Recent work in this project got the Best Paper Award in ISLPED'18, and Pearl Cohen Poster Award in 2019 UCLA Bioscience Innovation Day.
Moore's law has driven the exponential growth of information technology for over 50 years, during which the ever-increated computing power has made a huge impact on human lives. On the other hand, 50 years has passed since the Society of Neuroscience was founded, and the pace of research in understanding how brain works has never stopped. Under this big picture, this project focuses on creating high performance and energy efficient computation tools for neuroscientific research, and brain-inspired computing.
Miniaturized fluorescence microscope is an emerging technique in observing the activity of a large population of neuros in vivo. It opens up new research opportunities for neuroscientific experiments and understanding how brain works. We are working with leading team in this field at UCLA and focus on the backend real-time processing for the calcium imaging data collected from a 3-gram head mounted miniature fluorescence microscope http://miniscope.org/, as the figure shows. Technical challenges we are facing include: Acceleration for real-time calcium image analysis, which consists of pipelined processing stages such as motion correction, cell segmentation and neural activity extraction; Energy-efficient embedded computing on miniaturized device with limited power supply.
Recent advancements in microscopes allowing imaging an entire brain of a mouse in a couple days. However due to the intensive resolution and scale, computational processing of these images can take months. Although the results of these computational pipelines are analyzed offline after imaging, this processing can severely limits the workflow and hypothesis iteration speed. In our collaboration project with PIs Dr. William Yang (UCLA Neuroscience) and Dr. Hongwei Dong (USC Neuroscience) that was awarded an NIH U01 BRAIN grant, we accelerate such processing to study brain cell features of Huntington’s disease. The major run-time costs in these pipelines arise from converting dense three-dimensional images to a set of edges and nodes. These graphs offer a sparse and natural representation of neurons which have collections of branches whose pattern and connections determine their function or pathology. Graph processing algorithms are known for their acceleration difficulty due to their low computation to memory access ratio and their load-balance irregularity during parallelization. It is also increasingly more common for image pipelines to require additional specific image segmentation tasks such as cell body detection which are accomplished with more common neural network approaches. At a modest 30x magnification of a single mouse brain, the data size is still 20-30 teravoxels, therefore we focus on designs that optimize data-flow and prioritize computation at the edge, for example on the imaging workstation itself.
Broader impacts from these research projects shed light on establishing real-time closed-loop feedback control system, speeding up neural reconstruction of the entire brain of mouse, developing energy-efficient computing architecture, and accelerating the pace of neuroscience research and education worldwide.
 Zhe Chen, Hugh T. Blair, Jason Cong. LANMC: LSTM-Assisted Non-Rigid Motion Correction on FPGA for Calcium Image Stabilization, 27th ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA), Feb 24-26, 2019.
 Zhe Chen, Andrew Howe, Hugh T. Blair, Jason Cong. CLINK: Compact LSTM Inference Kernel for Energy Efficient Neurofeedback Devices, International Symposium on Low Power Electronics and Design (ISLPED), July 23-25, 2018. (Best Paper Award)
 H.T. Blair, J. Cong, D. Wu. FPGA Simulation Engine for Customized Construction of Neural Microcircuit. Proceedings of the 2013 International Conference on Computer-Aided Design (ICCAD), pp. 607-614, November 2013.
News: National Science Foundation will help UCLA spread technology behind miniscope.
Please read the full press at the following link: