasia is that the thalamus plays a role in alerting the language cortex such that the language cortex, in effect, goes to sleep. Thalamic aphasia can occur even with a right thalamic lesion in a left-handed patient, indicating that hemispheric language dominance extends to the thalamic level. Although some skeptics have attributed thalamic aphasia to pressure on adjacent structures and secondary effects on the cortex, cases of thalamic aphasia have been described with small ischemic lesions, especially those involving the paramedian or anterior nuclei of the thalamus in the territory of the tuberothalamic artery. Because these lesions produce little or no mass effect, such cases indicate that the thalamus and its connections play a definite role in language function (Carrerra and Bogousslavsky, 2006).

Lesions of the left basal ganglia and deep white matter also cause aphasia. As in thalamic aphasia, the first syndromes described were in basal ganglia hemorrhages, especially those involving the putamen, the most common site of hypertensive intracerebral hemorrhage. Here the aphasic syndromes are more variable, but most commonly involve global or Broca-like aphasia. As in thalamic lesions, ischemic strokes have provided better localizing information than hemorrhage cases. The most common lesion is an infarct involving the anterior putamen, caudate nucleus, and anterior limb of the internal capsule. Patients with this lesion have an “anterior subcortical aphasia syndrome” involving dysarthria, decreased fluency, mildly impaired repetition, and mild comprehension disturbance. This syndrome most closely resembles Broca aphasia, but with greater dysarthria and less language dysfunction. Fig. 12A.5 shows imaging findings in an example of this syndrome. More restricted lesions of the anterior putamen, head of caudate, and periventricular white matter produce hesitancy or slow initiation of speech but little true language disturbance. More posterior lesions involving the putamen and deep temporal white matter, referred to as the temporal isthmus, are associated with fluent paraphasic speech and impaired comprehension, resembling the features of Wernicke aphasia. Small lesions in the posterior limb of the internal capsule and adjacent putamen cause mainly dysarthria, but mild aphasic deficits may occasionally occur. Finally, larger subcortical lesions involving both the anterior and posterior lesion sites produce global aphasia. A wide variety of aphasia syndromes can thus be seen with subcortical lesion sites. Nadeau and Crosson (1997) presented an anatomical model of basal ganglia involvement in speech and language, based on the known motor functions and fiber connections of these structures. Evidence from PET indicates that basal ganglia lesions affect language, both directly and indirectly, via decreased activation of cortical language areas.

Fig. 12A.5 Magnetic resonance imaging (MRI) study of the brain with slices in the axial, coronal, and sagittal planes in subcortical aphasia. The lesion is an infarction involving the anterior caudate, putamen, and anterior limb of the left internal capsule. The patient presented with dysarthria and mild, nonfluent aphasia with anomia, with good comprehension. The advantage of MRI in permitting visualization of the lesion in all three planes is apparent.

The insula, a cortical structure that shares a deep location with the subcortical structures, may also be important to speech and language function. Dronkers (1996) reported that involvement of this area is closely associated with the presence of apraxia of speech in aphasic patients. Hillis and colleagues (2004), however, in MRI studies of brain in acute stroke patients, found that the left frontal cortex correlates more with speech apraxia than does the insula.

In clinical terms, subcortical lesions do produce aphasia, although less commonly than cortical lesions do, and the language characteristics of subcortical aphasias often are atypical. The presentation of a difficult-to-classify aphasic syndrome, in the presence of dysarthria and right hemiparesis, should lead to suspicion of a subcortical lesion.

Pure Alexia without Agraphia

Alexia, or acquired inability to read, is a form of aphasia, according to the definition given at the beginning of this chapter. The classic syndrome of alexia, pure alexia without agraphia, was described by the French neurologist Dejerine in 1892. This syndrome may be thought of as a linguistic blindfolding (or “pure word blindness”) in that patients can write but cannot read their own writing. On bedside examination, speech, auditory comprehension, and repetition are normal. Naming may be deficient, especially for colors. Patients initially cannot read at all; as they recover, they learn to read letter by letter, spelling out words laboriously. They cannot read words at a glance as normal readers do. By contrast, they quickly understand words spelled orally to them, and they can spell normally, both orally and in writing. Some patients can match words to pictures, indicating that some subconscious awareness of the word is present, perhaps in the right hemisphere. Associated deficits include a right hemianopia or right upper quadrant defect in nearly all patients and frequently a deficit of short-term memory. There usually is no hemiparesis or sensory loss.

The causative lesion in pure alexia is nearly always a stroke in the territory of the left posterior cerebral artery, with infarction of the medial occipital lobe, often the splenium of the corpus callosum, and often the medial temporal lobe. Dejerine postulated a disconnection between the intact right visual cortex and left hemisphere language centers, particularly the angular gyrus. Fig. 12A.6 is an adaptation of Dejerine’s original diagram; Fig. 12A.7 shows an MRI of a patient with alexia without agraphia. Geschwind later rediscovered this “disconnection” hypothesis. Although Damasio and Damasio found splenial involvement in only 2 of 16 cases, they postulated a disconnection within the deep white matter of the left occipital lobe. As in the disconnection hypothesis for conduction aphasia, the theory fails to explain all the behavioral phenomena, such as the sparing of single-letter reading. A deficit in immediate memory span for visual language elements or an inability to perceive multiple letters at once (simultanagnosia) also can explain many features of the syndrome. Typical features of pure alexia without agraphia are presented in Table 12A.7.

Fig. 12A.6 Horizontal brain diagram of pure alexia without agraphia, adapted from that of Dejerine, 1892. Visual information from the left visual field reaches the right occipital cortex but is “disconnected” from the left hemisphere language centers by the lesion in the splenium of the corpus callosum.

Fig. 12A.7 Magnetic resonance image of the brain obtained using a fluid-attenuated inversion recovery (FLAIR) sequence. The patient was an 82-year-old man with alexia without agraphia. The infarction involves the medial occipital lobe and the splenium of the corpus callosum, within the territory of the left posterior cerebral artery.

Table 12A.7 Bedside Features of Pure Alexia without Agraphia

Alexia with Agraphia

The second classic alexia syndrome, alexia with agraphia, described by Dejerine in 1891, may be thought of as an acquired illiteracy in which a previously literate patient is rendered unable to read or write. The oral language modalities of speech, naming, auditory comprehension, and repetition are largely intact, but in many cases the patient demonstrates a fluent paraphasic speech pattern with impaired naming. This syndrome thus overlaps Wernicke aphasia, especially in cases in which reading is more impaired than auditory comprehension. Associated deficits include right hemianopia and elements of the Gerstmann syndrome: agraphia, acalculia, right-left disorientation, and finger agnosia. The lesions typically involve the inferior parietal lobule, especially the angular gyrus. Etiologic disorders include strokes in the territory of the angular branch of the left middle cerebral artery and mass lesions in the same region. Characteristic features of the syndrome of alexia with agraphia are summarized in Table 12A.8.

Table 12A.8 Bedside Features of Alexia with Agraphia

Aphasic Alexia

In addition to the two classic alexia syndromes, many patients with aphasia have associated reading disturbance. Examples already cited are the “third alexia” of Broca aphasia and the reading deficit of Wernicke aphasia. Neurolinguists and cognitive psychologists have divided alexias according to breakdown in specific stages of the reading process. The linguistic concepts of surface structure versus the deep meanings of words have been instrumental in these new classifications. Four patterns of alexia (or “dyslexia” in British usage) have been recognized: letter-by-letter, deep, phonological, and surface dyslexia. Fig. 12A.8 diagrams the steps in the reading process and the points of breakdown in the four syndromes. Letter-by-letter dyslexia is equivalent to pure alexia without agraphia. Deep dyslexia is a severe reading disorder in which patients recognize and read aloud only familiar words, especially concrete, imageable nouns and verbs. They make semantic or visual errors in reading and fail completely in reading nonsense syllables or nonwords. Word reading is not affected by word length or regularity of spelling; one patient, for example, could read “ambulance” but not “am.” Most cases are characterized by severe aphasia, with extensive left frontoparietal damage.

Fig. 12A.8 Neurolinguistic model of the reading process. Evidence from studies of the alexias points to three separate routes to reading: 1, the phonological (or grapheme-phoneme conversion) route; 2, the semantic (or lexical-semantic-phonological) route; and 3, the nonlexical phonological route. In deep dyslexia, only route 2 can operate; in phonological dyslexia, 3 is the principal pathway; in surface dyslexia, only 1 is functional.

(Adapted with permission from Margolin, D.I., 1991. Cognitive neuropsychology. Resolving enigmas about Wernicke’s aphasia and other higher cortical disorders. Arch Neurol 48, 751-765.)

Phonological dyslexia is similar to deep dyslexia, with poor reading of nonwords, but single nouns and verbs are read in a nearly normal fashion, and semantic errors are rare. Patients appear to read words without understanding. The fourth type, surface dyslexia, involves spared ability to read laboriously by grapheme-phoneme conversion but inability to recognize words at a glance. These patients can read nonsense syllables but not words of irregular spelling, such as “colonel” or “yacht.” Their errors tend to be phonological rather than semantic or visual (e.g., pronouncing “rough” and “though” alike).

Agraphia

Like reading, writing may be affected either in isolation (pure agraphia) or in association with aphasia (aphasic agraphia). In addition, writing can be impaired by motor disorders, apraxia, and visuospatial deficits. Isolated agraphia has been described with left frontal or parietal lesions.

Agraphias can be analyzed in the same way as alexias (Fig. 12A.9). Thus, phonological agraphia involves the inability to convert phonemes into graphemes or to write pronounceable nonsense syllables in the presence of ability to write familiar words. Deep dysgraphia is similar to phonological agraphia, but the patient can read nouns and verbs better than articles, prepositions, adjectives, and adverbs. In lexical or surface dysgraphia, patients can write regularly spelled words and pronounceable nonsense words but not irregularly spelled words. These patients have intact phoneme-grapheme conversion but cannot write by a whole-word or “lexical” strategy.

