摘要
fMRI Study in Insight Problem Solving Using Japanese Remote Associates Test Based on Semantic Chunk Decomposition Hitoshi Terai (terai@is.nagoya-u.ac.jp) Graduate School of Information Science, Nagoya University, Nagoya, Aichi, 464–8601 Japan CREST, Japan Science and Technology Agency, Chiyoda, Tokyo, 102–8666 Japan Kazuhisa Miwa (miwa@is.nagoya-u.ac.jp) Graduate School of Information Science, Nagoya University, Nagoya, Aichi, 464–8601 Japan Kazuaki Asami (asami@cog.human.nagoya-u.ac.jp) Graduate School of Information Science, Nagoya University, Nagoya, Aichi, 464–8601 Japan studies are premised on using English speaking participants. Therefore, insight tasks based on a wide variety of languages are required to promote the development of research in this field. Thus, our research purpose is to develop a Japanese version of the RAT for use in brain research using examples from the standard RAT task widely used in current neuroscience studies. In addition to developing the task, we tried to reveal the relationship between brain activities and the process of insight problem solving, including both processes of an im- passe and evoking an emotional experience when solutions are found. Abstract This study aims to develop a Japanese version of the remote associates test required to decompose semantic chunks for use not only in behavioral studies but also in brain researches. Fur- ther, this study attempted to reveal the relationship between the process of solving insight problems and brain activities. Re- sults of the behavioral data show that the solution time was sig- nificantly longer in the chunked than in the non-chunked con- dition. The imaging data identified the following brain activi- ties. First, the right and left cingulate gyri, related to conflict monitoring, were more activated during the process of search- ing for a target in the chunked than in the non-chunked condi- tion. Second, the left posterior cingulate gyrus was more acti- vated when the participants could find a target by overcoming constraints as semantic chunks related to emotional process. Keywords: fMRI; insight; remote associates test; chunk de- composition. Japanese RAT Required to Decompose Semantic Chunks Introduction Reproductive thinking is the application of previously ac- quired knowledge and efficiently solves typical problems. However, in insight problems, such reproductive thinking forms mental sets, preventing solution and leading problem solvers to an impasse (Dominowski & Dallob, 1995; Ohls- son, 1992; Smith, 1995). Thus, problem solving is accom- plished by overcoming such reproductive thinking. Problem solvers often have an “aha!” experience when solving insight problems (Davidson, 1995; Metcalfe, 1986a,b; Metcalfe & Wiebe, 1987). Such characteristics of insight process, in- cluding an impasse and sudden attainment with the emotional experience, have been studied through psychological exper- iments. Moreover, in recent years, many studies revealed the mechanisms between characteristic insight processes and brain functions (Dietrich & Kanso, 2010). Insight problems generally used for related researches (e.g., nine dots problem, candle problem, triangle of coins) require at least several minutes to solve them. Further, once they are solved, they cannot be reused as problem solving tasks among the same participants. In contrast, when employ- ing brain imaging studies, verbal problems including ana- gram tasks, riddles, and the remote associates test (RAT) have been used. This is owing to the fact that these studies require not only the use of problems that can be solved within several tens of seconds but also can be used repeatedly for the same participants. However, almost all tasks used in brain imaging Insight problem solving characterized by an impasse and the suddenness of solution with emotional experience represents a radical representation change. Problem solvers have to re- construct their erroneous mental representations constructed at an early stage of insight problem solving, whereas they can take step-by-step analytic approaches in non-insight prob- lems. Such representation change in the insight process has been interpreted based on the theories such as the transi- tion of problem spaces and the chunk decomposition by con- ducting psychological experiments (Kaplan & Simon, 1990; Knoblich et al., 1999; Ohlsson, 1992). Familiarity with a class of objects and events leads to the creation of patterns as chunks that capture recurring constellations of features or components. Preserving the mental efforts by using chunked knowledge contributes to efficient problem solving. How- ever, if the available chunk does not work in a way that is helpful vis-`a-vis finding a solution, it might work to prevent solving the problem. Moreover, once it is constructed, it is difficult to decimate and an impasse might result (Knoblich et al., 1999; Ohlsson, 1992). RAT consists of sets of three words drawn from a mutually remote associate cluster. Problem solvers are required to find a fourth word which could serve as a specific kind of asso- ciative connective link between these disparate words (Med- nick, 1962). One example might be a set of three problem words: “arm,” “coal,” “peach.” The answer to the example is the word “pit.” The answer word generate three words or