Bryan P. Bergeron

Morphing as a means of generating variation in visual medical teaching materials

In computer-based medical education, there is frequently a need to present students with pictorial data representative of the natural variation associated with disease presentations as well as the progression of disease within an individual. Because of the difficulty in acquiring such data, image acquisition is often the most resource-intensive phase of multimedia courseware development. In light of the resource demands associated with image content, many courseware designers do not make opportune use of image data, but rely instead upon text descriptions to provide variation in content. The resulting lack of adequate pictorial content often lessens the overall impact of the courseware. To overcome constraints imposed by the difficulty in acquiring pictorial content of sufficient richness, a methodology of generating variation in visual teaching materials has been developed through the use of morphing. These techniques have general applicability in creating variation in pictorial teaching materials in a variety of image-intensive domains.

Clinical skill-building simulations in cardiology: HeartLab and EkgLab

HeartLab and EkgLab are two simulation-based programs designed to teach medical students the essentials of the auscultatory cardiac examination and of electrocardiogram interpretation, respectively. The issues considered throughout the development of these projects, namely implementation language selection, program architecture, simulation design, patient models, and the approach to verification, are applicable to the design of any simulation-based system.

Data qualification: logic analysis applied toward neural network training

For neural networks to develop good internal representations for pattern mapping, noise in the training set data must be controlled. Because of the many difficulties associated with manually validating training data, we have focused on using decision table techniques as a practical, domain-independent means of optimizing training set formulation. Decision tables provide a variety of mechanisms whereby training set data can be processed to remove ambiguity, contradictions, and other noise. In addition to serving as data filters, decision tables can be used in the evaluation of neural network training.

HeartLab: a multi-mode simulation for teaching cardiac auscultation

A threaded interpretive language (TIL), Forth, was utilized to im plement HeartLab, a simulation-based program designed to teach medical students the essentials of the auscultatory cardiac exam. The simulation output, high-resolution bit-mapped graphics cou pled with realistic, computer-generated heart sounds, makes full use of the Apple Macintosh sound and graphics capabilities. Multimode simulation of this type and construction of an ap propriate patient model create a number of design challenges for which this approach proved to be particularly well suited.

Academia, privacy and modern information technology: partnering with industry in the modern economy

Authors and educators were once thought of as relatively independent purveyors of knowledge. Mter fulfilling their teaching obligations, academicians were encouraged disseminate their ideas as widely as possible, in exchange for peer recognition credit toward tenure, or evidence of scholarl y work for grant applications. However, modem economic pressures, fueled by advances in information technology, have permanently transformed the academician’s perception of what is private, what is publishable, and the degree to which information both professional and personal, should be made freely available on the Intemet and other digital media for non-targeted dissemination. Academicians and academic institutions as a whole have become highly protective of their increasingly valuable knowledge assets.The result of this paradigm shift is a redefinition of the contemporary role of academia, especially as it relates to the potentially lucrative associations with business and industry. In order to survive, many academicians must become knowledge workers and brokers who look not only to government and private grants for support, but increasingly to partners in business and industry who are willing to cooperate in mutually beneficial and profit-generating ventures.

Challenges associated with providing simulation-based medical education

Computer-based instructional programs in medicine, especially those that make use of graphics and simulation technology, are discussed. A number of impediments to this mode of teaching medical students are discussed, including: (1) the lack of knowledge about the learning process; (2) nonadaptive tutorial strategies; (3) difficulties associated with assessing simulation validity; (4) questionable transferability; (5) the lack of intuitive user interfaces; (6) the high cost of developing quality educational software; (7) the inability to customize applications to suit local teaching practices; (8) the high cost of computer hardware; and (9) ineffective evaluation methodologies.<<ETX>>

Imageview: A High-Level Authoring Tool for Repurposing Multimedia Content

The teaching of medicine lends itself to the use of a rich mixture of sounds, text, graphics, images, and animated sequences to illustrate concepts. Often, these multimedia teaching materials ar available only in limited contexts, such as during a specific lecture or within a particular software program. As a result, students have little opportunity to review non-textual information, and courseware authors must obtain their own images and other content that may in turn be encapsulated within a software program. In response to these and other limitations of traditional content management, we have developed an infrastructure and tools supportive of making image-based content available both for student review and for courseware authors to use as a repurposeable resource.

Computer-aided clinical problem solving as an educational paradigm for teaching preclinical cardiac pathophysiology

Clinical problem solving exercises have long been used as a way of providing simulated experience with clinical medicine. Multimedia capabilities now available extend the potential for realism of these exercises. In addition, with the use of hypermedia methodologies, it is now possible to link components of a clinical case — e.g., the tests, the images (or sounds or tracings) of the actual results, or the textual descriptions of findings — with other related information. This information can amplify on the tests themselves, the range of possible results that could alternatively have been obtained, the differential diagnosis of any particular finding, or the profiles of diseases in the differential, and can provide discussions and references about the diseases. Thus the case problem solving exercise may have potential for more than providing practice with making clinical judgments; it may offer an alternative paradigm for accessing much of the same medical content traditionally delivered in didactic lectures and in textbook presentations. We have implemented an authoring environment for case problem solving exercises that enables access to multimedia presentation of clinical results, and related clinical and pathophysiologic content material. Case simulations designed with this approach are currently being used in teaching first-year medical students in the HST track of the Harvard Medical School.