Decoding the cascading architecture of creative ideation: an ERP study

Creative ideation (CI) refers to the process of generating original ideas in response to open-ended problems, with idea generation and evaluation being two key cognitive processes involved in CI. However, the current theoretical debate concerns how these two processes support CI. One perspective posits a serial model where idea generation and evaluation occur sequentially, while another argues for a parallel model where these stages operate simultaneously. To clarify this debate, the present study recorded EEG data from 150 participants performing the alternative uses task (AUT) and applied temporal-generalization multivariate pattern analysis (MVPA) to examine whether the temporal evolution pattern of CI is stable or dynamic. Results showed that the temporal-generalization pattern fell between the diagonal and rectangular patterns, suggesting partial overlap between the cognitive processes underlying CI. These findings suggest that CI does not strictly conform to the predictions of a serial or parallel model but instead aligns more closely with the cascading model, wherein the initiation of a subsequent process occurs before the completion of the preceding one, resulting in a partial overlap between them. Moreover, univariate analysis showed that the cascading of CI mainly occurred during the late component (LC) and post-LC stages, representing the temporal unfolding of idea generation and subsequent evaluation. Together, the present study clarifies the theoretical debate between the serial and parallel models of CI, reveals the cascading architecture of CI, and advances the mechanistic understanding of creative cognition.