Decoding of concepts within and across semantic categories
No Thumbnail Available
Currently, there is no agreement about how conceptual knowledge is represented in the brain. Studies on semantic dementia suggest that the anterior temporal lobe (ATL) functions as the semantic hub, combining information distributed throughout the brain and constructing abstract semantic representations. However, functional neuroimaging studies argue that the semantic hub is located more posteriorly, suggesting that the posterior medial temporal gyrus (pMTG) might be a better candidate for the semantic hub than the ATL. The goal of this functional magnetic resonance imaging (fMRI) study was to examine the temporal lobe’s role in the organisation of semantic knowledge in the brain. We expected to find a posterior-to-anterior gradient of specificity of semantic representations, where more coarse categorical information is processed in the pMTG and more detailed representations are computed in the ATL. We used spoken and written Dutch words representing either basic or subordinate names for categories of animals and buildings (e.g., dog/house, retriever/chihuahua). The combined results from the whole brain standard univariate analysis and searchlight-based multivariate pattern analysis (MVPA) revealed distributed brain activation associated with the semantic processing and within-/across-category discrimination of individual written and spoken words. However, our results were challenged by high variability across subjects, and possible limitations in the experimental design. Therefore, we were not able to obtain an expected activation in the ATL.
Faculteit der Sociale Wetenschappen