Myo Action: Accelerating Voluntary Actions via Electromyography and Muscle Stimulation

Paper Title

Myo Action: Accelerating Voluntary Actions via Electromyography and Muscle Stimulation

Publication Info

• Topic area: Human-computer interaction and haptic assistance for reaction-time acceleration.
• Keywords: Electromyography (EMG), electrical muscle stimulation (EMS), reaction-time acceleration, sense of agency, haptic assistance, decision-based actions, user intention, human-computer interaction, motor control, wearable devices.

Background and Problem

• Problem / challenge: Existing EMS-based systems for accelerating reaction-time often override user intention, leading to a loss of agency. These systems also require extensive timing calibration and fail in decision-involving tasks (e.g., go/no-go trials).
• Significance: Reaction-time acceleration has critical applications in sports, gaming, safety, and training. Preserving user agency while achieving acceleration is essential for usability, ethics, and user acceptance.
• Motivation and related work: Prior EMS systems relied on preemptive stimulation, which diminished agency, especially in decision-based tasks. While EMG has been used for movement detection, no system has successfully combined EMG and EMS to accelerate reaction-time while preserving agency.

Solution

• Proposed approach: Myo-Action, a system that uses EMG to detect the onset of muscle activation and triggers EMS to accelerate voluntary movements while preserving user agency.
• Novelty:
  1. Combines EMG and EMS to achieve reaction-time acceleration without overriding user intention.
  2. Operates with ultra-low latency (∼ 290 μs) to ensure timely intervention.
  3. Demonstrates preserved agency in decision-based tasks (e.g., go/no-go trials).
  4. Eliminates the need for pre-calibration of stimulation timing.
• Procedure and key techniques:
  • EMG detects early neural activation in a target muscle before movement occurs.
  • EMS induces faster muscle contraction compared to voluntary control.
  • A low-latency hardware system processes EMG signals and triggers EMS within ∼ 290 μs.
  • The system includes analog filtering, optical switching, and adaptive signal processing to ensure reliable and precise operation.

Results

• Concrete findings:
  1. Myo-Action reduced reaction-time by ∼ 23 ms compared to voluntary action (p < 0.0001, Cohen’s d = 0.85).
  2. Preserved agency in decision-based tasks, with significantly higher agency scores for both go (5.3 vs. 3.5) and no-go (6.1 vs. 1.5) trials compared to baseline EMS.
  3. Detection accuracy of EMG-triggered EMS was high (F1 = 0.95).
• Advantage over baselines:
  • Reaction-time acceleration without pre-calibration, unlike baseline EMS systems.
  • Significantly higher agency preservation in decision-based tasks compared to preemptive EMS.
• Experiments / evaluation:
  • Participants: 12 individuals (7 male, 4 female, 1 non-binary; average age = 23.9 years).
  • Tasks: No-decision reaction-time, decision-based reaction-time (go/no-go), and exploratory applications (pen-catching and video-gaming).
  • Metrics: Reaction-time, agency scores (7-point Likert scale), and detection accuracy.
• Limitations and future work:
  • Current implementation supports only one muscle and lacks movement classification.
  • Limited generalizability due to a small, homogeneous participant sample.
  • System performance in dynamic environments and multi-muscle interactions remains unexplored.
  • Future work includes advanced EMG processing, multi-channel support, and evaluations in more dynamic scenarios.

Summary

Myo-Action introduces a novel method for accelerating reaction-time using EMG-triggered EMS while preserving user agency. The system achieved a ∼ 23 ms reduction in reaction-time and demonstrated significantly higher agency in decision-based tasks compared to baseline EMS. By eliminating the need for pre-calibration and ensuring low-latency operation, Myo-Action offers a promising direction for haptic assistance in applications such as sports, gaming, and safety. Future work will address multi-muscle interactions, dynamic environments, and broader user evaluations.