Christian Grévisse

Radioanatomy, short for radiographic anatomy, is the study of anatomy through medical imaging. Its early‐stage introduction into medical curricula has been recommended in the literature. As with many other medical courses, it has seen a shift toward blended learning, including assessment on learning management systems such as Moodle, one advantage being automatic or at least assisted grading. The majority of previous studies in the realm of radioanatomy report only on the usage of multiple choice questions, due to several challenges related to computer‐based assessment. Nonetheless, we encourage radioanatomy teachers to include a more diverse set of question types. We consolidated the lessons learned during our experience over three academic years of carrying out summative assessments in radioanatomy courses on Moodle. Among others, we discuss technical aspects such as image optimization. Providing a lexicon for standardized answers fosters automatic grading. A student survey supports the idea of using stack visualizations for better image interpretation. We finally underline the importance of collaboration between different stakeholders to ensure a smooth assessment preparation, execution, and analysis. These findings offer valuable insights for improving e‐assessment in radioanatomy and potentially other medical courses.

The growing demand for online education raises the question of which learning resources should be included in online programs to ensure students achieve their desired learning outcomes. By automatically identifying the core concepts in educational materials, teachers can select coherent and relevant resources for their courses. This work explores the use of Large Language Models (LLMs) to identify core concepts in educational resources. We propose three different pipelines for building knowledge graphs from lecture transcripts using LLMs and ontologies such as DBpedia. These knowledge graphs are then utilized to determine the central concepts (nodes) within the educational resources. Results show that LLM-constructed knowledge graphs when guided by ontologies, achieve state-of-the-art performance in core concept identification.

Multiple choice questions are heavily used in medical education assessments, but rely on recognition instead of knowledge recall. However, grading open questions is a time-intensive task for teachers. Automatic short answer grading (ASAG) has tried to fill this gap, and with the recent advent of Large Language Models (LLM), this branch has seen a new momentum. We graded 2288 student answers from 12 undergraduate medical education courses in 3 languages using GPT-4 and Gemini 1.0 Pro. GPT-4 proposed significantly lower grades than the human evaluator, but reached low rates of false positives. The grades of Gemini 1.0 Pro were not significantly different from the teachers’. Both LLMs reached a moderate agreement with human grades, and a high precision for GPT-4 among answers considered fully correct. A consistent grading behavior could be determined for high-quality keys. A weak correlation was found wrt. the length or language of student answers. There is a risk of bias if the LLM knows the human grade a priori. LLM-based ASAG applied to medical education still requires human oversight, but time can be spared on the edge cases, allowing teachers to focus on the middle ones. For Bachelor-level medical education questions, the training knowledge of LLMs seems to be sufficient, fine-tuning is thus not necessary.

Assessment is an essential part of education, both for teachers who assess their students as well as learners who may evaluate themselves. Multiple-choice questions (MCQ) are one of the most popular types of knowledge assessment, e.g., in medical education, as they can be automatically graded and can cover a wide range of learning items. However, the creation of high-quality MCQ items is a time-consuming task. The recent advent of Large Language Models (LLM), such as Generative Pre-trained Transformer (GPT), caused a new momentum for automatic question generation solutions. Still, evaluating generated questions according to the best practices for MCQ item writing is needed to ensure docimological quality. In this article, we propose an analysis of the quality of LLM-generated MCQs. We employ zero-shot approaches in two domains, namely computer science and medicine. In the former, we make use of 3 GPT-based services to generate MCQs. In the latter, we developed a plugin for the Moodle learning management system that generates MCQs based on learning material. We compare the generated MCQs against common multiple-choice item writing guidelines. Among the major challenges, we determined that while LLMs are certainly useful in generating MCQs more efficiently, they sometimes create broad items with ambiguous keys or implausible distractors. Human oversight is also necessary to ensure instructional alignment between generated items and course contents. Finally, we propose solutions for AQG developers.

Dear EditorI read with great interest the article by Siddiqui et al. (2023) on twelve tips for introducing E-Portfolios in health professions education. As a computer scientist and responsible for ...

Technology Enhanced Learning has seen a number of technologies during its history, from instructional video cassettes to smartphones and virtual reality environments. Inevitably, with technologies getting obsolete due to replacement or lack of support, TEL solutions face the issue of becoming non-functional. In addition, limited resources may force faculties to abandon certain solutions instead of recreating them with newer technologies. In this paper, we present a case study of an animal neurophysiology virtual lab that has recently suffered from the technology obsolescence phenomenon, both at software and hardware levels. Most importantly, the end of life of Flash was threatening the continuity of an e-learning course. We analyze the issues they were facing, establish a list of requirements for a possible solution and present the implemented changes. We finally perform a risk analysis of obsolescence for the technologies used in the new version of the virtual lab. ^{1 1} 1978-1-6654-6521-2/22/$31.00 2022 IEEE