Purchase this article with an account.
Ariana Familiar, Edmund Chong, Won Mok Shim; Interpolated visual features during apparent motion are represented in primary visual cortex. Journal of Vision 2014;14(10):286. doi: https://doi.org/10.1167/14.10.286.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
While we perceive apparent motion (AM) between two static objects, our visual system interpolates an intermediate object, which is not present in the bottom-up input, on the illusory path of AM (Kolers & von Grunau, 1976; Hidaka et al., 2011). Recent neuroimaging studies using pattern classification methods demonstrate that the intermediate representations on the AM path can be decoded in regions of V1 retinotopically mapped to the AM path, suggesting that such interpolated features are represented at the earliest stages of cortical processing (Chong et al., 2011 VSS). However, it remains unclear whether the same population of neurons which encode this intermediate feature when it is actually presented are recruited during AM. Using fMRI and a forward encoding model, we examine the population-level orientation tuning in the region of V1 corresponding to the AM path while subjects view rotational AM induced by successive presentation of a right- (45o) and left-tilted (135o) grating at the upper and lower corners in the right visual field respectively. In the region of V1 corresponding to the AM path, we find clear orientation tuning functions that peak at the supposed intermediate orientation (0o), and gradually decrease as the distance from this orientation increases. Furthermore, this tuning is similar to the orientation tuning profile evoked by actual presentation of the intermediate grating on the AM path. Such orientation tuning is not found when subjects visually imagine the same stimuli, nor when the percept of AM is abolished by presenting the two gratings simultaneously. Our results suggest that non-stimulated region of early visual cortex (V1) represents features of a dynamic stimulus during AM even when those features are not present elsewhere in the bottom-up input, and therefore such features must be interpolated through top-down processing.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only