Reimagining Anatomy Education with AI and VR

Virtual Reality (VR) is changing the way we learn, especially in fields like anatomy. This technology allows students to interact with 3D models of the human body in a way that textbooks and lectures cannot. Traditional methods of learning anatomy, which often involve reading, listening to lectures, or dissecting cadavers, have limitations. These include lack of interactivity, high costs, and ethical issues. With VR, students can engage with anatomy in a more hands-on manner.

Despite the benefits of VR, not much research has been done on using virtual assistants powered by artificial intelligence (AI) to help students learn anatomy. This paper introduces a new VR environment where users can interact with a virtual AI assistant that can answer questions about anatomy. This assistant can communicate verbally, making the learning process more engaging.

#Background

Anatomy Education is crucial for medical students. They must understand the structures and functions of the human body to perform medical procedures effectively. Traditionally, students learn through textbooks, lectures, and cadaver dissections. However, these methods can be expensive and may not always provide the interactive experience that students need.

There are different ways to test anatomy knowledge, like written and oral exams. Tools like Anderson's modified Bloom's taxonomy provide a framework for creating questions that vary in complexity, helping students learn better.

In recent years, VR has emerged as a useful tool for anatomy education. It offers a more engaging environment where students can interact with 3D models. With VR, students can learn without the ethical issues tied to using cadavers. It also allows for collaborative learning experiences, which are in line with educational principles.

Unfortunately, many VR systems use fixed scenarios that do not adapt to individual learning needs. This is where Generative AI can significantly improve the learning experience. Unlike conventional virtual assistants that follow rigid scripts, AI can generate more natural and engaging conversations, making the interaction feel more human-like.

#What is Generative AI?

Generative AI refers to technology that can produce text, images, or other media based on user input. In education, chatbots powered by generative AI can answer students' questions in real time, providing more personalized assistance. These AI systems use large amounts of data from various sources, enhancing their ability to deliver accurate and relevant information.

In this study, we developed an immersive VR environment that includes a generative AI virtual assistant designed to help students learn anatomy. The assistant can respond to different types of questions and engage in verbal communication.

#Research Goals

The study aims to achieve three main goals:

1. Assess User Performance when interacting with the generative AI assistant in various configurations.
2. Evaluate Subjective Experiences of participants, such as usability and mental workload.
3. Identify potential benefits and limitations of using generative AI in anatomy education.

#Methodology

#Participants

A total of 16 participants were selected for the study. They were recruited from the University of Delaware and had to meet certain criteria, such as being 18 years old and having normal or corrected vision. The participants underwent a process to familiarize themselves with the VR equipment before the study began.

#VR Environment

The VR environment was created using the Unity game engine. It includes interactive 3D models of human anatomy. Participants could grasp, resize, and rotate the models during the training session. The VR setup used a Valve Index headset and controllers, allowing users to move around the virtual space easily.

#AI Integration

The generative AI service, ChatGPT, was integrated into the VR environment to serve as the virtual assistant. This assistant could answer questions and engage with participants in conversation. The assistant was designed to provide responses that were animated through lip synchronization and facial expressions, making the interaction more realistic.

#Study Design

The study was conducted using a within-subject design, where each participant experienced both the avatar and screen-based configurations of the virtual assistant. The questions posed to participants varied in cognitive complexity-some were straightforward knowledge-based questions, while others required more in-depth analysis.

#Data Collection

Data was collected on several aspects:

• Task completion time: The time it took participants to answer questions.
• Number of interactions: How many times participants asked the virtual assistant for help.
• Scores: Whether the answers provided were correct or incorrect.

Participants also completed questionnaires to evaluate their feelings about usability, task load, and sense of presence within the virtual environment.

#Results

#User Performance

The study primarily focused on how well participants performed while interacting with the generative AI assistant.

• Task Completion Time: Both configurations of the virtual assistant had similar completion times. However, participants tended to respond faster to knowledge-based questions when using the avatar configuration.
• Number of Interactions: Participants had more interactions with the virtual assistant in the avatar configuration, especially when answering analysis-based questions.
• Scores: Participants scored higher on knowledge-based questions when using the avatar configuration compared to the screen configuration.

#Subjective Experiences

The participants provided feedback on their experiences:

• Usability: Participants generally found both configurations usable, with SUS scores indicating that both were effective for learning.
• Task Load: Responses indicated that participants felt mental demand was higher than physical demand when using the virtual assistant.
• Sense of Presence: Participants reported feeling a sense of presence in the virtual environment, with some aspects scoring higher than others.

#Feedback

In interviews, participants expressed preferences for the avatar due to its immersive qualities but noted some issues with voice synchronization. They highlighted a need for better planning of questions to ensure smooth interactions.

#Discussion

The results indicate that VR with a generative AI assistant can enhance anatomy education by providing a more interactive and engaging learning experience. Participants showed a preference for the avatar configuration for its ability to create a more lifelike interaction.

While both configurations had their advantages, the study hints at the potential for a combined approach that utilizes the strengths of each. The avatar's presence could be beneficial for complex questions, while the screen could aid in simpler tasks.

#Limitations

The study was limited by its small sample size of 16 participants. This restricts the generalizability of the findings. There were also variations in how participants phrased their questions, affecting the responses they received.

Furthermore, some participants faced challenges with the speech recognition technology, particularly those with accents or who spoke quickly. These issues may have impacted their ability to receive correct answers.

#Future Research

Based on the findings, future research should consider:

1. Larger Sample Size: Increasing the number of participants will help validate the results.
2. Improved AI Responses: Further work is needed to ensure that the AI can handle complex analysis questions effectively.
3. User Experience: Understanding the varying levels of participant experience with VR and AI will help in designing future studies.
4. Expanded Features: Incorporating more advanced features, such as emotion recognition and gestural feedback, could improve engagement.

#Conclusion

This study highlights the potential of using generative AI in a VR environment to enhance anatomy education. By combining the immersive qualities of VR with intelligent virtual assistants, students can enjoy a more interactive and personalized learning experience. As technology continues to evolve, it is essential to explore its benefits and limitations further, ensuring that educational methods keep pace with advancements in AI and VR.