Database and tools for analysis of topographic organization and map transformations in major projection systems of the brain.

Integration of dispersed and complicated information collected from the brain is needed to build new knowledge. But integration may be hampered by rigid presentation formats, diversity of data formats among laboratories, and lack of access to lower level data. We have addressed some of the fundamental issues related to this challenge at the level of anatomical data, by producing a coordinate based digital atlas and database application for a major projection system in the rat brain: the cerebro-ponto-cerebellar system. This application, Functional Anatomy of the Cerebro-Cerebellar System in rat (FACCS), is available via the Rodent Brain WorkBench (http://www.rbwb.org). The data included are x,y,z-coordinate lists describing exact distributions of tissue elements (axonal terminal fields of axons, or cell bodies) that are labeled with axonal tracing techniques. All data are translated to a common local coordinate system to facilitate across animal comparison. A search capability allows queries based on, e.g. location of tracer injection sites, tracer category, size of the injection sites, and contributing author. A graphic search tool allows the user to move a volume cursor inside a coordinate system to detect particular injection sites having connections to a specific tissue volume at chosen density levels. Tools for visualization and analysis of selected data are included, as well as an option to download individual data sets for further analysis. With this application, data and metadata from different experiments are mapped into the same information structure and made available for re-use and re-analysis in novel combinations. The application is prepared for future handling of data from other projection systems as well as other data categories.

AI in medical education--another grand challenge for medical informatics.

The potential benefits of artificial intelligence in medicine (AIM) were never realized as anticipated. This paper addresses ways in which such potential can be achieved. Recent discussions of this topic have proposed a stronger integration between AIM applications and health information systems, and emphasize computer guidelines to support the new health care paradigms of evidence-based medicine and cost-effectiveness. These proposals, however, promote the initial definition of AIM applications as being AI systems that can perform or aid in diagnoses. We challenge this traditional philosophy of AIM and propose a new approach aiming at empowering health care workers to become independent self-sufficient problem solvers and decision makers. Our philosophy is based on findings from a review of empirical research that examines the relationship between the health care personnel's level of knowledge and skills, their job satisfaction, and the quality of the health care they provide. This review supports addressing the quality of health care by empowering health care workers to reach their full potential. As an aid in this empowerment process we argue for reviving a long forgotten AIM research area, namely, AI based applications for medical education and training. There is a growing body of research in artificial intelligence in education that demonstrates that the use of artificial intelligence can enhance learning in numerous domains. By examining the strengths of these educational applications and the results from previous AIM research we derive a framework for empowering medical personnel and consequently raising the quality of health care through the use of advanced AI based technology.

Health care workers and their needs: the forgotten shadow of AIM research.

The field of AI in Medicine (AIM) seems to have accepted that decision support is, and will be, needed within most medical domains. As society calls for cost-effectiveness, and human expertise or expert guidance are not always available, decision support systems (DSSs) are proposed as the solutions. These solutions, however, do not necessarily correspond with the basic needs of their targeted users. We will show this through a review of the literature related to health care workers and the various factors that have an influence on their performances. Furthermore, we will use these empirical findings to argue that the AIM community must go beyond its decision support philosophy, whereby the gaps in human expertise are filled in by the computer. In the future, joint emphasis must be placed on decision support and the promotion towards independent and self-sufficient problem solving. In order to implement this paradigm change, the AIM community will have to incorporate findings from the research discipline of AI in Education.