BioPhotonics spoke with Weijian Yang, an associate professor in the department of electrical and computer engineering at the University of California, Davis and a leader of a team that developed a two-photon fluorescence microscope that captures high-speed neural activity. The system, described in Optica, uses a line of light, rather than a dot of light, to scan and illuminate the brain area where neurons are active. This technique accelerates image collection. What are the advantages of collecting information on neuronal activity 10× faster than traditional two-photon microscopy? Traditional two-photon microscopy typically records neuronal activity using calcium indicators at speeds of 30 to 60 Hz, which is sufficient for imaging a single plane. When imaging multiple planes in depth, the speed decreases proportionally with the number of planes. Increasing the imaging speed allows us to compensate for this loss by capturing the fast dynamics of neuronal signals in a 3D volume and studying how neurons interact within a large network. Increasing the imaging speed also benefits voltage imaging. When a neuron fires, the calcium signal can last for hundreds of milliseconds or longer, even though the actual spike only lasts for milliseconds. Voltage indicators have much faster dynamics and can thus more faithfully represent neuronal activity. To take full advantage of voltage indicators, we need imaging speeds of at least several hundred hertz, ideally reaching into the kilohertz range. So, line illumination versus point illumination for adaptive sampling is accomplished with a 920-nm femtosecond laser and a digital micromirror device for spatial modulation? Two-photon microscopes shine a small dot of femtosecond laser light, typically in infrared wavelength, and scan it across the entire sample area, collecting signals dot by dot. Once it finishes scanning a single frame, it repeats the process for the next frame. This method is slow, resulting in a low frame rate, and it could risk damaging the brain by depositing excessive light and heat. Compared with a single dot of light, a short line can sample a larger area of neurons at once, significantly speeding up the imaging process. Additionally, our approach only images the regions of interest, which are the neurons. This is achieved using a digital micromirror device, a chip containing thousands of tiny mirrors that can be individually controlled. This technology allows us to dynamically pattern the light beam, enabling precise targeting of active neurons without illuminating unnecessary areas and thus minimizing heat generation. This is why we call our method adaptive sampling, as the sampling adapts to the underlying neuronal structure of the brain tissue. Your research refers to trying to follow brain activity during learning. What are your next steps? There are two main directions for our next steps. The first involves further tool development and demonstration: specifically, voltage imaging and volumetric imaging. For voltage imaging, it allows for the measurement of electrical signals across neuron membranes, providing a direct and extremely rapid readout of neural activity. Incorporating voltage imaging would enable us to observe transient electrical changes in neurons with greater temporal precision. By combining our method with beam multiplexing techniques and using a faster scanner, the imaging speed could reach kilohertz levels, making it suitable for voltage imaging. For volumetric imaging, it will capture 3D images of neural activity over time, which is crucial for understanding the complex spatial arrangements and neuronal interactions within large neural networks. Implementing remote focusing techniques will be necessary to achieve high-resolution volumetric imaging efficiently. The second direction is to apply this technique in real neuroscience applications. One example is to study how the functional structure and organization of the neuronal network evolve during learning. Specifically, we want to understand how knowledge or memory is created and represented in the neuronal network and how this representation changes over time. Our tools will enable us to record brain activity during the animal’s learning behaviors.