AbstractWhile the multiple-demand network plays an established role in cognitive flexibility, the role of default mode network is more poorly understood. In this study, we used a semantic feature matching task combined with multivoxel pattern decoding to test contrasting functional accounts. By one view, default mode and multiple-demand networks have opposing roles in cognition; consequently, while multiple-demand regions can decode current goal information, semantically-relevant default network regions might decode conceptual similarity irrespective of task demands. Alternatively, default mode regions might show sensitivity to changing task demands like multiple-demand regions, consistent with evidence that both networks dynamically alter their patterns of connectivity depending on the context. Our task required participants to integrate conceptual knowledge with changing task goals, such that successive decisions were based on different features of the items (colour, shape and size). This allowed us to simultaneously decode semantic category and current goal information using a whole-brain searchlight decoding approach. As expected, multiple-demand regions represented information about the currently-relevant conceptual feature, yet similar decoding results were found in default mode network regions, including angular gyrus and posterior cingulate cortex. Semantic category irrespective of task demands could be decoded in lateral occipital cortex, but not in most regions of default mode network. These results show that conceptual information related to the current goal dominates the multivariate response within default mode network. In this way, default mode network nodes support flexible memory retrieval by modulating their response to suit active task goals, alongside regions of multiple-demand cortex.Significance StatementWe tested contrasting accounts of default mode network (DMN) function using multivoxel pattern analysis. By one view, semantically-relevant parts of DMN represent conceptual similarity, irrespective of task context. By an alternative view, DMN tracks changing task demands. Our semantic feature matching task required participants to integrate conceptual knowledge with task goals, such that successive decisions were based on different features of the items. We demonstrate that DMN regions can decode current goal, alongside multiple-demand regions traditionally associated with cognitive control. The successful decoding of goal information plus largely absent category decoding effects within DMN indicates that this network supports flexible semantic cognition.
Recognition during recall failure: Semantic feature matching as a mechanism for recognition of semantic cues when recall fails
