Concluding Remarks

While most organ systems have become less mysterious with the advent of powerful molecular, cytological, and functional approaches, with the brain things are almost going in the opposite direction. Each new finding in molecular, cellular, and functional domains adds additional levels of complexity. The challenge is that more information does not necessarily lead to more understanding unless we are able to correlate the diverse data into a single framework. My aim is to build multimodal imaging tools and analytical approaches that integrate diverse streams of data into a single framework.

- Toward whole-brain connectomics: Brains are challenging to image because they function by means of synaptic connections that are not easily seen without the aid of nanoscopic imaging devices (i.e., electron microscopes). To obtain synapse-level connectivity maps of the brain, we use serial section electron microscopy (EM), owing to its high resolution to resolve neuron-to-neuron connections. At the same time, each neuron in the brain projects its axon over millimeters or even centimeters in all dimensions. Identifying the synaptic connectivity between neurons requires nanoscale resolution but over vast volumes. This juxtaposition of big and small presents a supreme challenge. While whole-brain imaging with EM is feasible in small animals like C. elegans and Drosophila, comprehensive mapping of a mammalian brain, the ultimate goal of connectomics, is hindered by the lack of a large-volume brain staining method.

I have developed a fixation method that preserves extracellular space through the entire mouse brain, a key for high fidelity reconstruction and improved staining efficiency, and a novel EM staining strategy that increases the volume limit by 500 times, from 1 mm³ to 500 mm³ (i.e., the volume of an adult mouse brain), making the whole mouse brain connectomic imaging possible. By integrating time-lapse X-ray microCT in the method development, I have significantly advanced our understanding of the chemical and physical processes involved, ensuring scalability to even larger brains, such as primate brains.

- Interrogating molecular identities of brain cells with ultrastructure-preserved correlated light and electron microscopy (CLEM): Neurons comprising the network display extraordinary molecular diversity, underpinning their distinct way of processing and transmitting information. However, inferring the specific biochemical and electrophysiological types of neurons from EM images is difficult. Without molecular knowledge of the neurons, determining their roles in the circuit becomes challenging.

I have demonstrated that small engineered immunoprobes, such as nanobodies and single-chain variable fragments of antibodies (scFv), can pass the membrane without using detergent to permeabilize cells, a detrimental procedure for EM imaging. These methods, therefore, allow for the superimposition of multiple immunolabels onto ultrastructure-preserved EM images to reveal the molecular identities of brain tissues (i.e., our current record is 5 colors in one tissues volume), thus enabling us to decipher the working principles of EM-reconstructed networks.

Future Direction: An Integrated Imaging Platform to Study the Brain. My independent research aims to integrate various imaging modalities, which are typically challenging to combine, into unified pipelines. This will allow for the correlation of diverse data sets to investigate changes in brain circuits over the lifespan and interruptions caused by disorders. The plan entails:

Aim 1: Cross-region connectivity mapping to improve the throughput of connectomic imaging.

Aim 2: Expansion X-ray microscopy to enable comparative structural analysis.

Aim 3: Superplexed CLEM to investigate molecular and ultrastructural diversity of brain cells

Outreach Aim: Timestamped electron microscopy to superimpose neuronal activity information

A critical shortcoming of single modalities, be they based on one imaging modality like electron microscopy, or molecular analysis such as transcriptomics, or one assay of activity such as Calcium imaging, is that they exist in an isolated domain without context. I believe my efforts to combine structural, molecular, and activity information into one dataset is absolutely necessary to better understand how changes in brains over the lifespan come about.

Teaching Interests. I have had the opportunity to teach both entry-level undergraduate and advanced graduate courses, notably as a teaching assistant (TA) for Nanomaterials Chemistry & Engineering at the University of Texas at Austin and as head TA for Principles of Chemistry at Tsinghua University. My responsibilities in these courses included teaching sections, preparing handout notes for students, grading assignments, meeting with students during office hours, and coordinating the work of TAs.

Given my background and training, I am prepared to teach a broad range of introductory courses related to chemical/materials engineering and microscopy. In addition, I would welcome the opportunity to teach more advanced courses in my area of expertise. Finally, I would be delighted to develop new interdisciplinary courses according to departmental curricular needs.