White Matter Tractography and Human Brain Connections Using GPUs (Moises Hernandez-Fernandez from the Oxford Centre for Functional MRI of the Brain)

Moisés Hernández-Fernandez from the Oxford Centre for Functional MRI of the Brain, University of Oxford, UK will present a seminar on the
1st of June 2016, (at 1pm) entitled

“White Matter Tractography and Human Brain
Connections Using GPUs”

Abstract

Understanding the human brain is one of the key scientific challenges of the 21st century. A key component for gaining insight into brain functional mechanisms is understanding the underlying anatomical and structural organisation and how functional subunits are interconnected at the global and local scale. Studying brain organisation (and its potential disruptions) is also a necessary first step for investigating the etiology, progression and pathology of neurological and psychiatric diseases, such as multiple sclerosis, schizophrenia and Alzheimer’s disease. Diffusion MRI (dMRI) uniquely allows us to study such organisation, reveal the wiring diagram of the brain, and provide estimates of tissue microstructure non-invasively and in-vivo. Despite its great potential and its wide use, the analysis includes the very time-consuming statistical modelling of four-dimensional datasets that restrict the potential and the clinical applicability of the technology. We present a novel parallel framework for analysis of dMRI data that exploits the immense computational power provided by modern GPUs and allows accelerations of up to two orders of magnitude when comparing GPU performance with a single-threaded CPU implementation.

This talk details the research that led to the NVIDIA 2016 GPU Centre of Excellence Achievement Award.

About the speaker

Moisés Hernández-Fernandez is a DPhil student at the Oxford Centre for Functional MRI of the Brain, University of Oxford. His research focuses on high performance computing applied to the analysis of diffusion MRI (dMRI) data. He is interested in how parallel computer architectures, such as Graphics Processing Units (GPUs), can be used in scientific applications that require very high computational resources. His current work uses GPUs for resolving tissue microstructural patterns and for estimating long-range brain connectivity.

This seminar is open to all and will start at 1pm in the OeRc Conference Room
(Room 278)

****Lunch will be made available****