The overall goal of the Computational Neuroscience Core Facility is to provide the resources necessary for the Program Project to develop and analyze biologically meaningful models of neural function. To achieve this goal the Core provides: 1) computational resources such as high-performance computers, desktop workstations, centralized data storage and archieving, account management, programming expertise, printers, and plotters; 2) administrative support such as research assistance, clerical assistance and grants management; and 3) scientific consultation by forming an External Scientific Advisory Committee (ESAC) and by sponsoring visits by leading neuroscientist who serve as consultants to the individual research Projects. In addition, the Core facilitates collaborations among the research Projects, in part, by sharing data and technology among the Projects. In the proposed funding period, the Core will implement a two-year program for establishing a 32-node 'Beowulf' cluster (high performance computing cluster, HPCC). The HPCC will increase the computational capabilities of the Core from 5 to 200 GFLOPS (2x10[9] Floating-point Operations per Second), and will substantially reducing operating costs. To help ensure that the Project Leaders, research fellows and graduate students remain at the cutting edge of computational neuroscience, the Core will invite leaders in the field to visit and consult with members of the Program Project. For example, Drs. Bruce McNaughton and John Tyson have agreed to visit during Year 1 of the proposed funding period. In addition, the Core has assembled an ESAC whose members include Drs. Daniel Johnston, Eve Marder, John Rinzel and Terrence Sejnowski. The ESAC will provide annual reviews of the Program Project as well as: 1) advise us on the direction and scope of our work; 2) advise us on training our graduate students and research fellows; and 3) discuss their work and its implications for improving our studies. This combination of state-of-the-art computational resources, personnel resources, and scholarly environment has proven to be effective in producing high quality computational studies of neural function and for providing the best training possible for our graduate students and postdoctoral fellows.
Showing the most recent 10 out of 62 publications
