Visual illusions, the tricks you can play on your own sight, often yield insights into how our visual system operates. One of the most compelling visual illusions is the so-called “waterfall illusion.” Gaze at a waterfall for a prolonged period of time and then glance at the adjacent static rocks: the rocks appear to be moving upwards. (If you don't have a waterfall handy, a web search will yield many online versions of the illusion.)
While it has been proposed that this visual after-effect reflects fatigue or adaptation of certain areas of the brain, the neuronal basis of the illusion remains unclear.
Previous studies suggested that a specific group of neurons in the visual cortex may be more active when the effect is perceived than when it has faded. In today's issue of Neuron, Huk and colleagues challenge this view and report some interesting findings that may lead researchers to reconsider that theory.
First, the authors show that the neuronal changes traditionally associated with the illusion are not a direct consequence of the perception of the waterfall. Rather, they may depend on giving that perception one's full attention: they cannot be detected when subjects are “distracted” by a simultaneous task of similar difficulty while perceiving the illusion.
The researchers then show that the illusion is caused by the adaptation of some neurons following repeated stimulation. As we watch the water fall downward, the neurons that are selective for downward direction adapt and become less responsive. When we then look at the static rock, the upward-selective neurons will be relatively more active, leading us to perceive that the rocks are moving upward.
Taken together, these results are exciting because they shed light on the neuronal basis of the waterfall illusion and more generally on the visual perception of motion.
In addition, as Huk himself points out, they show that perception does not always correspond to an increase in activity in the relevant brain areas, but may result from relative differences between the responses of different neurons within these areas. The accompanying Preview by Rees discusses the implications of this work.
(Reference: Volume 32 Number 1 October 11, 2001 Neuron)
[Contact: Alexander C. Huk]
11-Oct-2001