Research Report Color Opponency in Synaesthetic Experiences Danko Nikolic ´, 1,2 Philipp Lichti, 1 and Wolf Singer 1,2 1 Department of Neurophysiology, Max-Planck Institute for Brain Research, Frankfurt am Main, Germany, and 2 Frankfurt Institute for Advanced Studies, Johann Wolfgang Goethe University, Frankfurt am Main, Germany ABSTRACT—Grapheme-color synaesthesia is a rare condi- tion in which perception of a letter or a digit is associated with concurrent perception of a color. Synaesthetes report that these color experiences are vivid and realistic. We used a Stroop task to show that synaesthetically induced color, like real color, is processed in color-opponent channels (red-green or blue-yellow). Synaesthetic color produced maximal interference with the perception and naming of the real color of a grapheme if the real color was opponent to the synaesthetic color. Interference was re- duced considerably if the synaesthetic and real colors en- gaged different color channels (e.g., synaesthetic blue and real red). No dependence on color opponency was found for semantic conflicts between shape and color (e.g., a blue lemon). Thus, the neural representation of synaesthetic colors closely resembles that of real colors. This suggests involvement of early stages of visual processing in color synaesthesia and explains the vivid and realistic nature of synaesthetic experiences. In a classic Stroop (1935) task, the subject is instructed to name the ink color of a word that refers to a color. Reaction times are longer if the meaning of that color word differs from the ink color (e.g., if ‘‘red’’ is written in blue ink) than if the meaning and the ink color are the same. The Stroop task can also be applied to study grapheme-color synaesthesia, in which perception of a letter or a digit is associated with concurrent perception of a color; synaesthetes report that these perceptions are vivid and realistic. In the case of synaesthesia, the time needed to name the ink color of a grapheme is shorter if the synaesthetic color of that grapheme is the same as the ink color (congruent condition) than if the two colors are different (incongruent condition; Dixon, Smilek, Cudahy, & Merikle, 2000; Mattingley, Rich, Yelland, & Bradshaw, 2001; Odgaard, Flowers, & Bradman, 1999; Palmeri, Blake, Marois, Flanery, & Whetsell, 2002). These findings indicate interactions (interference, facilitation, or both) between the real and the synaesthetic colors, provide empirical evidence that these synaesthetic experiences are real, and suggest that synaesthesia occurs without voluntary control (Cytowic, 1989; Hubbard, Arman, Ramachandran, & Boynton, 2005; Palmeri et al., 2002). However, it is unknown at which level of processing the interactions between real and synaes- thetic colors occur. Studies based on functional magnetic resonance imaging provide evidence for activation of the color-specific area V4/V8 of the visual cortex during synaesthetic experiences (Hubbard et al., 2005; Hubbard & Ramachandran, 2005; Nunn et al., 2002; Sperling, Prvulovic, Linden, Singer, & Stirn, 2006). Neurons of these areas are responsible for processing of visual stimuli and have color-opponent receptive fields: Cells excited by red are inhibited by green (and vice versa), and cells excited by blue are inhibited by yellow (and vice versa; Chichilnisky & Wandell, 1999; Hubel & Livingstone, 1987; Hurvich & Jame- son, 1955, 1957; Zeki, 1980). In the present study, we exploited this color opponency of hue-selective cells to identify the site of interactions between real and synaesthetic colors. We modified the Stroop task by creating two types of incongruent conditions. In one, the real color was opponent to the synaesthetic color (i.e., red vs. green or blue vs. yellow). We hypothesized that in this case, the two colors would likely involve the same color channels (Zeki, 1977). In the other incongruent condition, the real color was not opponent to the synaesthetic color (e.g., red vs. blue), and we hypothesized that the two colors would be processed by different channels (i.e., independently). We also hypothesized that if synaesthesia-induced colors are processed by opponent channels, there will be longer reaction times in the color-naming task when synaesthetic and real colors are opponent than when they are independent. Address correspondence to Danko Nikolic ´, Department of Neuro- physiology, Max-Planck Institute for Brain Research, Deutschor- denstrause 46, 60528 Frankfurt am Main, Germany, e-mail: danko@ mpih-frankfurt.mpg.de. PSYCHOLOGICAL SCIENCE Volume 18—Number 6 481 Copyright r 2007 Association for Psychological Science