Adaptive Electrical Muscle Stimulation Improves Muscle Memory

Research Background and Problem

• Identified Issues and Challenges: The authors highlight the shortcomings of current electrical stimulation-based systems in teaching motor skills. These systems typically employ static guidance strategies, meaning they consistently apply guidance regardless of the user's learning progress, failing to adapt to the user's learning curve. This approach does not dynamically adjust to specific user errors, even though dynamic feedback is considered crucial for skill acquisition.

• Significance: Motor skill learning requires feedback to ensure behavioral accuracy. While static feedback is effective in the early stages, its efficiency and user experience may decline as skills become more advanced. Additionally, many modern force-feedback devices (e.g., exoskeletons or robotic arms) are bulky and challenging to integrate into daily use scenarios, whereas electrical stimulation devices offer advantages in wearability and miniaturization.

• Research Motivation and Related Work: Although prior studies have validated the effectiveness of EMS systems in motor training, most of these studies employed static guidance methods without dynamically adapting to user needs. This research aims to address this gap by examining whether dynamically adjusted EMS guidance strategies can enhance learning outcomes, particularly in the formation of motor memory.

Solution

• Proposed Method or Solution: The authors designed an "adaptive electrical stimulation system (adaptive-EMS)" that dynamically adjusts guidance strategies based on the user's error rate during the learning process. The system categorizes guidance into three stages based on the user's error rate:

  1. Demonstration Stage: For users with a high error rate (error rate > 50%), the system directly demonstrates the complete action sequence.
  2. Correction Stage: When the error rate falls between 50% and 25%, the system corrects actions only when the user makes a mistake, preventing errors and enabling the user to perform the correct action.
  3. Warning Stage: For users with an error rate below 25%, the system provides warnings when mistakes occur but does not correct the actions, encouraging independent learning.

• Innovations:

  • Introduced a dynamic feedback adjustment mechanism for the first time, building on traditional static-EMS.
  • Proposed a stepwise intervention reduction strategy, transitioning from high to low intervention to help users achieve self-directed learning.
  • Designed the strategy based on a learning curve model, adjusting guidance intensity in real-time according to the error rate.

• Implementation Steps:

  1. Develop an algorithm to automatically adjust learning stages.
  2. Use custom electrical stimulation hardware and interfaces to control electrical stimulation for users' finger movements.
  3. Conduct comparative experiments using two learning strategies (static-EMS and adaptive-EMS).

Research Findings

• Specific Results Achieved:

  1. Users' final recall performance ("muscle memory") under the adaptive-EMS condition was significantly better than under static-EMS.
  2. The adaptive-EMS system was perceived as more effective by the majority of participants (approximately 90%) and was more engaging than the static strategy.
  3. The learning time and number of sessions required for both strategies were similar, indicating that the improved learning outcomes of the adaptive strategy were not solely due to extended training time.

• Advantages:

  • Learning Effectiveness: In subsequent unaided recall tests, the error rate under adaptive-EMS was significantly lower than that under static-EMS.
  • User Experience: Participants reported greater engagement in learning and stronger motivation when correcting errors due to the interactive nature of the system.

• Experimental or Evaluation Results:

  • Accuracy Improvement: The average error rate decreased from 6.2 under static-EMS to 3.67 under adaptive-EMS.
  • Subjective Feedback: Participants generally felt that the dynamically adjusted guidance strategy better facilitated independent error correction and learning.

• Limitations and Future Directions:

  • Some participants expressed a desire to manually adjust the guidance stages, suggesting that the pre-set stage transition mechanism may not fully suit all users.
  • Certain participants felt that the corrective feedback lacked sufficient contextual information, potentially causing initial confusion.
  • The effectiveness and limitations of EMS in specific tasks (e.g., rhythm-based tasks) remain underexplored, warranting further investigation in future studies.

Conclusion

By employing the adaptive-EMS system, this study demonstrates how dynamically adjusted guidance strategies can effectively enhance motor memory learning outcomes and improve the user learning experience. The findings provide guidance for designing more intelligent physical teaching systems in the future and suggest that customizing learning stages based on user needs and enhancing the contextual information of feedback could further optimize system performance.