What Rapid Eye Movement Reveals

Michele Rucci wants to look deep into your eyes. And when he does, he sees something amazing in those windows to your soul: they’re jittering around like popcorn kernels in a hot air popper. A disconcerting image? Perhaps. But Rucci, a Boston University professor of psychological and brain sciences and director of BU’s Active Perception Laboratory, has found that these coordinated microscopic jitters allow us to see fine spatial detail like letters on a page. His research, published in December 2015 in Current Biology, reveals that problems with visual acuity may actually stem from inaccurate motor control of your eyes, head, and neck—not damage to the eye. His research may eventually lead to a clearer understanding of vision disorders.

BU Research spoke to Rucci about his vision of vision and why movements so tiny may have impacts so large.

BU Research: What questions does your lab want to answer about vision? And why are these things not known yet?

Rucci: The main question is: How is it that some light gets in our retina, activates retinal receptors, and gets transformed into what we see? We recognize a face; we recognize objects. But all we have to start are neural responses signaling light characteristics at individual points in the scene. How do we put them together? It’s the basic mechanics of vision, but it’s a really, really hard question.

Well, photoreceptors in the retina fire neurons and the signals go to the brain, right?

Right—we are such visual creatures that we don’t even realize that there is a question there. It seems so effortless. You open your eyes and you see. So you wonder what is the question? But how does our brain transform individual pixels into an image of, say, a dog? How do you know that these points go together into the same object and these others belong to another object? How are we able to interpret this?

That’s a big question. How did people first start thinking that head movements, eye movements, and micro-movements of the eye itself fit into the picture?

First the question was: Why is it so hard to replicate human capabilities in machines? What are we missing? And it’s been really a struggle. One aspect that is clearly missing is the behavior of the observer: our eyes are constantly active, always moving. And that movement actually helps us extract the information. I’ll give you an example. Let’s just consider the scene in front of me. If I look at one point and I move a little bit, that gives me information of three-dimensionality, what we call parallax. So already behavior is introducing new cues, new pieces of information. We have this idea that we don’t need motion to see. But if in the laboratory we remove every kind of motion, a procedure known as retinal stabilization, the image will literally fade away.

That’s weird.

We’re not aware of it, but that’s how it is. We reconstruct a scene by moving our eyes around.

We unconsciously focus on different points to construct an image?

Exactly. Your eyes jump from one point to the next. You make these very rapid eye movements, called “saccades,” which separate brief periods of fixation in which you grasp information. But the term “fixation” is inaccurate: small eye movements incessantly occur, even during these periods.

What does “saccade” mean?

I think it comes from the way horses move their heads. It’s the French word for “jerk.”

And how rapid are they?

They’re very, very fast. We stop at each point for like 200 to 300 milliseconds, and then we very rapidly jump to another point. So, three times a second, we move our eyes. It’s jumping, jumping, jumping, jumping. And the eyes continually jitter even in the periods in between saccades. We can’t stop our eyes: incessant eye movements occur even when we attempt to maintain steady gaze on a single point. We call the movements microscopic because they’re very difficult to record, but they’re actually very large compared to the size of individual receptors on the retina.

Is it just your eyeball moving, or your whole head, or both?

Well, it’s the eyes, but as you’re trying to keep your head fixed, you’re actually also making very small head movements. So the two things combine together to create a lot of retinal motion.

It would seem like it would ruin your vision, all this movement.

Exactly! So the first question is: Why don’t you see the world blurred? Right? It actually gets more interesting.

So what do the micro-movements do, the little jitters?

The standard idea is that they help refresh the image by preventing neural adaptation. To me, it always seemed a sort of superficial level of explanation. The interesting question is: how do they actually help you see? And they’re revealing a lot. In 2007, we discovered that this jitter contributes to vision of fine spatial detail, and in 2010, that some of these movements known as microsaccades are very precisely controlled. It’s amazing, the level of control that we have on these eye movements. They are reshaping the stimulus on the retina in a very interesting way.

So that was a surprising discovery, that we had that level of control over it?

Yes. So microsaccades are controlled, and now in this latest paper, we are finding out that drift is also controlled—that’s the jittery motion that is there all the time.

Do the head and eyes move the same way, or move differently but in conjunction somehow?

What we have found in this last study is that the head and the eyes move together even at this microscopic scale. So you’re not aware of making very small head movements but the eye will partially but not completely compensate for the movement of the head. It’s as if the amount of motion on the retina is designed to be at a fixed level.

And by coordinating everything like that, that allows you to see a coherent image?

Well, it allows you to start extracting information. One important point is: How do you know where edges of objects are? How do you start extracting this information? The processing of edges is believed to occur up in the cortex. It’s a complicated operation. It requires a few steps. But everything we know is based on assumptions that are very static, which are implicit to the way vision is traditionally studied in the laboratory. What is starting to emerge is that these small eye movements reformat the input to the retina; they transform it into something that is spatial-temporal. This transformation turns out to be tuned to the characteristics of the natural world and starts a very important function, which is the function of extracting useful information—including edges.

Are there specific visual disorders or diseases that mess this up?

Usually these very small movements are not considered because they are so difficult to record. They are known to exist, they are known to be important, but in most vision science they are thrown under the carpet because how do we deal with them? In fact, they’re even regarded as a nuisance because you can’t control them. So most of what we know about vision comes from experiments in which we take an image, we flash it; most neurophysiological data come from experiments with paralyzed animals. So we know practically nothing about how these small movements vary in different pathological conditions. It’s a completely unknown world out there. What this research is leading to is the possibility that visual acuity is not just a visual phenomenon; it’s really a visual-motor achievement. If you don’t make microsaccades that are precise, if you don’t control your eye jitter (drift), you’re not going to be able to thread a needle well. So you need to have that level of control. Now, if you cannot thread the needle, is that because your visual acuity went bad or because you’re now controlling your eye movements less accurately? The second part is not even considered, so far.

Is that what you’re thinking about for your next line of research?

Yes, exactly. That’s how we are evolving.