AI-Generated Mnemonic Images Improve Long-Term Retention of Coronary Artery Occlusions in STEMI: A Comparative Study (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Medical students often face challenges in retaining complex information about coronary artery occlusions when using traditional ECG diagrams. Visual mnemonics may enhance learning by making information more memorable and easier to recall. </sec> <sec> <title>OBJECTIVE</title> This study aimed to evaluate the effectiveness of artificial intelligence (AI)-generated mnemonic-based images overlayed on a 12-lead ECG in improving long-term retention and student preference compared to traditional ECG diagrams. </sec> <sec> <title>METHODS</title> Mnemonic images were created using DALLE-3 and refined for educational use. A comparative study was conducted with 275 first-year medical students across six campuses. Forty students in the experimental group (n=40) received lectures and mnemonic-based images, while 235 students in the control group received traditional ECG diagrams only (n=235). Exam performance was assessed using a weekly and final course exam, and student engagement was measured using the Situational Interest Survey for Multimedia (SIS-M) with a 78% response rate (n=31). </sec> <sec> <title>RESULTS</title> All control group sites showed a significant drop in successive exam scores which was not observed in the experimental group suggesting better long-term retention. Furthermore, experimental group students showed higher preference of the mnemonic images, as shown by high average scores for triggered situational interest (M = 4.685, SD = 0.432, p&lt;.001), maintained interest (M = 4.60, SD = 0.48, p&lt;.001), maintained-feeling interest (M = 4.459, SD = 0.616, p&lt;.001), and maintained-value interest (M = 4.742, SD = 0.472, p=.006). Additionally, 80% of students in the experimental group preferred the mnemonic-based images over traditional ECG diagrams, indicating enhanced engagement and preference. Qualitative feedback highlighted increased impact on retention, engagement, and preference for the mnemonic images while also showing challenges related to difficulty understanding the images and timing of image presentation. </sec> <sec> <title>CONCLUSIONS</title> The use of AI-generated mnemonic images improved student engagement, retention, and interest by leveraging the strengths of visual memory aids. These findings suggest that generative AI can be a valuable tool in creating effective mnemonic-based learning resources, with the potential to enhance outcomes in medical education. </sec>

Ähnliche Arbeiten