Our visual world is composed of many different strengths of light. Compared to the average background intensity of light, there are more dark details than bright details in a visual scene. Evolution has carefully adapted the visual system of many animals, including humans, to this difference between dark and bright details. Human observers are more sensitive to dark details and respond more quickly when those details change. Consistent with this, cells in the early visual cortex have a higher level of activation to dark stimuli (than to bright), and the response latency fordark–sensitive cells (called OFF cells) is shorter than that of bright–sensitive cells (called ONcells). A second visual phenomena that has fascinated humans for thousands of years are afterimages. The natural philosopher and father of Biology, Aristotle observed around 330B.C.: “…the perception continues in the sensory organs, … not merely while they are actually engaged in perceiving, but even after they have ceased to do so”. He was describing the situation where after viewing a bright or dark object for some seconds, even after moving our eyes away from it a negative version of the object is still perceived (Fig. 1). Because the ON and OFF cells early in the visual system respond vigorously to the removal of opposite–sign stimuli (Fig. 2A), previous researchers have assumed that the post-stimulus response of ON and OFF cells may be the neuronal basis of afterimage perception (Fig. 2B). However, neuronal correlates of afterimages are observed in many other brain areas, thus the origin of the afterimage is still controversial. 

Figure 1: Examples of the negative afterimage illusion for the ION Neuroscience building (A) and the ION logo (B).Stare at the red cross on the left for about 30 seconds without moving your eyesand try not to blink; then stare at the red crossto the right:you should perceive the originalcolors,however the physical image is pure grey.Thedifferent strengths of colors you perceive only exist in your mind due to adaptation of different neural channels.

In a recent paper entitled “Asymmetries of Dark and Bright Negative Afterimages Are Paralleled by Subcortical ON and OFF Post‐Stimulus Responses” publishedin the Journal of Neurosciencein January 2017, researchers from Dr. WANG Wei’s lab at ION took advantage of the known differences between the perception of dark and bright stimuli to test whether such asymmetries also exist for bright and dark afterimages, and then probe the underlying neuronal mechanisms generating afterimage perception. The researchers performedmatched psychophysical experiments in human observers and neuronal single‐unit and S–potential recordings in a core animal model for early visual processing, the cat lateral geniculate nucleus of the thalamus (LGN, the first brain structure to receive input directly from the eye).

Figure 2: A, Cartoon ON cell responses to the onset of a bright stimulus and offset of a dark stimulus, and vice versa for OFF cells. B, Removal of bright stimulus after adaptation causes a dark afterimage, while removal of a dark stimulus after adaptation causes a bright afterimage. One plausible hypothesis for negative afterimages suggests an origin in subcortical responses to the opposite‐sign stimulus offset. C, Perceptual cancellation paradigm for measuring the perceived contrast of afterimages. A physical bright or dark disc was presented for 4 seconds, then variable strength nulling pedestal was presented. D, Afterimage？pedestal ratio for 8 subjects. All the subjects show stronger afterimage perception for dark stimuli (more red than blue shaded area of the curve). E, Bayesian estimation of the population tuning curvesfor theAfterimage？pedestal ratio. Posterior distribution plus 95% confidence interval shows these are highly significantly different. F, Averaged PSTH to the offset of a bright stimulus for OFF cells and the offset of a dark stimulus for ON cells. Note that average firing rate of ON cells is larger than OFF cells, and only in the LGN this difference was maintained for several seconds. G, Distributions of relative response strength of ON and OFF cells to the stimulus offset.

In human subjects, the researchers used a technique called perceptual nulling, where the illusory percept of the afterimage is cancelled by a physical dark or brightstimulus called a pedestal (Fig. 2C). Theydiscovered that asymmetries do indeed exist between bright and dark negative afterimages induced by dark and bright stimulus discs (Fig. 2D-E). Dark afterimages lookweaker and are perceived for less time than bright afterimages. The researchers then sought to determine the neural substrates, and performed neuronal recordings at the equivalent time period to afterimages in humans (post-stimulus responses). They found ON and OFF cells revealedclosely matched asymmetries with respect to response strength and duration (Fig. 2F-G). Comparing the responses from the eye (the retinal S-potentials) to those in the LGN (relay cell single units), they found only LGN responses matched the duration statistics observed in human observers. Aristotle speculated that “the previous perception remains when we have turned our gaze from sunlight to darkness … owing to the motion excited by the light still subsisting in our eyes”. This work helps to pin down this post-stimulus mechanism that Aristotle hypothesized so long ago.It not only suggests that subcortical differences between ON and OFF channels help to explain the perceptual difference humans observe betweenbright and dark afterimages, it further supports the general notion of a subcortical neural correlatefor the perception of negative afterimages.

This work was mainly conducted by graduate student LI Hui under the supervision of Dr.WANG Wei at the Institute of Neuroscience (ION), Chinese Academy of Sciences. This work was supported by National Natural Science Foundation of China grants (31571078 and 81601628), 973 Project and the State Key Laboratory of Neurosciences (SKLN).