Poster C138, Sunday, March 26, 5:00 – 7:00 pm, Pacific Concourse
Does the brain have a domain-general mechanism for representing mental models?
Katherine L. Alfred1, Andrew C. Connolly1, Joshua S. Cetron1, David J. M. Kraemer1; 1Dartmouth College
Research on deductive reasoning has highlighted the usefulness of constructing mental models to represent knowledge about the world. Neuroimaging research has indicated the specific involvement of superior parietal cortex (SPC) during tasks that require deductive reasoning with mental models. These results are consistent with the notion of a domain-general deductive reasoning mechanism that relies on spatial computations in parietal cortex. However, this research has left open at least two key questions: 1) Does SPC activity reflect the mental model or deductive reasoning (or both)?, and 2) Is this parietal reasoning mechanism domain-general? A previous study from our lab used multivariate representational similarity analysis (RSA) to reveal the structure of neural patterns of activity in SPC with respect to a newly-learned mental model generated through linear syllogistic reasoning. These results indicated that RSA can be used to examine the structure of a newly-learned representational space and that SPC supports the representation of a mental model even outside of an explicit reasoning task. However, the content in this previous study, as in most deductive reasoning studies, was visuospatial in nature. Here we replicate and extend these findings using a task in which participants generate three distinct mental models (linear hierarchies) through a transitive reasoning task. The relationships within the three models range from being expressly visuospatial to entirely abstract. Whereas univariate fMRI results reveal that models generated from all three tasks elicit activity in SPC, RSA reveals more fine-grained similarities and differences in the neural representations of the models across different domains.
Topic Area: THINKING: Reasoning