The neural substrates of inferential and referential semantic processing.

INTRODUCTION: A distinction has been proposed, on theoretical grounds, between referential and inferential semantic abilities. The former account for the relationship of words to the world, the latter for the relationship of words among themselves. The hypothesis of, at least partially, different neurological underpinnings for this distinction has been supported by the presence of double dissociations in neurological patients between tasks that can be considered to tap the cognitive processes involving these two different classes of semantic knowledge, such as, for example, picture naming (referential) and naming to a verbal definition (inferential). METHODS: We report here the results of a functional magnetic resonance experiment, contrasting the pattern of brain activity associated with, respectively, "referential" (picture naming, word-to-picture matching) and "inferential" (naming to definition, word-to-word matching) tasks. RESULTS: All tasks activate an extensive set of brain areas involving both hemispheres, corresponding to the "common semantic network". In addition, left hemispheric temporal areas are selectively engaged by the inferential tasks. Conversely, a specific activation of the right fusiform gyrus is associated with the referential tasks. CONCLUSIONS: These findings suggest that while inferential tasks, as compared with referential tasks, engage additional processing resources subserved by left hemispheric language areas involved in lexical retrieval, referential tasks (as compared with inferential tasks) recruit right hemispheric areas generally associated with nonverbal conceptual and structural object processing. These findings are compatible with the double dissociations reported in neurological patients.

Conceptual knowledge in the brain: fMRI evidence for a featural organization.

The organization and representation of conceptual knowledge in the brain remains a controversial issue in terms of both neuropsychological and imaging evidence. We report the results of a functional magnetic resonance study in which the role of the most debated dimensions (domain and feature type) was evaluated through a concept-feature verification task. The scope of the task was to eliminate serious methodological concerns that weighed down previous imaging research in this area, and to allow more definitive conclusions regarding the specific contribution of these dimensions. The results show differential patterns of brain activity according to feature type (both motion and visual form/surface features) but not according to concept domain (living vs. nonliving things). These findings are in accord with a modality-specific account of conceptual knowledge organization in the brain, in which specific kinds of features (e.g. form, color, motion, etc) have differential importance for representing different concepts.

The neural substrate of naming events: effects of processing demands but not of grammatical class.

Grammatical class is a fundamental property of language, and all natural languages distinguish between nouns and verbs. Brain activation studies have provided conflicting evidence concerning the neural substrates of noun and verb processing. A major limitation of many previous imaging studies is that they did not disentangle the impact of grammatical class from the differences in semantic correlates. In order to tease apart the role of semantic and grammatical factors, we performed a functional magnetic resonance imaging study presenting Italian speakers with pictures of events and asked them to name them as 1) Infinitive Verb (e.g., mangiare [to eat]); 2) Inflected Verb (e.g., mangia [she/he eats]); and 3) Action Noun (e.g., mangiata [the eating]). We did not find any verb-specific activation. However, reliable left inferior frontal gyrus (IFG) activations were found when contrasting the Action Noun with the Infinitive Verb condition. A second-level analysis indicated then that activation in left IFG was greatest for Action Nouns, intermediate for Inflected Verbs, and least for Infinitive Verbs. We conclude that, when all other factors are controlled, nouns and verbs are processed by a common neural system. In the present case, differences in left IFG activation emerge as a consequence of increasing linguistic and/or general processing demands.

The role of semantics and grammatical class in the neural representation of words.

On the basis of neuropsychological and functional imaging evidence, meaning and grammatical class (particularly the verb-noun distinction) have been proposed as organizational principles of linguistic knowledge in the brain. However, previous studies investigating verb and noun processing have been confounded by the presence of systematic correlations between word meaning and grammatical class. In this positron emission tomography study, we investigated implicit word processing using stimuli that allowed the effects of semantic and grammatical properties to be examined independently, without grammatical-semantic confounds. We found that left hemisphere cortical activation during single-word processing was modulated by word meaning, but not by grammatical class. Motor word processing produced significant activation in left precentral gyrus, whereas sensory word processing produced significant activation in left inferior temporal and inferior frontal regions. In contrast to previous studies, there were no effects of grammatical class in left inferior frontal gyrus (IFG). Instead, we found semantic-based differences within left IFG: anterior, but not posterior, left IFG regions responded preferentially to sensory words. These findings demonstrate that the neural substrates of implicit word processing are determined by semantic rather than grammatical properties and suggest that word comprehension involves the activation of modality-specific representations linked to word meaning.

Semantic similarity and grammatical class in naming actions.

