Insight problem-solving is a creative cognitive process that involves overcoming mental impasses and discovering novel solutions. However, the neural mechanisms underlying insight, particularly in spatial manipulation task remain poorly understood. In this study, we investigated the brain dynamics involved in solving matchstick arithmetic problems, a type of spatial insight problem, using functional magnetic resonance imaging (fMRI). We applied two complementary analysis methods: (1) a general linear model (GLM) to identify brain regions associated with different problem-solving strategies (quick, analytical, and insight-based solutions), and (2) a hidden Markov model (HMM) to estimate discrete brain states during the problem-solving process. Our findings reveal that the default mode network (DMN) was more active during insight-based solutions than during quick and analytical strategies, whereas the executive control network (ECN) exhibited increased activation during quick and analytical solutions. These results suggest distinct roles for the DMN and ECN in spatial insight problem-solving. Additionally, we characterized specific brain states for each problem-solving strategy using HMM and examined their relationships with behavioral performance and state transitions. Our results suggest that insight problem-solving involves dynamic interactions between large-scale brain networks, with different strategies corresponding to different brain states. Notably, the high variability in brain state dynamics observed during the prolonged process of insight solutions may reflect increased cognitive flexibility. This study offers novel insights into the neural basis of spatial insight and its temporal dynamics.
OGAWA et al. (Sat,) studied this question.