Haldan Keffer Hartline, a Prize, and Two Accidents - Limulus, Lateral Inhibition, and the Mach Band Effect

You might think that once we measure voltages off of one nerve fiber, we can add up the rest of them and get the signal sent to the brain: not so, and Hartline would demonstrate this. Hartline often joked that he worked on the visual system of a blind animal (since Limulus is almost totally blind) and that he studied light sensitivity in the dark. Vision research often requires a constant level of illumination, and since light casts shadows, especially if a researcher’s head gets in the way, darkness is the normal state of the lab. Along these lines, John Dowling recounts a second Hartline-story of accidental discovery. The lab was dark, Hartline had his nose to the bench, and in walked an ignoramus who flipped on the lights. Immediately Hartline noticed that not all of the optic nerve fibers increased their firing frequency when exposed to this light source. Instead he discovered an array of inhibitory cells, which reduced their signaling frequency when exposed to light.

It might not be immediately clear how this work is so important, but the discovery of specialized cell types that work in parallel to solve a complex task is extraordinary, and had wide implications for neurobiology in general. These inhibitory and excitatory cells allowed for an elegant solution to the problem of edge detection. Again, it isn’t obvious to anyone who doesn’t study biological or computer vision why a solution to edge detection is so deeply important, however from a computational perspective this simple task poses a fractious design problem. However it is somewhat more straightforward if you are passingly familiar with a digital camera. The contrast between two objects in the visual field allows a camera, or an eye, to make guesses about the borders between objects that might otherwise be nearly indistinguishable. When you “sharpen” a picture you’ve taken, you’re modifying the picture, creating a representation that is less faithful to the real world you’re photographing. While it is less faithful, it may well be more useful. It helps you distinguish boundaries between objects, which might be more important than being able to precisely discriminate between colors.

//creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons

Hartline’s accidental discovery of the excitatory and inhibitory array didn’t only offer up a solution to edge detection, it resolved the Mach band effect. The Mach band effect is an optical illusion, noticeable when two similarly shaded shapes come into contact with each other. When this happens, the darker of the gray shapes will appear darker gray, and the lighter of the gray shapes appear lighter.   As we’ve already seen, the inhibitory arrays allow for edge detection by sacrificing fidelity to the “real” color and shading properties of the object. The laterally down regulated firing of these cell arrays down regulate so that contrast between objects is more obvious when the objects are close together. That is what the cells in your eye are doing when the Mach bands touch (Hartline and Ratliff 1957; Hartline 1956; Hartline and Ratliff 1958).