New research by Postdoc, Ashesh Dhawale and Bence Ölvecsky suggests errors resulting from variability in motor function are a feature, not a bug, of our nervous system and play a critical role in learning. The study published in Current Biology addresses the issue of how the brain regulates variability which is necessary for learning, but not useful when a successful action needs repeating.
Dhawale, Ölvecsky and team looked at data from approximately 3 million rat trials and found that rats regulate their motor variability based on the outcomes of the most recent 10 to 15 attempts at a task. Depending on the outcome of previous trials, rats will either increase their variability if trial has gone poorly, or limit their variability if trial has gone well. They also studied performance-dependent variability by training rats to press a 2D joystick towards a target angle. The researchers discovered rats that were regularly rewarded had low variability and if they performed less well in later trials their variability increased. The Harvard Gazette
Image: Still image from Joystick Angle Task Trial video on YouTube.com courtesy of Ölvecsky and Dhawale