The neural substrates of conscious color perception demonstrated using fMRI

Abstract

It is well established that seeing color activates the ventral occipital cortex, including the fusiform and lingual gyri, but less is known about whether the region directly relates to conscious color perception. We investigated the neural correlates of conscious color perception in the ventral occipital cortex. To vary conscious color perception with the stimuli-remaining constant, we took advantage of the McCollough effect, an illusory color effect that is contingent on the orientation of grating stimuli. Subjects were exposed to a specific combination of chromatic grating patterns for 10 min to induce the McCollough effect. We compared brain activities measured while the subjects viewed achromatic grating stimuli before (PRE) and after the induction of the McCollough effect (POST) using functional magnetic resonance imaging (fMRI). There were two groups: one group was informed that they would perceive illusory color during the session (INFORMED group), whereas the other group was not informed (UNINFORMED group). The successful induction of the McCollough effect was confirmed in all subjects after the fMRI experiment; nevertheless, only approximately half of the UNINFORMED subjects had been aware of the color during the POST session, while the other half had not. The left anterior portion of the color-selective area in the ventral occipital cortex, presumably V4α, was significantly active in subjects who had consciously perceived the color during MR scan. This study demonstrates the activity in a subregion of the color-selective area in the ventral occipital cortex directly related to conscious color perception.

Introduction

Subjective color experience is primarily elicited by the neuronal processing of the wavelength composition of the light that reaches the eyes. Electrophysiological studies in nonhuman primates have revealed that chromatic interaction was found at multiple levels from LGN (Kastner et al., 1992), V1 (Wachtler et al., 2003), V4 (Zeki, 1983), and inferior temporal cortex (Komatsu et al., 1992). Visual areas V1/V2 are primarily involved in the earlier stages of color processing, in which the presence and the intensity of different types of wavelengths are registered, whereas responses of cells in V1 are mildly modulated by the surrounding wavelength composition (Wachtler et al., 2003). In the higher order visual areas V4, however, cells are shown to respond to colored patches based on what they are perceived, regardless of the wavelength composition of the patches (Zeki et al., 1983). These results suggest that the activities in the higher order visual cortex reflect the perception more strongly than the early visual cortex.

Human neuroimaging studies have demonstrated that passive viewing of colored stimuli and simple color discrimination activate the fusiform and lingual gyri Corbetta et al., 1991, Gulyas and Roland, 1994, McKeefry and Zeki, 1997. Using multicolored abstract or natural scenes whose wavelength composition or illumination intensity changes continually, without altering the perceived color of the individual patches, Bartels and Zeki (2000) found two active zones in the fusiform gyrus, V4 proper and V4α. Because of the coactivation by the color constancy tasks, they grouped them as V4 complex, designating it as the color center of the human brain. Detailed retinotopic study using functional MRI and phase encoding technique by Wade et al. (2002) revealed that the color center includes the fourth visual field (hV4) that is anterior to the V3v: they observed that the color stimulation activated the V1, V2, hV4, and further anterior ventral occipital region.

Recent studies have shown that illusory color effects, which can elicit color perception in the absence of chromatic stimuli, also activate a portion of the color center Barnes et al., 1999, Hadjikhani et al., 1998, Sakai et al., 1995. This suggests that the activity in the color center may be necessary for color perception, which is consistent with lesion studies showing that achromatopsic human patients lose color perception following damage to fusiform and lingual gyri Damasio et al., 1980, Meadows, 1974, Zeki, 1990. Less is known, however, about how activity in the color center relates to conscious color perception.

To examine how activity in the color center relates to conscious color perception, we utilized an illusory color effect called the McCollough effect (McCollough, 1965), in which the intensity of color perception varies while the stimuli remain constant. To induce the McCollough effect, two orthogonally oriented grating patterns, such as a green-and-black horizontal grating and a magenta-and-black vertical grating, are viewed alternatively for a few minutes (Fig. 1c). After such induction, black-and-white test patterns are presented to subjects. White section of the horizontal grating appears to the viewer to be tinged pink, whereas that of the vertical grating appears green, complementary to the color of the induction stimuli. It is known that the McCollough effect is an automatically and inevitably induced phenomenon.

As the McCollough effect can induce color perception without chromatic stimuli, it is well suited for the investigation of the conscious color perception. Our hypothesis was that the induction of the McCollough effect and conscious illusory color perception are different processes and hence neural representations are different. If this hypothesis is true, there may be subjects who are not consciously aware of the illusory color whereas the McCollough effect has been induced. To prove this hypothesis, we examined subjects who were not provided a priori information of the McCollough effect (UNINFORMED group). Neural activities were measured before and after the induction of the McCollough effect using functional magnetic resonance imaging (fMRI). Half of the subjects did not perceive illusory color during fMRI session after the induction (UNAWARE group) and others did (AWARE group), whereas psychophysical test after fMRI confirmed the induction of the McCollough effect in all subjects. Then we recruited another group who was informed that they would perceive illusory color during the session and was required to pay attention to the color (INFORMED group). All subjects in the INFORMED group perceived illusory color during the fMRI session. Common activation across all groups was in the posterior portion of the color center on the left, presumably V4, which may represent the induction of the McCollough effect. On the other hand, common activation between AWARE and INFORMED groups, but not in the UNAWARE group, was found in the anterior portion of the color center, presumably V4α, on the left. Hence, the left V4α may represent the perception of the illusory color.