The eye presents a paradox to experimental neuroscientists. On one hand, it’s accessible, its function is well understood, and its inputs can be precisely controlled, says neurobiologist William Newsome of Stanford University. “On the other hand, it is very difficult to record its electrical signals while visual behavior is actually taking place.” That’s because the hardware for electrical recordings—rigid electrodes—aren’t compatible with small, spherical, constantly moving rodent eyes.
To record retinal cell activity, researchers tend to remove the eye from the animal, dissect the retina,...
Measuring 1.5 mm by 0.8 mm and containing 16 individual electrodes, the polymer and metal meshes are injected, one per mouse eye, toward the rear of the vitreous body, where they unfurl to coat the retina. Tiny wires connected to the meshes extend out of the corner of the animals’ eyes for attachment to an external recording device. The animals are restrained, their heads immobilized, for sessions of light stimulation and recording, but they are free to move and behave normally between sessions. Remarkably, the meshes have little effect on vision, and after a few weeks they detach from the retina.
Long-term, in vivo recording ability opens a range of new research avenues, says Lieber, whose team has used the meshes to measure changes in retinal ganglion cell activity over the course of several day/night cycles.
It’s a “fantastic” innovation, says Marla Feller of the University of California, Berkeley, who studies the organization of retinal neural circuits during development but was not involved in the research. “It was exciting to see how successful it was at being able to record from the same cells over multiple days.” Not only might the method be useful for following individual cells through retinal maturation, she says, but also for examining which cells die and which survive during eye diseases such as glaucoma. (Science, 360:1447–51, 2018)
Multi-cell retinal recordings | Method | Viable recording time | No. of simultaneously recording electrodes | Invasive? |
Ex vivo | The eye is removed, and the retina is dissected and laid flat on an array of microelectrodes. | 8 to 12 hours for mouse retinas | Approximately 500 | Highly. Animal sacrificed |
In vivo | Mesh electrode is injected into the eye of an anesthetized mouse. | Several weeks | 16 so far but could theoretically be more | Minimally. Visual function negligibly affected |