Italian speakers were asked to name pictures of actions (e.g. "bere", to drink). Pictures were presented at the same time as distracter words that were semantically related or unrelated to the picture names, and were of the same or different grammatical class (verbs or nouns). Half of the participants named the actions as verbs in citation form, the other half as verbs inflected for third person singular or plural. We found a reliable semantic interference effect. Crucially, we also observed a significant interaction between naming context and grammatical class: naming latencies were slower for verb distracters in the inflected form condition, but not in the citation form condition. The results are taken to provide evidence for the separability of semantics and grammatical class.

Questioning the living/nonliving dichotomy: evidence from a patient with an unusual semantic dissociation.

In this article the authors describe a patient (J.P.) whose category-specific naming deficit eluded the classical dichotomies between living versus nonliving items or visual versus functional attributes. At age 22, he had herpes simplex encephalitis followed by a left temporal lobectomy. J.P. was tested on measures of visual perception, category naming, fluency, and name-picture matching. He showed a severe impairment naming and identifying fruits, vegetables, and musical instruments. His performance with animals and birds was spared inconsistently, meaning that even the preserved categories were, at some point, affected. J.P.'s unusual deficit supports the hypothesis that semantic knowledge is organized in the brain on the basis of object properties, which can cut across the living-nonliving categorical distinction.

The breakdown of semantic knowledge: insights from a statistical model of meaning representation.

Investigations of patients with semantic category-specific deficits have revealed a wide range of performance and variability in categories that are impaired or spared; this variability presents a challenge to accounts of category specificity. Accounts based only on impairment to semantic features of a particular type (e.g., visual), as well as accounts based only on featural properties (e.g., feature intercorrelations), are insufficient to explain the variability of patients' performance. A first goal of the paper is to discuss how a hybrid account incorporating both a level of organization according to feature types (a level of nonlinguistic conceptual representations) and a level of organization dictated by featural properties may provide a more comprehensive account of the cases reported in the literature. The second and most novel goal of the study reported here is to derive from our hybrid account a series of novel predictions concerning the representation and impairment of a different domain of knowledge: knowledge of actions and events, a domain of knowledge that has received remarkably little attention to date.

An investigation of semantic errors in unimpaired and Alzheimer's speakers of Italian.

Semantic errors are a common type of slip of the tongue for normal speakers; they are also considered to be the hallmark of progressive diseases that affect semantic memory such as Alzheimer's Disease (AD) and semantic dementia. In unimpaired speakers, semantic errors have been shown to be affected by syntactic variables. For example, Marx (1999) has shown that speakers of a gendered language such as German tend to substitute a target with another word that shares the same grammatical gender with the target more often than chance would predict. This finding suggests that errors occur at a level at which lexical information about the target is activated and retrieved. Here, we assess whether such an effect of syntactic variables (gender) also holds in experimental tasks designed to induce errors in unimpaired speakers of Italian (another gendered language) and whether this effect can also be observed in the semantic errors produced in the same task by AD patients. We found that for the unimpaired speakers, the gender of the target word constrained the error committed, while this was not the case for the AD patients. We take this finding to suggest a different locus for the errors in the two populations: while the semantic errors by the unimpaired speakers occur because of mis-selection of a lexical entry due to lexical competition among semantically similar words, the errors by the AD patients occur because of insufficient activation of lexical representations.

Role of the anterior temporal lobe in repetition and semantic priming: evidence from a patient with a category-specific deficit.

Neuroimaging studies in healthy participants have implicated anterior temporal lobe regions and the fusiform gyrus in repetition priming and semantic priming. Only the investigation of patients with selective lesions, however, can establish the necessity of these particular regions. To this end, we administered three tests of repetition priming (pseudoword identification; masked-form priming; category-exemplar generation) and a test of semantic priming to a patient (J.P.) with a category-specific deficit stemming from bilateral damage to the anterior fusiform gyrus and anterior temporal regions. On all of the repetition priming tasks, J.P. showed priming effects within 1 S.D. of 10 age- and education-matched CON; ANOVAs indicated no interaction between group and prime condition. These findings suggest that the anterior fusiform and anterior temporal lobe are not required for these priming effects. J.P. also showed normal repetition priming even for items that he had never been able to name or to provide semantic information about. On the semantic priming task, J.P. showed normal levels of priming across categories. When we separately analyzed his priming for items he could never name or access information about versus items that he had been able to name on at least two testing sessions, we found priming for the latter items, but not for the former. This result suggests that category-specific deficits resulting from damage to the anterior temporal lobes may disrupt the automatic, rapid access of semantic information of some items.