Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Simultaneously Uncovering the Patterns of Brain Regions Involved in Different Story Reading Subprocesses

Figure 4

Map of the patterns of representation compared with the regions involved in sentence processing: our method recovers similar regions and differentiates them according to which information process they represent.

a- Adapted from [8]: Top: A recently published probabilistic overlap map showing where sentence reading generates greater neural activity than perceiving nonword letter strings. The value in each voxel indicates how many of the 25 individual subjects show a significant (at , FDR-corrected) effect for the SentencesNonwords contrast. Bottom: The main functional parcels derived from the probabilistic overlap map using an image parcellation (watershed) algorithm, as described in more detail in [7]. b- Results obtained by our generative model, showing where semantic, discourse, and syntax information is encoded by neural activity. Note this model identifies not just where language processing generates neural activity, but also what types of information are encoded by that activity. Each voxel location represents the classification done using a cube of voxel coordinates, centered at that location, such that the union of voxels from all subjects whose coordinates are in that cube are used. Voxel locations are colored according to the feature set that can be used to yield significantly higher than chance accuracy. Light green regions, marked with (1), are regions in which using either semantic or syntactic features leads to high accuracy. Dark gray regions, marked with (2), are regions in which using either dialog or syntactic features leads to high accuracy.

Figure 4