Localization of multiple dipoles: mathematical programming approaches.

Although there has been progress in EEG and evoked potential analysis, the identification of underlying neural activity has eluded researchers despite its importance. This paper introduces the application of mathematical programming techniques to the "inverse" problem of three-dimensional localization of brain activity from scalp potentials. Preliminary computer experience with these optimization methods reported herein, yields encouraging results on simulated data and points new directions for research.

Computer and information networks.

The most basic conclusion coming out of the EDUCOM seminars is that computer networking must be acknowledged as an important new mode for obtaining information and computation (15). It is a real alternative that needs to be given serious attention in current planning and decision-making. Yet the fact is that many institutions are not taking account of networks when they confer on whether or how to replace their main computer. Articulation of the possibilities of computer networks goes back to the early 1960's and before, and working networks have been in evidence for several years now, both commercially and in universities. What is new, however, is the unmistakable recognition-bordering on a sense of the inevitable-that networks are finally practical and here to stay. The visionary and promotional phases of computer networks are over. It is time for hard-nosed comparative analysis (16). Another conclusion of the seminars has to do with the factors that hinder the fuller development of networking. The major problems to be overcome in applying networks to research and education are political, organizational, and economic in nature rather than technological. This is not to say that the hardware and software problems of linking computers and information systems are completely solved, but they are not the big bottlenecks at present. Research and educational institutions must find ways to organize themselves as well as their computers to work together for greater resource sharing. The coming of age of networks takes on special significance as a result of widespread dissatisfactions expressed with the present computing situation. There is a feeling that the current mode of autonomous, self-sufficient operation in the provision of computing and information services is frequently wasteful, deficient, and unresponsive to users' needs because of duplication of effort from one installation to another, incompatibilities, and inadequate documentation, program support, and user assistance. Complaints about the relative lack of uniform standards and the paucity of information on what programs and data are available and how to get and use them are commonplace. The human tendency, when beset by problems such as these, is to seek a savior in the next new technology-networks in this case. But networking does not in and of itself offer a solution to current deficiencies. What it does offer is a promising vehicle with which to bring about important changes in user practices, institutional procedures, and government policy that can lead to effective solutions. Thus more critical than whether networking is developed and applied is how it is developed and applied. For example, networking emphasizes the need for standards and good documentation. Unless effective mechanisms are developed and strong measures taken in networking to ensure that suitable standards and documentation are developed, present inadequacies could get worse, not better.