Hierarchical artificial grammar processing engages Broca's area.

The present fMRI study investigates the neural basis of hierarchical processing using two types of artificial grammars: one governed by rules of adjacent dependencies and the other by rules of hierarchical dependencies. The adjacent dependency sequences followed the rule (AB)(n), at which simple transitions between two types of syllable categories were generated (e.g. A(1)B(1)A(2)B(2)). The hierarchical syllable sequences followed the rule A(n)B(n), generating a center-embedded structure (e.g. A(2)A(1)B(1)B(2)) the learning of which required the processing of hierarchical dependencies. When comparing the processing of hierarchical dependencies to adjacent dependencies, significantly higher activations were observed in Broca's area and the adjacent rim of the ventral premotor cortex (BA 44/6) in addition to some several other cortical and sub-cortical regions. These results indicate that Broca's area is part of a neural circuit that is responsible for the processing of hierarchical structures in an artificial grammar.