Update

What's right about the neural organization of sign language? A perspective on recent neuroimaging results

Since the discovery of mirror neurons, there has been a great deal of interest in understanding the relationship between perception and action, and the role of the human mirror system in language comprehension and production. Two questions have dominated research. One concerns the role of Broca's area in speech perception. The other concerns the role of the motor system more broadly in understanding action-related language. The current study investigates both of these questions in a way that bridges research on language with research on manual actions. We studied the neural basis of observing and executing American Sign Language (ASL) object and action signs. In an fMRI experiment, deaf signers produced signs depicting actions and objects as well as observed/comprehended signs of actions and objects. Different patterns of activation were found for observation and execution although with overlap in Broca's area, providing prima facie support for the claim that the motor system participates in language perception. In contrast, we found no evidence that action related signs differentially involved the motor system compared to object related signs. These findings are discussed in the context of lesion studies of sign language execution and observation. In this broader context, we conclude that the activation in Broca's area during ASL observation is not causally related to sign language understanding.

Sign languages are complex human languages that exhibit linguistic structuring at all levels: phonology (form), morphology, syntax, and semantics. The neural systems that support signed and spoken languages are similar, but not identical. Sign production recruits the left inferior frontal gyrus (IFG), including Broca's area, as well as the left superior parietal lobule. Sign comprehension engages bilateral superior temporal cortex and IFG. Parietal cortex appears to play a unique role in the expression of spatial language due to the iconic use of the hands to depict the location and movement of objects.

Our current understanding of the neural systems underlying linguistic communication has been largely derived through the study of spoken languages. However, human languages are not limited to the oral–aural modality as naturally occurring manual-visual sign languages attest. The existence of sign languages used in deaf communities provides an opportunity to assess the contributions of general linguistic processing and the effects of the modality of language expression in crafting the neural systems for language. Neuroimaging and aphasia studies of deaf signers reveal great commonalties in the neural systems used for sign and speech and provide evidence for a core neurobiological substrate for human linguistic communication. Also observed are cases of modality-specific patterns of brain activation and modality-specific language impairments. These force us to entertain the possibility of cross-linguistic differences in the neural instantiation of language. Taken together, these studies advance our understanding of the neurobiological substrates of human language.

Using functional magnetic resonance imaging (fMRI), we neuroimaged deaf adults as they performed two linguistic tasks with sentences in American Sign Language, grammatical judgment and phonemic-hand judgment. Participants’ age-onset of sign language acquisition ranged from birth to 14 years; length of sign language experience was substantial and did not vary in relation to age of acquisition. For both tasks, a more left lateralized pattern of activation was observed, with activity for grammatical judgment being more anterior than that observed for phonemic-hand judgment, which was more posterior by comparison. Age of acquisition was linearly and negatively related to activation levels in anterior language regions and positively related to activation levels in posterior visual regions for both tasks.

Vocabulary acquisition is such a major aspect of language learning in children, but also in adults when learning a foreign language, that a dedicated vocabulary learning device may exist within the language organ. To identify the relevant brain systems, we performed regional cerebral blood flow measurements in normal subjects while they were learning a list of neologisms or a list of word–nonword pairs. Structures implicated in phonological short-term memory (Broca's area, left temporo-parietal junction) were steadily activated during nonwords learning, while the left temporal lobe neocortical and paralimbic structures (parahippocampal region), associated with long-term memory, contributed to learning in a time-dependent manner, with maximal activation at the beginning of the process. The neural system specifically activated when learning new vocabulary was strongly lateralized to the left hemisphere. This evidence refines current models of memory function and supports theories which emphasise the importance of phonological competence in hemispheric dominance for language.

We examine the hemispheric organization for the production of two classes of ASL signs, lexical signs and classifier signs. Previous work has found strong left hemisphere dominance for the production of lexical signs, but several authors have speculated that classifier signs may involve the right hemisphere to a greater degree because they can represent spatial information in a topographic, non-categorical manner. Twenty-one unilaterally brain damaged signers (13 left hemisphere damaged, 8 right hemisphere damaged) were presented with a story narration task designed to elicit both lexical and classifier signs. Relative frequencies of the two types of errors were tabulated. Left hemisphere damaged signers produced significantly more lexical errors than did right hemisphere damaged signers, whereas the reverse pattern held for classifier signs. Our findings argue for different patterns of hemispheric asymmetry for these two classes of ASL signs. We suggest that the requirement to encode analogue spatial information in the production of classifier signs results in the increased involvement of the right hemisphere systems.