Storage of the Motion After-Effect

Omnibrain had a story on the motion after-effect a while back, including a link to a rather nice demonstration. To get an idea of the effect, just look at the centre of the moving pattern for 20 seconds or so. When you look away, things should look fairly weird for a while.

So why does this happen? The most popular way to explain this effect is to use a so-called ratio model (e.g., Barlow & Hill, 1963). The model itself is simple, but it does rely on a bit of prior knowledge of how neurons work. Skip two paragraphs ahead if you don’t need to be reminded of the basics.

Imagine that for each bit of the retina, there is a neuron that fires when there is movement in one direction, and a complementary neuron that fires in response to movement in the opposite direction. Most likely, there’s a bunch of these pairs, each responding maximally to a particular direction of motion. By the way, these neurons are probably nowhere near your retina, unless you’re a rabbit – Barlow and Hill thought they could be found in the lateral geniculate nucleus, though most researchers today would probably prefer to place them in the primary visual cortex, or in area MT.

When you view the trippy moving stimulus, each receptive field only gets stimulation in a single direction (since the receptive field is small). So in a given spot, only a couple of the neurons have been firing like mad – namely, the ones that respond to that direction of motion. The rest have been firing at their base rate (neurons generally fire spontaneously at a low rate in the absence of any stimulation). As you get tired of the irritating techno track and decide to look away, the stimulated neurons go into a refractory period, that is, their sensitivity to stimulation drops for a while, which reduces their firing rate.

Knowing this, the basic idea behind the ratio model is simple: if your brain decides which way the world is moving by comparing the ratio of firing between these direction-sensitive neurons that are tuned in opposite directions, the ratio will shift in favour of the opposite direction as you look away, because while the neurons that responded to the opposite direction are firing at base rate as they have done all along, the neurons that did respond to the direction of movement are now firing at less than base rate.

And indeed, Barlow and Hill (1963) showed that in the retina of the rabbit (where the ganglion cells are direction-sensitive, unlike human ganglion cells), something like this seemed to be happening.

The ratio model is such a tidy explanation that most investigators paid no mind to a few papers published by Spigel (1960, 1962) in an obscure journal. The ratio model predicts that it doesn’t really matter what you do after adapting to the moving stimulus. As long as you don’t view the stimulus again, the refractory period only lasts for a few seconds – the underlying biochemical process is well understood, and really doesn’t allow for any exceptions.

While Spigel made no explicit reference to the ratio model, his results violate this prediction. In a nutshell, Spigel had his participants adapt to a rotating stimulus. Then, he stopped the rotation and had the participants view a static version of the stimulus (note that this is slightly different from merely looking away from the stimulus). The participants reported when they no longer experienced the after-effect. Next, the participants adapted to the rotating stimulus again, but this time following adaptation, they were placed in complete darkness for the same length of time that they previously experienced the after-effect. After this, they once again viewed the static version of stimulus.

The design of this experiment may seem a bit contrived, so maybe it’s a good idea to test it on yourself to see how it works. Open the moving stimulus and this image of the stimulus in separate tabs or windows. Make sure you have a stopwatch and a quick way of switching from the moving stimulus to the static stimulus (for example, alt-tab for windows, or command/control-[tab number] for tabs). If your browser automatically resizes oversized images, make sure you view this image in full size. View the moving stimulus for 30 ompa-ompas, switch to the image of the stimulus, and start your stopwatch. Stop it when the image is no longer moving, and record the duration of the effect. Next, view the moving stimulus for another 30 ompa-ompas. After this, look away from the screen and start your stopwatch. When an equal amount of time to the original effect duration has passed, look back at the image of the stimulus (time the effect again if you want to be scientific). Lo and behold, the effect is (hopefully!) still there. If you want to be properly scietific, you may also wish to test yourself in the first condition one more time to check that the after-effect still lasts for roughly the same length of time, thus ruling out any possible practice effects.

In the original study, Spigel (1960) found that following the time spent in darkness, the participants still reported an after-effect. The duration of the effect was significantly shorter than in the original condition, but it was nevertheless there. Spigel (1962) found that the storage effect also appeared if the participants simply viewed a blank white surface between adaptation and testing, and later investigators found that essentially, the storage effect appears pretty much regardless of what the participants view between adaptation and testing (Thompson & Wright, 1994) – this is pretty much what you just did.

It’s very tricky to explain Spigel’s (1960, 1962) findings with a ratio model. There’s no reason why this purported imbalance in firing rates between opposing neurons should essentially disappear during the pause between adaptation and testing, only to re-appear when you view the static stimulus again. While no one seems to have tested how long the after-effect can be stored at most, Masland (1969) found that 15 minutes of adaptation produced an after-effect that was visible when viewing the static stimulus after around 24 hours. There is no way that the adaptation of single neurons can explain an effect that lasts so long.

Finally, a caveat: I think there may be two different processes that produce motion after-effects. One mediates the immediate effect you get when you look away from the screen after viewing the trippy stimulus that Omnibrain found. This effect goes away after a few seconds, and while it does come back when looking at the static version of the stimulus, the effect is not that that strong. The other process mediates the after-effect you get when viewing a static version of the adapted stimulus. It’s worth noting that some stimuli (particularly rotating ones) don’t produce much of an immediate effect, but you do get a nice after-effect when you view the static version. This suggests to me that the ratio model may not be plain wrong – but it’s only part of the story.

Unfortunately, no one seems to know what the other part is (but see van de Grind et al, 2004 for a good but heavy-going attempt to make sense of it all).

References
Barlow, H.B., & Hill, R.M. (1963). Evidence for a Physiological Explanation of the Waterfall Phenomenon and Figural After-Effects. Nature, 200, 1345.

Masland, R.H. (1969). Visual Motion Perception: Experimental Modification. Science, 165, 819-821.

Spigel, I.M. (1960). The Effects of Differential Post-Exposure Illumination on the Decay of the Movement After-effect, Journal of Psychology, 50, 209-210.

Spigel, I.M. (1962). Contour Absence as a Critical Factor in the Inhibition of the Decay of a Movement Aftereffect. Journal of Psychology, 54, 221-228.

Thompson, P., Wright, J. (1994). The Role of Intervening Patterns in the Storage of the Movement Aftereffect. Perception, 23, 1233-1240.

Van de Grind, W.A., van der Smagt, M.J., & Verstraten, F.A.J. (2004). Storage for Free: A Surprising Property of a Simple Gain-Control Model of Motion Aftereffects. Vision Research, 44, 2269-2284.