Hermann Thomas, Grautoff Steffen, Tielking Friederike, Persson Jan, Diebner Hans H, Tiesmeier Jens
Ambient Intelligence Group, Faculty of Technology, Bielefeld University, 33619 Bielefeld, Germany.
Emergency Medical Services, District of Herford, Amtshausstr. 3, 32051 Herford, Germany.
Sensors (Basel). 2025 Jul 12;25(14):4373. doi: 10.3390/s25144373.
We introduce a novel technique for the sonification/auditory representation of a 12-lead electrocardiogram (ECG), the standard diagnostic method for the detection of ST elevation myocardial infarction (STEMI). Our approach to ST elevation sonification conveys the detailed variation of the ST segment to enable differentiated, correct interpretation and severity without consulting a visual display. We present a variety of novel sonification designs and discuss their benefits and limitations. As part of an emergency training program, a cohort of 44 medical students (5th academic year) participated in a classification study in which the diagnostic accuracy of the participants was determined with regard to audibly presented ECG sequences of different STEMI severity levels. Regarding the classification of sonified ECG sequences, the discrimination of isoelectricity (IE, the healthy class) from all other (STEMI) classes combined yielded a perfect classification of all 660 classification instances (sensitivity = specificity = 1). With respect to the individual classification of all five classes (IE, inferior/anterior, and moderate/severe STEMI), an overall accuracy of 0.82 (0.79, 0.85) and an intraclass coefficient of κ=0.77 were estimated.
我们介绍了一种用于12导联心电图(ECG)的声化/听觉呈现的新技术,这是检测ST段抬高型心肌梗死(STEMI)的标准诊断方法。我们对ST段抬高进行声化的方法传达了ST段的详细变化,以便在不参考视觉显示的情况下进行差异化、正确的解读和判断严重程度。我们展示了各种新颖的声化设计,并讨论了它们的优点和局限性。作为应急培训计划的一部分,一组44名医学生(第5学年)参与了一项分类研究,在该研究中,根据不同STEMI严重程度水平的可听呈现的ECG序列来确定参与者的诊断准确性。关于声化ECG序列的分类,将等电位(IE,健康类别)与所有其他(STEMI)类别组合进行区分,对所有660个分类实例实现了完美分类(敏感性 = 特异性 = 1)。对于所有五个类别(IE、下壁/前壁以及中度/重度STEMI)的个体分类,估计总体准确率为0.82(0.79,0.85),组内相关系数κ = 0.77。