Neural Representation of Rules at Different Hierarchical Levels

Humans use rules to organise their actions to achieve specific goals. For simple situations, single rules can be used to link a sensory stimulus to its appropriate response. More complex situations, however, require the application of multiple rules organised in hierarchies, where high-level rules influence the application of low-level rules. Several theories suggest that different regions along the anterior-to-posterior axis of Prefrontal Cortex (PFC) control rules at progressively higher levels in the hierarchy of action control. In this study, we directly investigated in which areas rules at different hierarchical levels are represented. Participants (N = 16) had to simultaneously represent and apply two rules: one being a low-level stimulus-response rule, the other being a higher-level modifier rule. We used multivariate pattern analysis (MVPA) to separately explore which brain areas encode information about the identity of both the active low-level and high-level rule. The results show that rules at different hierarchical levels are encoded in distinct regions within a large-scale brain network comprising mainly parietal and lateral prefrontal areas. However, we did not find that areas were organised in an anterior-to-posterior gradient. These results are in line with recent findings showing that complex rules are decomposed in their defining features, where each feature is then represented in brain structures appropriate for the type of information conveyed. By contrast, our results do not support the existence of a gradient in which increasingly abstract rule representations are controlled by progressively more anterior areas in PFC.