Gaining Perspective on How We See in 3-D
Interview with Bruce Cumming, M.D., Ph.D.
By Allyson T. Collins, M.S.
NEI Science Writer/Editor
Bruce Cumming, M.D., Ph.D.
Chief, Laboratory of Sensorimotor Research
National Eye Institute
The recent blockbuster movie "Avatar" may be best experienced with 3-D glasses, but in reality, our own two eyes help us see everything in three dimensions. Bruce Cumming, M.D., Ph.D., came to the NEI Laboratory of Sensorimotor Research in 2000 to study that innate process.
In a recent interview in an office stocked with a thousand pairs of 3-D paper glasses (bought in bulk for research purposes) and five computer screens, Cumming talked about using a combination of laboratory experiments and high-level math to identify the features of the brain that help us see depth in everything around us.
Why did you choose a career in science?
I had always been interested in how the body works. Quite early on in my life, I had decided that I wanted to go to medical school. When I was a medical student, I remember reading papers about how our visual system adapts if we put on goggles that magnify objects. It was just amazing to me--both that our bodies are able to do this, and at how far researchers had come in understanding the process in the brain. At that point, researchers could almost write equations describing what the brain is doing to adjust to the goggles. I thought, 'This is the future.' From that moment on, I wanted to do research on how the brain works.
Why did you decide to stick with basic research instead of clinical medicine?
I believe that basic research gives us everything we have--medical knowledge and treatments. I thought that the biggest contribution I could make to humanity is to discover something through basic research that could be applied to clinical medicine. In science, we need everyone, but I think that if you want to shoot high, basic research is where you can really make the biggest impact on human health in the long run.
How does your research, which involves the brain, relate to vision?
Seeing doesn't happen in the eye. The eye, from my perspective, is a necessary first step in vision because it provides electrical signals that the brain uses to see. But if you were to take away the eyeball and replace it with an appropriate piece of electronics, we would still see just fine. The process of seeing is very complicated. We don't just look at a picture with the retina and see; all seeing is done in the brain.
So what aspects of vision are you studying in the brain?
I work on three-dimensional vision because it's a good model for understanding part of how the brain allows us to see. For 3-D vision, the brain processes different visual information from each of our eyes to give us a sense of depth. My ultimate objective is to write down an equation to explain how these 3-D visual signals are generated. I'm also trying to figure out the relationship between these signals and our internal sensation of depth to prove that signals in a particular part of our brain actually make us feel that one object is in front of another.
Is it really possible that how we see could be described with a math equation?
The math behind this is called quantitative modeling. About 20 years ago, a relatively simple equation was developed to explain a lot about how we see, in terms of the connections between brain cells. You could imagine that any equation would be a bit simple compared to the process of seeing, but we use this equation as a springboard for testing different hypotheses about vision. Each time we devise an experiment based on the equation, we realize that either the brain can do something more sophisticated than what the equation says, or the equation doesn't explain in enough detail how we see. Through this type of modeling, we have learned that cells signal depth in the early part of visual processing, in the visual cortex area of the brain. These brain cells are more clever than we thought, but they don't do enough to completely explain depth perception. So the equation has captured something very important, but we have to modify it and test another hypothesis because there is still information missing. You might think that we're going around in circles, but if we show that part of the equation is either wrong or right, we're getting a step closer to understanding vision.
What's your ultimate career goal?
I'm looking at 3-D vision as a model system to answer a major question: how is it that the firing of millions of neurons in the brain can help us perceive visual information? I could get a computer to give me information about depth and tell me whether one object is in front of or behind another. In fact, much of the background information we have about 3-D vision comes from studies of computer vision and robots. But I wouldn't say for a minute that the computer is seeing anything. I'm trying to understand what it is to actually see something--to have a conscious sensation and perceive a scene--rather than just have an image on the retina.