Self-organized critical noise amplification in human closed loop control.

Area of Science:

• Neuroscience
• Control Theory
• Human Motor Control

Background:

• Human closed-loop control tasks, such as standing or balancing, show non-Gaussian fluctuations with long-tailed distributions.
• The underlying cause of these fluctuations in human behavior remains unknown.

Purpose of the Study:

• Investigate if self-organized critical noise amplification, arising from adaptive controllers with finite memory stabilizing unstable dynamics, causes these non-Gaussian fluctuations.
• Formulate and test a realistic model of adaptive closed-loop control incorporating memory constraints and delays.

Main Methods:

• Developed a theoretical model of adaptive closed-loop control with memory and delay constraints.
• Conducted psychophysical experiments where human participants balanced an unstable target on a screen.
• Compared model predictions with experimental data on human control dynamics.

Main Results:

• The model successfully reproduced the long-tailed distributions observed in human behavior.
• The model also replicated other characteristic features of human control dynamics.
• Fine-tuning the model allowed for the identification of subject-specific control system parameters.

Conclusions:

• Self-organized critical noise amplification in adaptive control systems with finite memory may explain non-Gaussian fluctuations in human motor control.
• The human nervous system appears to estimate system parameters online with high efficiency, using short observation periods.