Steering Group Report: Brain Science as a Mutual Opportunity for the Physical and Mathematical Sciences, Computer Science, and Engineering. Held in Arlington, Virginia on August 21-22, 2006

摘要

Humankind now stands at a special moment in its long history of thinking about the brain, a moment of revolutionary change in the kinds of questions that can be asked and the kinds of answers that can be achieved. Fundamental shifts include: The Scope and Scale of Experimental Investigations: Instead of one- or few-at-a-time measurements, it is becoming possible to measure brain structure, chemistry, and activity simultaneously at many locations with high specificity and spatial/temporal resolution. The Character of Theoretical Understanding: Instead of mainly bottom-up or top-down models and theories, it is becoming possible to formulate comprehensive multi-scale models that are both bottom-up and top-down and include relevant dynamics at different spatial and temporal scales. The Ways in Which Knowledge Can Be Used: Applications for the emerging multi-disciplinary knowledge about the brain abound: In large-scale neural simulations, in robots and other engineered systems that mimic biological systems, and in brain-computer interfaces that enable bi-directional communication for next-generation neural prostheses. In this time of change there are significant unexploited opportunities for mutual scientific benefit between brain science and the physical and mathematical sciences, computer science, and engineering. Four broad areas of opportunity were identified: Because of its strong record of leadership in the physical and mathematical sciences, computer science, and engineering, NSF is well-positioned to enable and exploit the following opportunities: Opportunities in Instrumentation and Measurement: New instruments, probes, and experimental tools are needed for comprehensive measurement of the structure, chemistry, and activity of individual nerve cells and neural populations in functioning neural systems. Such tools will permit vastly improved experimental studies of neural dynamics that accompany development, learning, cognition, and behavior. Opportunities in Data Analysis, Statistical Modeling, and Informatics: The availability of immense quantities of high-resolution data in turn will demand new statistical tools and models, and new informatics capabilities for storage, representation, and modeling of high-throughput multiresolution data. New approaches for inferring association, linkage, and causality will be required. Opportunities in Conceptual and Theoretical Approaches: Advances in analysis and modeling of comprehensive multi-scale data will enable the exploration of much richer conceptual and theoretical approaches to understanding the brain at all levels. New mathematical approaches to understanding very high-dimensional, non-linear, non-stationary, multi-scale systems will be required. Opportunities in Building Brain-like Devices and Systems: Improved understanding of the brain, combined with advances in engineering capabilities, will permit revolutionary advances in neurally inspired computing and information processing, in the design of robots and other engineered systems that mimic biological capabilities, and in brain-computer interfaces that enable bi-directional communication with the brain in real time.