Finding the "Point of No Return"
When is a decision truly committed? This hardware-optimized closed-loop BCI system investigates the critical threshold between decision initiation and irreversible commitment.
Closed-Loop System Architecture
The experiment runs on a dual-PC setup for optimal performance. The task PC presents the random dot motion stimulus and handles participant responses, while the detector PC continuously monitors EMG signals for movement onset. When EMG onset is detected, a trigger is sent via hardware-synchronized connections to modify the visual stimulus with sub-millisecond precision. This closed-loop architecture enables us to probe exactly when actions become irreversible.
Research Question
When does a motor action become irreversible after initial motor preparation begins? Our experiment determines the exact timing of the "point of no return" in decision-making.
Methodology
Participants perform a dot-motion task while EMG activity is monitored. When muscle activity is detected, visual evidence changes after variable delays to test if actions can still be cancelled.
Technical Innovation
Hardware-synchronized parallel port timing ensures sub-millisecond precision, allowing us to pinpoint exactly when actions become committed.
A Taste of Results in Action
Real-time EMG detection algorithm performance from our pilot study, demonstrating the precision necessary to capture the moment of decision commitment.
Channel 33: EMG with Response Correctness & Button Delay Analysis. The system successfully detects the EMG onset signal (red vertical line) approximately 397ms before the physical button press.
Detection Latency
Ultra-fast algorithm detection time from actual EMG onset to detection trigger
Signal Transmission
Additional delay for transmission via LSL stream to the task PC
Total system latency: ~5ms
EMG-to-Button Window
Time between initial muscle activity and completed button press
Typical window for this type of perceptual task
Real-Time Intervention Capabilities
With our ~5ms end-to-end detection resolution, we can precisely modify the task in real-time as decisions unfold. The EMG detection runs on a dedicated PC and triggers stimulus changes through hardware-synchronized channels, allowing us to intervene at exact millisecond timings during the ~397ms window between initial muscle activity and completed button press. This unprecedented timing precision enables us to identify exactly when a decision crosses the "point of no return."