The non-opponent nature of colour afterimages.

Complementary colour afterimages have driven our understanding of human colour perception since the foundations of modern colour science. Despite their fundamental importance, decades of research have failed to establish the precise nature of colour afterimages and the neural mechanisms of adaptation at their origin. To date, it is unclear whether afterimage formation is caused by adaptation in the cone photoreceptors, of colour-opponent neurons in the subcortical pathway, or requires the assumption of yet unknown cortical mechanisms. To establish the neural mechanisms underlying afterimage formation, this study exploited the fact that different candidate mechanisms make fundamentally different predictions about the hue and saturation of afterimages. Using tailormade experimental paradigms, the exact colours perceived in afterimages were measured for a large range of inducers to test those predictions. Three experiments tested predictions of afterimage hue and saturation with varying inducer hues, and changes of afterimage hues depending on inducer saturation (Exp. 1.a: 8 colours, tested across 31 participants; Exp. 1.b: 24 colours, tested across 52 participants; Exp.2.a: 72 colours, replicated with 10 participants; Exp. 2b: 72 colours, replicated with 2 participants; Exp. 3: 48-216 colours, replicated with 5 participants). Results across all three experiments very consistently demonstrated that afterimage colours are not colour-opponent, as widely assumed, but closely follow a quantitative model of adaptation in the cone photoreceptors. These findings unequivocally establish the origin of afterimages along the hierarchy of neural processing, hence resolving all prevailing misconceptions and contradictions. By linking the perceptual nature and the neural origin of afterimages, the present paradigm also provides a technique for probing those neural mechanisms behaviourally and in first-person experience.