Representation and Integration of Semantic Information
One of the major challenges in the study of the human brain and mind is to understand the nature and organization of lexical semantic representations, or the representation of the meaning of words. There have been many empirical investigations into the nature of lexical semantic representations, but the question of how and where they are represented is still a matter of debate. In ERP priming studies with brain lesioned patients we compared priming for associatively (e.g., bread-butter) or purely semantically related words that are not associated (from the same semantic category, e.g., horse-bear). We assessed the N400 component of the ERP signal, which is an index of semantic processing difficulty and is sensitive to inter-word relationships. Patients with left IFG damage showed no reduction or delay of the N400 component of the ERP. In contrast, patients with lesions including posterior temporal areas of the left hemisphere showed reduced N400 effects for both types of relations. Finally, non-aphasic patients with lesions that included posterior temporal regions of the right hemisphere showed normal associative priming, but no N400 priming effect for purely semantic relations between words. This pattern of results shows that posterior temporal regions of the left and right hemisphere contribute to the processing of lexical semantic relations between words.
Our research also showed that the brain represents verbal and image based information separately. In our studies there were two key manipulations: 1) the semantic relationship of the prime and target (related or unrelated); and 2) the imageability of the words (high, e.g., banana or low, e.g., justice). We found that effects of imageability are independent of effects of context both in an ERP study and in an fMRI study. The fMRI study showed a network of areas that were more active to high imageable words than to low imageable words (i.e., imageability effect), and a separate network that showed greater activity in response to unrelated than related words (i.e., priming effect). These brain regions included middle temporal gyrus and inferior frontal gyrus. Importantly, the effects of priming and imageability did not interact, and the activated areas were non-overlapping. These results show that separate neural systems embody verbal and image based semantic representations.
We are currently investigating the nature of the image based semantic processing system by comparing imageability effects when critical words are processed without awareness (in the so-called attentional blink), and with awareness.
Exploring the Mind Speaker Series - Dr. Tirin Moore, Stanford
CMB Acting Director Andy Yonelinas Featured in Sacramento Bee