It turns out that what we see is not always what we think we see. The visual system is highly modular, and different modules are used to perform different tasks.
I’m reading a book called Sight Unseen: An Exploration of Conscious and Unconscious Vision, by Goodale and Milner. It provides a fascinating look at the visual system. In a nutshell, the authors found that we process visual information differently depending on how we will use that information. Hypotheses of separate visual systems have been proposed for years, but they particularly struck the researchers when they met Dee (a pseudonym). Dee was unable to distinguish shapes or orientations, but she could see colors and textures. During one experiment, where Goodale and Milner were testing how well she could see a pencil’s shape and orientation, Dee became frustrated. She reached out and took the pencil to examine it more closely.
Think about this a minute. Dee’s visual disability was so severe that her world was a mere blur of colors and textures; she could not even tell whether a line was horizontal or vertical. But somehow, without consciously knowing what the researcher held in his hand or how far away it was, she reached for it without hesitation, correctly angled her hand and opened it just wide enough for the object, and took it straight out of his fingers.
Why could Dee do this? It turns out that the brain has two major, independent “streams” along which visual information flows. The type of vision we’re most aware of is our representational visual system (the “ventral stream”), which allows us to examine and describe the world around us. However, we also have a separate visuomotor system (the “dorsal stream”) that allows us to do things like judge the distance, size, and orientation of objects when we pick them up. These calculations are performed almost instantaneously, and our conscious mind is not privy to them. Patients like Dee, who have damage to the ventral stream, cannot describe the size or shape of a block on the table in front of them, but they can pick it up smoothly, without hesitating or groping. In extreme instances, patients are able to accurately look or point toward lights they insist they cannot see! Conversely, patients with damage to the dorsal stream can describe the block or the light perfectly well but have trouble with manipulation: They fumble for the block or point at the light with as much consistency as a random guess.
| Figure: Locations of the dorsal stream (representational vision; green) and ventral stream (visuomotor system; purple) in the brain. Both streams originate in the primary visual cortex (blue). |
Disconnected visual circuits are found in other vertebrates as well. In the 1970s, neurobiologist David Ingle severed the main optic nerve of frogs and allowed it to regrow to the opposite brain hemisphere (so the right optic nerve was connected to the right hemisphere of the brain, as opposed to the left hemisphere as is normal). When presented with the stimulus of a prey, the frogs would reach out and snap at it—but in the opposite direction (so a prey presented on the right would elicit snapping on the left), as expected. However, the navigational abilities of the frogs were unimpaired. When presented with a barrier that they had to leap over, the frogs consistently leapt in the correct direction, regardless of whether the barrier appeared in front of the left eye or the rewired right eye. Ingle also performed the reverse experiment, where the nerves involved in navigation were severed and rewired, and those frogs crashed into barriers but caught flies accurately.
Even within the dorsal and the ventral streams, optic neurons are assigned in a highly modular fashion. For example, in our eyes, we not only have rod and cone cells that activate in the presence of light or color, respectively – we also have cells that activate for very specific stimuli, such as a horizontal line or a human face. This modularity is also present in the visuomotor system, where activation of the neurons is tied not only to the visual stimulus, but also to the person’s action. For example, some neurons activate only when a person reaches out and holds a vertical block edge-on. Reaching out and holding balls, or even blocks that are lying down, do not cause activation.
In evolutionary terms, a modular system makes sense. Organisms like euglenas have basic visual systems that allow them to navigate towards light, for example. When organisms such as frogs began needing vision to catch prey accurately, their brains evolved new pathways linking some of their optic neurons to motor neurons. This allowed the visual system to expand and evolve without losing older functions like light sensitivity. And even though both pathways use the same information, they process it in parallel, independent fashions.
What does this mean for us as humans? What we “see” of the world is only a tiny part of how we interact with our visual surroundings. Rather than having a single holistic visual system that receives information, interprets it, and uses it to paint landscapes, move around barriers, and grab objects, we instead have a series of separate systems that each interpret the input from our eyes in different ways. We use visual cues to accurately reach for and hold a petri dish versus a graduated cylinder, but we have no conscious awareness of how we use that information to guide our motions. This also makes the fact that we have such integrated behavior rather astounding. Think about how many different kinds of vision are involved in painting that landscape: representational vision to observe the landscape and match it to the painting; visuomotor control to select the correct brush size and paint tube and put the paintbrush to the canvas with just the right pressure. And within those, we have neurons that activate specifically for red, or blue, or a particular texture, while others activate only when we hold the brush and not when we hold the palette.
Once again, I’m convinced that evolution is truly crazy and truly amazing.
— Sonia Singhal
