Both approaches yielded representations of rule-based associations within left AG.
That is, while representations of low-level visual features were shared between days, representations of rule-based associations were not.
That is, while low-level visual features showed shared representations between days and were detectable on day 1, representations of rule-based associations were not.
First, we identified voxel patterns that discriminated between schema conditions (spatial vs. non-spatial), and that served as a marker for representations of rule-based associations.
Again, we did not find significant representations of rule-based associations, and low-level visual features were mostly represented in occipital regions (Table 6, lower part).
We did not find any significant representations of rule-based associations within the AG or any other brain region on day 1.
We did not find significant representations of rule-based associations within the AG or any other brain region on day 1 (critical cluster size = 75 voxels).
Results, Multi-voxel representations of schema components: "Next, we reasoned that if schema components converged in the AG only after consolidation, the multi-voxel representations of rule-based associations should not generalize from day 2 to day 1.
For example, it has been suggested that the frontal cortex mediates rule-selection in contrast to posterior representations of rule meaning (Bunge et al. 2003; Donohue et al. 2005).
Thus, applying a classifier (spatial vs. non-spatial) trained on data from day 2 to neural data of the transfer test should yield representations of rule-based associations within AG.
Thus, training a classifier on data from day 2 and testing it on neural data from the transfer test should yield representations of rule-based associations and low-level visual features within the AG.
