Enhancing Neuroscience Education: Using AI-Generated Quizzes to Support Student Learning
Aidan Schneider is a WashU graduate student in the Division of Biology & Biomedical Sciences. He recently completed the Scholar level of the Professional Development in Teaching Program, which required him to design and implement a Scholarship of Teaching and Learning (SoTL) research project. Aidan describes his project below.
As an aspiring researcher in Computational & Systems Neuroscience, I aimed to merge my interests in artificial intelligence with my commitment to teaching future scientists. I found an opportunity in an introductory neuroscience course for biology and psychology undergraduates. In this course, student concept proficiency is assessed very frequently—with a quiz every lecture. To keep up, students sought more practice materials. However, for students, creating them could be time-consuming and potentially error-prone. Therefore, I used ChatGPT to generate optional practice quizzes that accurately reflected the lecture content.
I developed a pipeline using YouTube and ChatGPT APIs to efficiently generate quizzes. Transcripts from previous lectures provided the reference for multiple-choice questions, each answer accompanied by a concise explanation. The instructors and I reviewed these for clarity and accuracy before sharing them with students. To evaluate how students took to these practice quizzes, I tracked student practice quiz responses and their grades across the semester. To understand the students’ views on these practice quizzes and AI education materials more broadly, I also surveyed them at the end of the semester.
Students quickly embraced the practice quizzes—over 80% took at least one, and on any given week, nearly half of the class would use them. High-performing students took practice quizzes more frequently. However, survey responses suggested that many factors contributed to engagement with a particular practice quiz, including curriculum difficulty, current performance, and available time. Overall, students found the practice quizzes to be a helpful, time-efficient study tool. They recommended integrating content from other parts of the class into questions, including pre-readings and discussion sections. More broadly, while opinions varied, most felt that AI-generated study materials were lacking in other courses—and many expressed a willingness to create such materials themselves if needed.
Developing and carrying out this study in collaboration with both biology instructors and scholars of teaching and learning was very enriching. Working with coursemaster Elise Walck-Shannon and instructors Ilia Katritch and Marta Celorrio Navarro deepened my appreciation for what it takes to create course materials that are responsive to student needs using cutting-edge technology. Successful integration of artificial intelligence into teaching requires a balance between automation and human oversight. However, when done correctly, even simple AI-generated materials can significantly support student learning. Through the SOTL seminar and collaboration with Lisa Kuehne, Rick Moore, and Denise Leonard at the CTL, I also learned how a scientific approach can be used to develop effective teaching strategies. This experience exposed me to the expansive field of education research and deepened my commitment to innovative teaching. Most importantly, it revealed how my combined expertise in biology and computer science not only fuels my work as a neuroscientist but also empowers me to be an effective educator in our rapidly evolving educational landscape.
