Anti-correlated stereograms

From A conversation about the brain
Jump to: navigation, search
In an anti-correlated random dot stereogram, (c), all the white dots in one eye match up with black dots in the other eye. People cannot perceive depth in this type of stereogram but neurons in V1 still respond differentially to different disparities added to the dots. Figure is from Parker (2007)[1].
Most people can see depth in random dot stereograms but if the contrast in one eye is reversed so that all the black dots become white and vice versa, they can't. The pattern just looks like a shimmery mess. To neurophysiologists, this type of pattern is interesting because neurons that are responsive to the disparity in an image continue to respond to the disparity in an anti-correlated pattern, despite the fact that the observer's perception is so different in the two cases[2]. This lack of correspondence between neural response and perception might seem perplexing at first. In terms of the idea explored in these pages, however, the fact that neurons might contribute to perception in one case (the correlated stereogram) but not in another (anti-correlated), is understandable. As in previous examples (a single neuron and vernier acuity), the extent to which a particular stimulus parameter or a particular neuron influences a perception/action/interpretation depends on the context. In the case of a stereogram, the rest of the stimulus seems to act as a 'gate' that allows perceptual access to the firing of a particular disparity-tuned neuron in one instance (the correlated stereogram) but not the other (anti-correlated). If the current sensory+motivational state travels along a familiar path, e.g. from coarse to medium and then fine scale match (as discussed earlier in relation to recognising a face or detecting a predator), then the current sensory+motivational state can land up in a Voronoi cell where a small number of neurons can determine the perceived depth of a part of the stimulus. The same is not true of an anti-correlated pattern. Although disparity-tuned neurons are stimulated, there is no orderly binocular structure that can be described at multiple scales (i.e. the hierarchical scale structure is quite different for the left and right eyes, so there is no way to compare the structures sensibly between the left and right eyes). Here is an illustration of the idea of different 'addresses' in the left and right eyes. A similar point was raised in relation to the behaviour of an animal being determined by the firing rate of a single neuron in some cases but not others.

Back to Neurophysiology.


  1. Parker, A. J. (2007). Binocular depth perception and the cerebral cortex. Nature Reviews Neuroscience, 8(5), 379-391.
  2. Cumming, B. G., & Parker, A. J. (1997). Responses of primary visual cortical neurons to binocular disparity without depth perception. Nature, 389(6648), 280-283