Center for Computing Research
James Bradley Aimone
|James Bradley Aimone|
Data-driven & Neural Computing
Sandia National Laboratories
P.O. Box 5800, MS 1327
Computational Function of Adult Neurogenesis
Much of my research has focused on developing computational models of adult neurogenesis, the process by which neural circuits in the brain add new neurons continuously throughout life. Adult neurogenesis has been implicated in the memory deficits observed in a number of neurological and psychiatric conditions, such as depression and traumatic brain injury. In the healthy brain, new neurons are believed to be linked to a number of aspects of episodic memory formation.
Prior to Sandia, I co-authored several papers using theoretical and computational techniques to describe the function of these new neurons in memory formation (Aimone et al., Nature Neuroscience 2006, Aimone et al, Neuron 2009, Aimone et al., Neuron 2011). At Sandia, my research has focused on ultra-large scale simulations on high performance computing platforms, enabling us to investigate the effect of neurogenesis in realistic scale systems.
Brain Inspired Computing
The computation that occurs in neural circuits is fundamentally distinct from conventional computing technologies. The brain utilizes highly parallel, non-linear algorithms to perform complex tasks, such as image recognition, memory formation, and decision making. Neural circuits are assembled in highly parallel, densely interconnected architecture in which memory, processing, and communication are effectively colocalized; providing considerable advantages in terms of processing speed, energy consumption, and reliability.
Successful development of neural inspired computing hardware has the potential to revolutionize how computing is used; particularly in data-centric problem domains relevant to both national security and commercial domains. In particular, my reserach focuses on the best practises for the development of neural algorithms based on modern neuroscience; including assessing the value of leveraging neural plasticity processes such as continuous neurogenesis in neural hardware.
Analysis and validation of biologically realistic neural simulations
Both the computational neuroscience and brain inspired computing work described above are increasingly pushing the boundaries of conventional computer simulations. In particular, the design and validation of neural models - whether intended to understand the brain or engineer new hardware - requries the development of metrics and procedures to quantify and validate the function of a neural system. This includes the development of new techniques for assessing information content and neural processing in high dimensional neural circuits as well as developing techniques in sensitivity analysis and uncertainty quantification of neural models both implemented on conventional hardware or novel neural architectures.
Ph.D. Computational Neuroscience, University of California, San Diego 2009
Thesis title: “Computational modeling of adult neurogenesis in the dentate gyrus”
Masters of Chemical Engineering, Rice University, Houston 2002
BS in Chemical Engineering, Rice University, Houston 2001
Selected Publications & Presentations
Awards & Recognition