Neural Correlates of Insight Phenomena

Difficult problems are sometimes solved in a sudden flash of illumination, a phenomenon referred to as “insight.” Recent neuroimaging studies have begun to reveal the neural correlates of the cognitive processes underlying such insight phenomena (Luo and Niki 2003; Jung-Beeman et al. 2004; Luo et al. 2004a, 2006; Mai et al. 2004; Lang et al. 2006). However, researchers have encountered a number of difficulties in applying neuroimaging methods to investigate insight. We will outline these difficulties, define general criteria brain-imaging studies of insight should meet, and then discuss in detail to what extent these criteria have been met in recent attempts to unravel the brain bases of insight.

One main difficulty that is well known from behavioral research is to produce insight phenomena in the laboratory. Even in purely behavioral studies it is hard to be certain whether participants in laboratory settings actually have insights or whether they solve problems in a more stepwise manner. A related problem is the small numbers of problems that are available to study insight (Bowden et al. 2005). In addition, the well-known classical insight problems, such as the nine-dot problem (Scheerer 1963), the two string problem (Maier 1930), and the candle problem (Duncker 1945), greatly vary in their sources of difficulty (Kershaw and Ohlsson 2004). This raises the question whether the data obtained in laboratory research can be generalized. Furthermore, studying how people solve single problems normally does not produce very reliable data. For this reason, recent experimental paradigms addressing the brain bases of insight have adopted so-called mini-insight problems (Bowden et al. 2005). These problems can normally be solved in a short period with or without external help and many exemplars can be created (Luo and Niki 2003; Jung-Beeman et al. 2004).