Neural processes underlying intuitive coherence judgments as revealed by fMRI on a semantic judgment task Ruediger Ilg, a,b, ⁎ Kai Vogeley, c,d Thomas Goschke, e Annette Bolte, f Jon N. Shah, g Ernst Pöppel, a and Gereon R. Fink c,g,h a Institute of Medical Psychology, Ludwig-Maximilians-Universität München, Germany b Department of Neurology, Technische Universität München, Germany c Brain Imaging Centre West, Research Centre Juelich, Germany d Department of Psychiatry, University of Cologne, Germany e Department of Psychology, Dresden University of Technology, Germany f Institute of Psychology, Braunschweig University of Technology, Germany g Institute of Neurosciences and Biophysics, Research Centre Juelich, Germany h Department of Neurology, University of Cologne, Germany Received 30 April 2007; revised 28 June 2007; accepted 9 July 2007 Available online 24 July 2007 Daily-life decisions and judgments are often made “intuitively”, i.e., without an explicit explanation or verbal justification. We conceive of intuition as the capacity for an effortless evaluation of complex situations on the basis of information being activated, but at the moment of decision not being consciously retrieved. Little is known about which neural processes mediate intuitive judgments and whether these are distinct from those neural processes underlying explicit judgments. Employing functional magnetic resonance imaging (fMRI) we show that intuitive compared to explicit judgments in a semantic coherence judgment task are associated with increased neural activity in heteromodal association areas in bilateral inferior parietal and right superior temporal cortex. These results indicate that intuitive coherence judgments activate neural systems that are involved in the integration of remote associates into a coherent representation and, thus, support the assumption that intuitive judgments are based on an activation of widespread semantic networks sparing a conscious representation. © 2007 Elsevier Inc. All rights reserved. Introduction Intuition can be defined as the ability to apprehend an idea or to make a judgment about stimulus properties without being able to refer explicitly to the knowledge or inferences underlying the judgment (Bowers et al., 1990; Bolte et al., 2003; Bolte and Goschke, 2005; Volz and von Cramon, 2006). For instance, when viewing a stranger’ s face we sometimes have the immediate impression that the person is trustworthy without being able to describe the facial features leading to this impression (Winston et al., 2002). Our decisions may be guided by intuitive emotional evaluations which are not based on analytic reasoning or explicit weighting of alternative options, but rather experienced as intuitive “gut feelings” (Bechara et al., 1997). While the concept of intuition has often been met with skepticism or considered as not amenable to scientific investigation, an accumulating body of empirical evidence indicates that our judgments and our decisions are often biased by preexisting knowledge that is activated during the judgment process, but at the moment of decision not consciously available (e.g., Dorfman et al., 1996; Bowden and Beeman, 1998; Cleeremans et al., 1998). These studies demystify the concept of intuition (Bolte and Goschke, 2005). One task that has been proven to be particularly useful for experimental investigations of intuitive judgments in the domain of semantic processing is a coherence judgment task originally introduced by Bowers et al. (1990) and Bolte et al. (2003). Using a modified version of the Remote Associates Test (Mednick and Mednick, 1967), these authors presented participants pairs of word triads, one of which was coherent in the sense that all three clue words had a common associate, the “solution word” (e.g., the triad “playing, credit, report” is weakly associated with the solution word “card”). The second triad of a given pair was incoherent, that is, the clue words had no common associate. The main finding was that participants were able to decide reliably above chance level which of the two triads was coherent even when they did not come up with the solution word, but made their judgments on an intuitive basis. This finding was replicated by Bolte et al. (2003), who presented only one word triad at a time and used a signal-detection www.elsevier.com/locate/ynimg NeuroImage 38 (2007) 228 – 238 ⁎ Corresponding author. Department of Neurology, Klinikum Rechts der Isar, Technische Universität München, Ismaninger Str. 22, 81675 München, Germany. Fax: +49 89 4140 4867. E-mail address: ilg@neuro.med.tum.de (R. Ilg). Available online on ScienceDirect (www.sciencedirect.com). 1053-8119/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2007.07.014