Revital Medenitzki, Ilana Dubovi E73 This study used the MediaPipe framework to collect multimodal movement data, focusing on head, hand, and torso motion (Quiñonez et al., 2022). Pose landmarks consisting of 33 skeletal points, including the shoulders, elbows, wrists, hips, knees, and ankles, were combined with selected FaceMesh landmarks representing the nose, chin, eyes, and mouth to estimate orientation and movement. Findings On average, students spent approximately 226 seconds (SD = 112) engaged in the procedural learning phase of the VR simulation. During this phase, they acquired and practiced procedural knowledge by performing tasks such as preparing medications for administration (e.g., selecting the correct drug and calculating the dosage), administering the medication to a virtual patient, and monitoring treatment effectiveness and potential side effects. Figure 1 illustrates an example of one student’s bodily engagement in the VR environment, showing sequences of hand movements, head turns, and torso rotations during a two-minute learning segment. Figure 1. Snapshots of student bodily movements during VR-based learning, captured with MediaPipe landmarks. Descriptive statistics (Table 1) indicate that approximately 53–56% of segment time was spent on hand and head movements, whereas about 35% was spent on torso rotations. In addition to bodily engagement, we also assessed indicators of cognitive engagement: blink rate (M = 12.8, SD = 7) and EDA rate (M = 27.8, SD = 15). Table 1. Bodily motion during learning with VR as measured by MediaPipe framework (n = 37) Count Total seconds of active motion Time percentage % Hand movements 1207 ± 712 121 ± 64 53% ± 7 Head turns 186 ± 104 125 ± 78 56% ± 12 Torso rotations 151 ± 87 77 ± 46 35% ± 10 15° torso turns 92 ± 54 - - 90° torso turns 1.3 ± 1.6 - -
RkJQdWJsaXNoZXIy Mjk0MjAwOQ==