The Mechanisms Underlying Colour Afterimages
Witzel, C.
Show abstract
Negative colour afterimages incorporate the most fundamental tenets of human colour perception (adaptation, opponency, complementary colours, context-dependence, dependence on the beholder), but despite their fundamental importance, the underlying mechanisms are controversial. The key question is whether the afterimages are attributable to adaptation in the cone photoreceptors, of colour-opponent neurons in the subcortical pathway, or require the assumption of yet unknown cortical mechanisms. The most common assumption in textbooks and contemporary research is that negative afterimages are cone-opponent. It has not previously been recognised that the role of those mechanisms can be distinguished because they make fundamentally different predictions about the hue and saturation of afterimages. To test these predictions, we developed experimental paradigms to measure the exact colours perceived in afterimages. The results reveal that afterimages do not align with cone-opponency but closely follow a well-founded model of cone adaptation (cone contrasts). Our findings establish that cone-adaptation is the sole origin of negative colour afterimages. The quantitative cone-contrast model provides a comprehensive, straight-forward explanation of the adaptive mechanisms underlying colour afterimages that resolves apparent contradictions and debunks wide-spread misconceptions. This model has far-reaching implications for longstanding mysteries about visual perception.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Violated predictions enhance the representational fidelity of visual features in perception 96%
- Temporal dynamics of human color processing measured using a continuous tracking task 94%
- The effect of illumination cues on color constancy in simultaneous identification of illumination and reflectance changes 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.