Neural Underpinnings Of Divergent Production Of Rules In Numerical Analogical Reasoning

Neural underpinnings of human creativity in diverse domains have been studied over decades. Creativity plays an important role in our work and life nowadays. Recent years researchers have generated increasing interest in the nature and mechanisms of creativity, particularly its neural mechanisms. Some neuroimaging studies have found that creative problem solving involves prefrontal (e.g., Fink et al.,2009; Green et al.,2012; Zhang et al.,2013), parietal (Fink et al.,2009; Sieborger et al.,2007), temporal (Chavez-Eakle et al.,2007; Fink et al.,2009), and visual regions (e.g., Howard-Jones et al.,2005). In these studies most of experimental tasks are verbal tasks, not numerical stimuli. Unlike verbal processing, number processing possesses different mental representations (Dehaene,2005) and different neural mechanisms (Jia et al.,2011; Kadosh et al.,2005; Piazza et al.,2002; Richland et al.,2007; Zhong et al.,2011). Very little is known about neural mechanisms of numerical creativity.In the present study, we employed the numerical analogical reasoning task and functional Magnetic Resonance Imaging (MRI) to investigate neural correlates of divergent production of rules in numerical analogical reasoning. Three different conditions were used to investigate neural bases of numerical divergent thinking in analogical reasoning:single solution analogical reasoning tasks, multiple solution analogical reasoning tasks, and recognition tasks. The format of analogical reasoning tasks in our study was " B? and a prompt indicating the type of a task was above each stimulus trial. There was only one answer in single solution analogical reasoning tasks. For example, a stimulus trial of analogical reasoning tasks was the only answer. On the contrary, there were two or more than two answers in multiple solution analogical reasoning tasks. For example,"4","3" and "2" were the answers of a numerical analogical reasoning task Each answer of analogical reasoning tasks corresponded to a specific rule that existed in analogical reasoning tasks. To prevent participants from using repeatedly the same several rules to solve all the tasks, we designed experimental tasks so that participants had to use a large number of different abstract quantity rules to complete multiple solution analogical reasoning tasks, apart from visual location relations appearing once in each of all the multiple solution analogical reasoning tasks. However, visual location relations appeared in each of all the single solution analogical reasoning tasks. Recognition tasks, as baseline, had a similar format of analogical reasoning tasks. The different was that participants were requested to identify whether the first three items (A, B, C) had the same digit or not. For example, a baseline stimulus needed participants to respond "yes", because in the first three items there was the same digit "4".Behavioral Data:during the scanning reaction time and completion rate were recorded. Statistical analyses of behavioral data indicated that there were significant difference in the mean reaction time for performing multiple solution analogical reasoning tasks (M=7538ms, SEM=325ms), single solution analogical reasoning tasks (M=3020ms, SEM=169ms) and baseline tasks (M=2958ms, SEM=147ms). F (2,19)=227.7, P<0.001. In addition, there were significant differences in completion rate for performing multiple solution analogical reasoning tasks (M=90%, SEM=2.1%), single solution analogical reasoning tasks (M=98%, SEM=0.2%) and baseline tasks (M=88%, SEM=1.5%). F (2,19)=13.14, p<0.001. This suggested that multiple solution analogical reasoning tasks involve more mental operations so that such tasks require longer reaction time and have lower completion rate.The present neuroimaging data on participants performed multiple solution analogical reasoning tasks versus single solution analogical reasoning tasks revealed that divergent production of rules involves significant activations at Brodmann area (BA)10in the right middle frontal cortex, BA40in the left inferior parietal lobule, and BA8in the superior frontal cortex.The results suggest that the right BA10and the left BA40are involved in the generation of novel rules, and the BA8is associated with the inhibition of initial rules in numerical analogical reasoning. The findings might shed light on neural mechanisms of creativity in numerical processing.