Abstract
It is still debated whether imagery involves a spatially specific form of mental representation, whether early visual areas are involved, and if so, what information is represented in early visual areas during imagery (Kosslyn, 2000; Pylyshyn, 2003). To address these issues, we used fMRI to investigate the spatial specificity and phenomenal content of imagery representations in retinotopically defined visual areas. Specifically, we manipulated the spatial location, contrast, and spatial frequency of both physical and imagined stimuli to evaluate if early visual areas were similarly modulated by these manipulations. Subjects either viewed or imagined seeing a flickering checkerboard 6 to the left or right of central fixation (size 4, temporal frequency 8 Hz) at one of three contrast levels (2, 10, 50%) or one of three spatial frequencies (check size 1, 3, or 6 ). Activations plotted on flatmaps revealed that the exact same regions of V1 and V2 were activated by imagery and visual stimulation, corresponding to the discrete peripheral location of the checkerboard lying along the horizontal meridian. The spatial specificity of these effects cannot be explained by confounding factors such as eye movements or blinks, and instead, strongly suggests that imagery activates spatially specific representations in visual cortex. Imagery of different contrasts levels led to responses of equal magnitude in early visual areas, whereas greater physical contrasts evoked greater fMRI responses. Interestingly, imagery of higher spatial frequencies led to stronger responses in V1, analogous to the effects of physical spatial frequency. Our results suggest that early visual areas such as V1 reflect both the spatial specificity as well as some aspects of the spatial content of visual imagery. These results further support the notion that early visual areas may play an important role in representing phenomenal experiences, even in the absence of relevant input.
NIH grants R01-EY14202 and P50-MH62196.