Visit ShanghaiTech University | 中文
HOME > News and Events > Seminars
Real-time High-resolution Functional Magnetic Resonance Imaging with GPU Parallel Computations
Date: 2016/10/21             Browse: 263
Seminar Topic: Real-time High-resolution Functional Magnetic Resonance Imaging with GPU Parallel Computations

Speaker: Zhongnan Fang
Time: Oct. 21, 3:30 p.m. - 4:30 p.m.
Venue: Room 1D-104, SIST Building

Abstract:  
Functional magnetic resonance imaging (fMRI) is a technique that enables non-invasive monitoring of brain activity by detecting changes in blood oxygenation levels. Recent advancements in MRI hardware and high performance computing have made it possible to achieve high spatiotemporal resolutions and real-time fMRI processing, respectively. However, it remains challenging to combine these two features at once due to the significant computational overhead. 

Another challenge facing real-time high-resolution fMRI is that current techniques do not provide the resolution needed for imaging small regions of the brain, such as cortical layers or hippocampal sub-regions. Compressed sensing (CS) is a potential method to address this problem, which features low hardware dependency and ease of integration with other high-resolution techniques. However, because CS reconstruction is an iterative and computationally intensive process, whether it can be performed fast enough for real-time fMRI is still an open question. 

In this talk, I will introduce a novel real-time high-resolution CS fMRI method to address these challenges. In the first part of the talk, a graphics processing unit (GPU) based high-throughput real-time fMRI system is introduced to overcome computational barriers associated with reconstruction of non-uniformly sampled images, motion correction and statistical analysis. The second part of the talk explores the feasibility and performance of CS fMRI, and demonstrates a high spatial resolution CS fMRI method. In the third part of the talk, methods from previous developments are combined and a high performance real-time high-resolution CS fMRI system is demonstrated. 
 
Biography:
Zhongnan Fang is a senior research scientist at LVIS corporation, Palo Alto, California, USA. He got his Bachelor's degree from Zhejiang University, Master of Science from University of California, Los Angeles, and Doctor of Philosophy from Stanford University. During his Ph.D. he was focused on developing new algorithms and methodologies for high spatial and temporal resolution functional magnetic resonance imaging (fMRI) using the compressed sensing and graphics processing unit (GPU) based parallel processing platforms. He joined LVIS corporation on 2015 and continues his research on novel medical imaging technologies.

Seminar 16076