Fig. 12A.9 Neurolinguistic model of writing and the agraphias. In deep agraphia, only the semantic (phonological-semantic-lexical) route (1) is operative. In phonological agraphia, route 2, the nonlexical phonological route, produces written words directly from spoken words. In surface agraphia, only route 3, the phoneme-grapheme pathway, can be used to generate writing.

Preliminary Evaluation

The bedside language examination is useful in forming a preliminary impression of the type of aphasia and the localization of the causative lesion. Follow-up examinations also are helpful; as in all neurological diagnosis, the evolution of a neurological deficit over time is the most important clue to the specific disease process. For example, an embolic stroke and a brain tumor both may produce Wernicke aphasia, but strokes occur suddenly, with improvement thereafter, whereas tumors produce gradually worsening aphasia.

Other Useful Tests

In addition to the bedside examination, a large number of standardized aphasia test batteries have been published. The physician should think of these tests as more detailed extensions of the bedside examination. They have the advantage of quantitation and standardization, permitting comparison over time and, in some cases, even a diagnosis of the specific aphasia syndrome. Research on aphasia depends on these standardized tests.

For neurologists, the most helpful battery is the Boston Diagnostic Aphasia Examination or its Canadian adaptation, the Western Aphasia Battery. Both tests provide subtest information analogous to that obtained with the bedside examination, and therefore meaningful to neurologists, as well as aphasia syndrome classification. The Porch Index of Communicative Ability quantitates performance in many specific functions, allowing comparison over time. Other aphasia tests are designed to evaluate specific language areas. For example, the Boston Naming Test evaluates a wide variety of naming stimuli, whereas the Token Test evaluates higher-level comprehension deficits. Further information on neuropsychological tests can be found in Chapter 34.

More specific diagnosis in the aphasic patient rests on the confirmation of a brain lesion by neuroimaging (Fig. 12A.10). The CT brain scan (discussed in Chapter 33A) revolutionized the localization of aphasia by permitting “real-time” delineation of a focal lesion in a living patient; previously, the physician had to outlive the patient to obtain a clinical-pathological correlation at autopsy. MRI provides better resolution of areas difficult to see on CT images, such as the temporal cortex adjacent to the petrous bones, and more sensitive detection of tissue pathology, such as early changes of infarction. The anatomical distinction of cortical from subcortical aphasia is best made by MRI. Acute strokes are visualized early on diffusion-weighted MRI.

Fig. 12A.10 A, Coronal T1-weighted magnetic resonance imaging study of the brain of a patient with primary progressive aphasia. Note the marked atrophy of the left temporal lobe. B, Axial fluorine-2 deoxyglucose positron emission tomography (PET) scans showing extensive hypometabolism in the left cerebral hemisphere, especially marked in the left temporal lobe.

The electroencephalogram (EEG) is helpful in aphasia in localizing seizure discharges, interictal spikes, and slowing seen after destructive lesions such as traumatic contusions and infarctions. The EEG can provide evidence that aphasia is an ictal or a postictal phenomenon and can furnish early clues in aphasia secondary to mass lesions or herpes simplex encephalitis. In research applications, electrophysiological testing via subdural grid and depth electrodes or stimulation mapping of epileptic foci in preparation for epilepsy surgery have aided in the identification of cortical areas involved in language.

Cerebral arteriography is useful in the diagnosis of aneurysms, arteriovenous malformations (AVMs), arterial occlusions, vasculitis, and venous outflow obstructions. In preparation for epilepsy surgery, the Wada test, or infusion of amobarbital through an arterial catheter, is useful in the determination of language dominance. Other related studies using language activation with functional MRI (fMRI) or PET are beginning to rival the Wada test for the study of language dominance (Abou-Khalil and Schlaggar, 2002).

Single photon emission CT (SPECT), PET, and functional MRI (fMRI; see Chapter 33C) are contributing greatly to the study of language. Patterns of brain activation in response to language stimuli have been recorded, mainly in normal persons, and these studies have largely confirmed the localizations based on clinicopathological findings in disorders such as stroke over the past 140 years. In addition, these techniques can be used to map areas of the brain that activate during language functions after insults such as strokes, and the pattern of recovery can be studied. Some such studies have indicated right hemisphere activation in patients recovering from aphasia (Cappa et al., 1997), whereas others have found that only left hemisphere activation is associated with full recovery (Heiss et al., 1999). A recent fMRI study (Saur et al., 2006) has suggested hypometabolism in the language cortex shortly after an ischemic insult, followed by increased activation of homologous areas in the contralateral hemisphere, and then a shift back to the more normal pattern of left hemisphere activation. Subcortical contributions to aphasia and language in degenerative conditions have been studied with PET. These techniques provide the best correlation between brain structure and function currently available and should help advance our understanding of language disorders and their recovery.