Lexicon, Neural Basis of

The lexical processing system is the collection of mechanisms that are used to store and retrieve our knowledge of the words of the language. Knowing a word means knowing its meaning, its phonological and orthographic forms, and its grammatical properties. How is this knowledge organized and represented in the brain? Two types of evidence have been used to answer this question. The major source of evidence has been the patterns of lexical deficits associated with brain damage in aphasic patients. More recently, functional neuroimaging methods -- POSITRON EMISSION TOMOGRAPHY (PET) and functional MAGNETIC RESONANCE IMAGING (fMRI) -- have played an increasingly important role. Evidence from neuropsychological and neuroimaging studies has converged on one widely shared conclusion: the mental LEXICON is organized into relatively autonomous neural subsystems in the left hemisphere, each dedicated to processing a different aspect of lexical knowledge.

One of the classic syndromes of APHASIA, anomia -- a deficit in retrieving words for production -- provides prima facie evidence for distinct representation of meaning and of lexical forms in the brain. Studies of anomic patients have shown that they are unable to produce the names of objects despite normal ability to recognize and define them, indicating a selective deficit in processing lexical forms. These patients tend to have more narrowly circumscribed damage, involving most often the left temporal lobe, but sometimes the parietal or frontal lobe or both. There is also evidence that the semantic system can be damaged independently of knowledge of lexical forms. The latter evidence has been obtained both with patients who have sustained focal brain damage due to strokes and patients with degenerative disorders such as Alzheimer's disease. Both types of patients make semantic errors (e.g., they might produce "table" in naming a chair or "tastes good, a fruit" in naming a pear) in all lexical processing tasks. Patients with selective damage to the semantic component of the lexicon typically have extensive left hemisphere damage involving the temporal, parietal, and frontal lobes. Converging evidence in support of the view that semantic information is distributed widely in the left hemisphere has been obtained in functional neuroimaging studies with PET.

Some brain-damaged patients are selectively impaired in retrieving only the orthographic form (e.g., the spelling of the word chair) or only the phonological form of words (e.g., the sound of the word chair). Patients of this type can be entirely normal in their ability to understand and define words, but fail to retrieve the correct word form in one, but not the other, modality of output. These patterns of performance attest to the autonomy of phonological and orthographic lexical forms from each other and from meaning. Converging evidence for this conclusion comes from functional neuroimaging studies which have shown that distinct brain regions are activated when neurologically intact participants are engaged in processing the phonological (frontal-temporal) versus the orthographic (parietal-occipital) forms of words.

Damage to the semantic system can lead to disproportionate difficulties with specific semantic categories. The most frequently observed category-specific deficits have concerned the contrast between living and nonliving things. However, the deficits can be quite selective, affecting (or sparing) only animals or only plant life. The lesion sites typically associated with these deficits include the left temporal lobe, the posterior frontal lobe, and the inferior junction of the parietal and occipital lobes. The existence of semantic category-specific deficits was originally interpreted as reflecting a modality-based organization of conceptual knowledge in the brain. It was proposed that visual and functional/associative properties are represented in distinct areas of the brain and that these two sets of properties are differentially important in distinguishing between living and nonliving things, respectively. On this view, selective damage to one of the modality-specific knowledge subsystems would result in a semantic category-specific deficit. However, recent results have shown that category-specific deficits are not the result of damage to modality-specific but rather to modality-independent knowledge systems. These results, and the fact that the reliable categories of category-specific deficits are those of animals, plant life, artifacts, and conspecifics, have led to the proposal that conceptual knowledge is organized into broad, evolutionarily determined domains of knowledge. Functional neuroimaging results with neurologically intact participants have confirmed that the inferior temporal lobe and parts of the occipital lobe are activated in response to animal pictures and words, whereas more dorsal areas of the temporal lobe and parts of the frontal lobe are activated in response to artifacts.

One of the classic features of the speech of some aphasic patients is agrammatic production -- a form of speech characterized by a relative paucity of function or closed-class words (articles, prepositions, auxiliaries, etc.). The disproportionate difficulty in producing closed-class words in some patients is in contrast to patients who show the reverse pattern of dissociation -- selective difficulty with open-class words (nouns, verbs, and adjectives). But, the dissociations of lexical processing deficits can be even more fine-grained than that: some patients are disproportionately impaired in producing verbs while others are disproportionately impaired in producing nouns, and some patients can be disproportionately impaired in comprehending one or the other class of words. Grammatical class effects can even be restricted to one modality of output or input. For example, there are patients who are impaired in producing verbs only in speaking (they can write verbs and can produce nouns both in speaking and in writing) and patients who are impaired in producing nouns only in speaking; and there are patients who fail to understand written but not spoken verbs. The fact that grammatical class effects can also be modality-specific implies a close link between word form and grammatical information. These results challenge the view that there exists a modality-neutral lexical node mediating between modality-specific lexical representations and word meaning. Damage to the left frontal lobe is typically associated with disproportionate difficulty in processing verbs and closed-class words, while damage to the left temporal lobe is associated with disproportionate difficulty in producing and comprehending nouns. Recent investigations with PET and event-related potentials (ERPs) have confirmed this general characterization of the roles of the frontal and temporal lobes in processing words of different grammatical classes.

Brain damage can also selectively affect different parts of words, revealing their internal structure. It is now well established that some aphasic patients have no difficulty in processing the stem of words (e.g., walk in walked) but fail to retrieve their correct inflectional suffixes (e.g. the -ed in walked), and that some patients can process normally the morphological affixes of words but not their stems. This double dissociation in processing different types of morphemes implies that the units of lexical representation in the brain are stems and inflectional affixes, and not whole words. Detailed single-case studies of aphasic patients have confirmed this conclusion, and have shown that difficulties in processing inflectional morphology tend to be associated with damage to more frontal areas of the left hemisphere, while difficulties in processing the stems of words are more likely to be associated with temporal lobe damage.

Although we still do not have a detailed understanding of the neural substrates of the lexicon, its general outlines are beginning to emerge, and it looks to be as follows: (1) the lexical processing system is distributed over a large area of the left hemisphere, involving the temporal, frontal, and parietal lobes; (2) different parts of the left hemisphere are dedicated to the storage and computation of different aspects of lexical knowledge -- meaning, form, and grammatical information are represented autonomously; and (3) within each of the major components of the lexicon, the semantic and lexical form components, there are further fine-grained functional and neural distinctions.

Additional links

-- Alfonso Caramazza

Lexicon, Neural Basis of

See also

Additional links

Further Readings