While problem-solving is central in our daily life, its underlying mechanisms remain elusive. The dominant theory states that one must restructure a problem (i.e., reorganize problem-related representations) to solve it. As empirical evidence supporting this mechanism is scarce, we used network science methodology to demonstrate the key role of restructuring in problem-solving. Individual semantic memory networks were estimated before and after participants attempted to solve a riddle. These networks represent the organization of solution-relevant and irrelevant words as nodes, with edges representing the strength of the relationship between them based on the participants’ relatedness judgments. Restructuring was quantified as the difference in semantic network metrics between pre- and post-solving phases. Successful problem-solving was predicted by local semantic network restructuring, only in edges and nodes assessed as helpful to solve the riddle. These results shed new light on the mental restructuring in problem-solving and provide a new method to quantify it.

How creative ideas emerge remains unclear. We designed an associative fluency task based on polysemous words to explore the neurocognitive correlates of semantic search related to creativity. We distinguished two search components by assessing clustering and switching performance from participant’s responses. We related these components to creativity, executive abilities, and semantic memory structure explored via semantic networks, and identified their predictive functional brain connectivity patterns. Clustering correlated with divergent thinking, and relied on interactions between control, salience, and attentional networks. Switching correlated with the ability to combine remote associates, memory structure, and executive abilities, and was predicted by connectivity between and within the default, control, and salience networks. These results suggest that switching captures interactions between memory structure and control processes guiding the search, while clustering may capture controlled processes related to persistent search. These findings shed new light on the neurocognitive mechanisms of semantic search supporting creativity.

The Remote Associates Test is a creativity task that consists in finding a word that connects three unrelated words. We examined EEG power in healthy participants performing an adaptation of this task that allowed us to explore two mechanisms: remote semantic combination by varying the associative strength between the cue words and the solution, and the insight phenomenon reported by the subjects. Time-frequency analyses revealed that associative remoteness was associated with early synchronization in alpha and beta bands in laterotemporal and temporoparietal clusters and late frontal activity in the theta band just before the participants found the solution. Insight was associated with synchronization in alpha and gamma bands in inferotemporal clusters and a frontal synchronization in the theta band just preceded the response. Our findings provide new insights on the dynamic mechanisms involved in this verbal creativity task.