The fruit fly brain has become one of the most popular model organisms to study neural computation and for relating brain structure to function. Many of the genes and proteins expressed in the mammalian brain are also conserved in the genome of Drosophila. Remarkably, the fruit fly is capable of a host of complex nonreactive behaviors that are governed by a brain containing only ~100000 neurons. The relationship between the fly's brain and its behaviors can be experimentally probed using a powerful toolkit of genetic techniques for manipulation of the fly's neural circuitry. Novel experimental methods for precise recordings of the fly's neuronal responses to stimuli and for mapping neurons and synapses in Drosophila nervous system have provided access to an immense amount of valuable data regarding the fly's neural connectivity map and its processing of sensory stimuli. These features coupled with the growing ethical and economic pressures to reduce the use of mammals in research, explain the growing interest in Drosophila-based brain models, not only to understand sensing, perception and neural computation but also to elucidate human neurodegenerative diseases such as Alzheimer's disease.
Despite significant investment and huge progress in understanding Drosophila neural circuits and the availability of excellent genomic and genetic community databases, a major obstacle in understanding the fly brain is the lack of communication/collaboration across the modelling community as well as lack of shared models, modelling tools and data repositories. Vast amounts of experimental data that have yet to be distilled into new models or used to validate and refine existing models, have been generated by labs around the world. Knowledge and information of the detailed neuroanatomy, neuron connectivity and gene expression of the adult Drosophila melanogaster brain has been made publicly available thanks to the efforts of earlier pioneering efforts.
This aim to develop an open source, modular software platform that will help researchers to work collaboratively and exploit the wealth of knowledge, data, models, and tools available to build and simulate a complete model of the fly brain. The software platform exploits relatively cheap supercomputing services that use Graphic Processor Units, which many academic institutions in the UK, US and worldwide haave adopted in recent years.
a) NVIDIA (Project Partner)
As many universities have invested in GPU-based commodity supercomputing services, NVIDIA is very interested to understand and meet the needs of the science community in order to develop successful products for very competitive market. It should be noted that big players like Intel and IMB are in direct competition with NVIDIA in the High-Performance Computing/Supercomputing market. As Project Partner NVIDIA will have direct, first-hand access to our results. The project will give the valuable insight into the current limitations of, particularly, their connectivity architecture for such applications. NVIDIA will also have access to novel software architectures that exploit parallelism, which will enable them to optimize and develop further their CUDA tools, libraries, languages and other development tools.
The Automatic Control and Systems Engineering Department at the University of Sheffield hosts the Rolls-Royce University Technology Centre in Control and Systems Engineering (RR-UTC). RR-UTC provides the company with the necessary technology to support the efficient production of world-class engine control and monitoring systems. The Centre recently achieved economic impact by developing the first radically new control laws for gas turbine engines for 30 years, which are now operating across the entire engine range, including the flagship Rolls-Royce Trent 1000 engines powering the Boeing Dreamliner.
One of the research areas being investigated within the Centre is the use of compressive sensing techniques for efficient, transmission and reconstruction of the signals from the sensors acquiring the data. This has the potential to reduce significantly the bandwidth required for data transfer from a vast array of sensors and pave the way for the implementation of wireless control systems. Another area of active research is in distributed decision support systems for health management of the fleet of engines, supported by high performance computing architectures. The proposed project is of particular interest because the brain employs exactly the type of low-energy, low bandwidth but highly robust coding strategies that are desired by the company. Moreover, by understanding the brain, which in effect is a highly sophisticated control system, it would be possible to develop alternative neuromorphic distributed control architectures which could be much cheaper to implement as well as more robust and fault tolerant.
|Description||We demonstrated that fly photoreceptors use nonlinear transformations of visual stimuli to encode efficiently edges present in temporal stimuli.
Specifically we showed recently that nonlinear encoding achieves near optimal rate/distortion bound.
|Exploitation Route||improve the design of artificial retinas|
Aerospace, Defence and Marine,Healthcare
|Country||United States of America|
|PI Contribution||We developed, and refined models of the retina for simulation on the Neurokernel platform.|
|Collaborator Contribution||Developed the initial open source simulation platform Neurokernel. Provided tools and resources for our group to use.|
|Impact||Open Science Prize Application This is a multi-disciplinary collaboration to develop models and a software platform for simulating the entire brain of the fruit fly: biology, computer science, control and systems engineering|
|PI Contribution||Evaluation of new GPU architectures|
|Collaborator Contribution||Training and technical support. In-kind contribution of £30K|
|Organisation||University of Oxford|
|Department||Department of Physiology, Anatomy and Genetics|
|Country||United Kingdom of Great Britain & Northern Ireland (UK)|
|PI Contribution||We are developing modelling and simulation tools to make experimentally plausible and accurate models of the fruit fly brain.|
|Collaborator Contribution||They support the drosophila brain project with data, evaluation & testing of the software, and to help validate models.|
|Impact||No outputs yet. Multi-disciplinary: biology, systems engineering, control engineering|