The eye is not just a lens that takes pictures and converts them into electrical signals, it is the first part of an elaborate system that leads to "seeing".   As with all vertebrates, nerve cells in the human eye separate an image into different image channels once it has been projected onto the retina and pre-sorted information is then transmitted to the brain as parallel image sequences. 

Perception by the brain is another matter.  For many animal species, "seeing" is one of the most important senses because every second the eyes record a huge number of impressions which are converted by the photoreceptor cells into electrical signals - that image processing begins in the retina of the eye, where nerve cells separate out the visual information in images featuring different content before transmitting them to the brain.

To understand something as complex as "seeing", scientists examined a somewhat simpler but extremely efficient system – the fruit fly's brain. Despite their tiny size, fruit flies are ideal candidates for such research because they are masters of visual processing, the number of nerve cells involved is manageable, meaning each individual cell can be examined, and genetic tools make it possible to block individual cells and analyze their role in the system.

Each individual eye in the fly's compound eye perceives "on" and "off" contrast changes. This information is split up right behind each individual eye (blue arrow). The L1 cells only transmit information about "on" edges ("light on"), the L2 cells only about "off" edges ("light off") to the output cells (green).  Credit: Max Planck Institute of Neurobiology / Schorner

Like vertebrates, the fly also transmits images directly from the sensory cells to various image channels. The information is then transmitted via a series of other cells to large, motion-sensitive nerve cells. These output cells of the motion vision system are responsible for visual flight control.

Scientists ascertained the early separation into different images by blocking certain cells using genetic engineering. They then exposed the flies to moving striped patterns in an LED arena and recorded the electrical responses of the large output cells. The various contrast changes, which occur as a result of the stripe movement, are perceived by the eye's photoreceptors. However, five nerve cells – the lamina cells L1 to L5 – are located directly behind every sensory cell.

"We had been wondering for a long time why there are so many cells, which of them transmit information to the motion vision system, and what kind of information this is," reports Alexander Borst of he Max Planck Institute of Neurobiology, who headed up the study. His team suppressed the activity of individual lamina cells while the flies watched the moving patterns. The experiments revealed that L1 and L2 cells are the main input channels into the flies' motion vision system. It got exciting as the team discovered that the cells transmit only certain partial information: L1 only responds if a dark/bright edge passes by (light on), while L2 only transmits information about a moving bright/dark edge (light off). This represents a clear parallel with the eyes of vertebrates, where ON and OFF bipolar cells, as they are known, also respond only to specific contrast changes.

"It can be no coincidence that we find this separation of contrast information in all vertebrates and now also in flies," says Borst. He has a hypothesis as to why this wiring has been maintained so consistently throughout evolution: it allows the brain to save energy. If only one cell relayed the information about the various contrast changes, it would have to maintain a basic membrane voltage, which would intensify in a "light on" situation and weaken in a "light off" situation. Such a basic membrane voltage requires energy. Having two cells is therefore more efficient, as only one needs to be active when "its" contrast change occurs.

Citation: Maximilian Joesch, Bettina Schnell, Shamprasad Varija Raghu, Dierk F. Reiff, Alexander Borst, 'ON and OFF pathways in Drosophila motion vision', Nature 468, 300-304 (10 November 2010) doi:10.1038/nature09545