Three-photon microscopy gives neuroscientists a deeper look at brain cells than previously able. Now, based on a substantial refinement of the technology, scientists at the Massachusetts Institute of Technology (MIT; Cambridge, MA) have conducted a study of stimulated neural activity in an awake mouse through every visual cortex layer and notably the mysterious "subplate" below.
In the study, the research team showed that as mice watched visual stimuli, their human observers could measure patterns of activity among neurons in all six layers of visual cortex and the subplate, providing new data about their role in how mammals process vision. Moreover, through a series of careful experiments, the researchers were able to show that the light they sent in, as well as the light that came back out, neither damaged, nor even altered, the cells they measured.
The research team's paper describes a new three-photon microscope optimized to deliver rapid, short, low-power pulses of light capable of reaching deep targets without causing any functional disturbance or physical damage, and then to detect the resulting fluorescence emitted by cells with high efficiency to produce images with sharp resolution and a fast frame rate.
The labs of co-corresponding authors Mriganka Sur, Newton Professor of Neuroscience in the Picower Institute for Learning and Memory at MIT, and Peter So, professor of mechanical engineering and biological engineering, have joined in pushing the frontiers of multiphoton microscopy. In the new study, they show they've now taken it far enough to study live neural activity. To do that, the team sought to refine many different parameters of both the laser light and the scope optics, based on meticulous measurements of properties of the brain tissue they were imaging. For instance, they not only measured the energy at which cells started to show overt damage (about 10 nJ), but also measured the power at which cells would start to behave differently, thereby producing data influenced by the measurement (2–5 nJ). With precision and purpose to deliver lower energy levels, the scientists optimized the scope to emit incredibly short pulses of light lasting for a pulsewidth of only 40 fs, and painstakingly arranged the optics to maximize the collection of the light that molecules, excited by the incoming laser energy, would emit back.
After carefully validating that the optimized three-photon microscope's measurements agreed with those of two-photon microscopes (in shallower layers of the cortex) and electrophysiology (which can go deeper, but blindly), the team set out to do direct visual observation of neural activity in all cortical layers of awake, behaving animals.
In the lab, they showed mice some grating patterns in 12 different rotated orientations and two directions of motion across a screen. With their optimized three-photon microscope, they watched neurons in each layer of the cortex—going more than a millimeter deep—to see how the cells reacted to this standard visual input. They could see the activity of the cells because they had engineered them to glow upon elevated calcium activity, using a label called GCaMP6s. They could see other tissues like blood vessels and white matter via a phenomenon called third-harmonic generation (THG).
Three-dimensional rendering of a sequence of 450 lateral three-photon images acquired with 2 μm increment from the visual cortex (layer 1 on the left to the subplate on the right); the green color represents GCaMP6s signal and the magenta color represents label-free THG signal generated in the blood vessels and myelin fibers in the white matter. Scale bar: 100 μm. (Image credit: Murat Yildirim et al.)
With the capability to see the deepest layers, they observed that layer 5 neurons are "broadly" tuned for orientation, meaning they respond to a wide variety of orientations, rather than just one or two specific ones. Layer 5 neurons also had more spontaneous activity than cells in other layers and more connections to deeper parts of the brain. Meanwhile, layer 6 neurons had somewhat sharper orientation tuning than neurons in other layers, meaning they are more specific in their response to distinct orientations.
Their most surprising finding was that the subplate, a thin layer of mostly neural "white-matter" fibers, was home to a population of neurons with patterns of activity that were weakly and broadly tuned to the visual input. The finding was revelatory, the researchers say, in that many neuroscientists believed that subplate neurons were mostly only active during development. The layer is also too thin to be measured with electrophysiology, says Murat Yildirim, a postdoctoral researcher and the study's co-lead author (postdoc Hiroki Sugihara is the other).
"So far, subplate neurons in the mature brain have not been studied at all due to the technical challenges of imaging these cells in vivo," the researchers wrote.
Full details of the work appear in the journal Nature Communications.