Neuroscience aims to understand the brain, an organ that distinguishes human as a species, defines us as individuals, and provides the intellectual power with which we explore the world. Similar to systems studied in physics, where many properties emerge from the interactions of their components, the functions of the brain arise from the interactions of neurons. The fundamental computational units of the brain, neurons communicate with one another via microscopic structures called synapses, through which they form circuits and networks specializing in different mental functions.
To understand the brain mechanistically, we need methods that can monitor the physiological processes of single synapses as well as the activities of a large number of networked neurons. Just as physics has long employed optical methods to manipulate and probe matter on the scales from the infinitesimal to the immense, we develop optical tools to study the brain over multiple length scales, from single proteins therein to the entire brain itself. We aim at developing robust and versatile methods that make immediate impacts on brain research.
imaging the brain at diffraction-limited and super-resolved resolution
in vivo images of dendritic spines at 600 micron depth
To image neurons and their networks in vivo, light has to travel through the brain, where its wavefront is distorted by refractive-index mismatches. The aberrated wavefront leads to an enlarged focal volume with reduced intensity, compromising image brightness, resolution and contrast. A few hundred microns below the surface of the brain, it is no longer possible to study neural computation with single synapse resolution. The Ji lab has pioneered the use of adaptive optics (AO) to measure and cancel out sample-induced aberration in microscopy (reviewed in 2017, 2018, 2021, and 2023).
Applying AO to two-photon fluorescence imaging, we have demonstrated diffraction-limited imaging performance throughout mouse cortex. We have also applied AO to other imaging modalities, including super-resolution and optical-sectioning structured illumination microscopy, three-photon fluorescence microscopy, third-harmonics generation microscopy, and single-photon confocal fluorescence microscopy. Our collection of work has established AO as an essential component for high-resolution imaging of complex tissues.
Imaging the brain at greater depth
Activity of lateral hypothalamic neurons imaged at 5 mm depth
Imaging depth in the mammalian brain is ultimately limited by light scattering. To image deep structures, currently the only viable option is microendoscopy, where a miniature probe is embedded in the brain for image relay. We systematically characterized microendoscopy performance and designed minimally invasive systems that allowed the chronic imaging of neurons down to 5.0 mm depth. We are now expanding microendoscopy applications to other modalities and systems.
imaging the brain at high speed
30-Hz volumetric imaging of dendrites and dendritic spines in vivo
To understand information processing in neurons and neural networks, we need to record neural activity at high rate throughout 3D volumes. Taking advantage of the fact that, during most in vivo brain imaging experiments, the structures of interest are stationary, we obtained volume information by scanning an axially extended focus in 2D. Using wavefront engineering to generate Bessel foci optimized to maintain high lateral resolution in vivo, we imaged the brains of a variety of species and demonstrated high-speed volumetric imaging of neurons and neural compartments with single synapse resolution, over the mesoscale, and vasculature dynamics in 3D. We have engineered Bessel-droplet foci that enabled us to study how single neurons transform their volumetric excitatory synaptic inputs into output using glutamate imaging.
We also applied an ultrafast linescanning method called FACED to achieve kHz-frame-rate two-photon fluorescence microscopy, enabling voltage imaging of >200 neurons simultaneously.