When scientists try to understand the patterns of brain activity that generate movements, they can’t simply match the activity in one set of neurons to the action that follows—just like no single piece of a car’s engine corresponds to the speed of the wheels. There’s a lot going on, both under the hood and inside the motor cortex, and seemingly simple movements like reaching for an object are the result of a complex, dynamic pattern of activity that varies over time.
New research from the University of Chicago provides an improved model for understanding how the brain’s movement engine works, explaining more than 90% of activity in the motor cortex during reaching movements. This new model could one day help improve brain computer interfaces (BCI) that control robotic prosthetic limbs.
“For decades, the approach people took was simply trying to relate the instantaneous activity of the brain to the movement of the body or the muscles, but that only explains a small fraction of what's going on in the motor areas of the brain,” said Matt Kaufman, PhD, Assistant Professor of Organismal Biology and Anatomy and the Neuroscience Institute at UChicago and senior author of the study, which was published in Nature Communications.
“This model could explain how the brain solves a problem that has been mysterious for a long time, which is how you go from thinking, ‘Okay, I want to make this movement,’ to determining what the correct control signals are and generating them,” Kaufman said.
Matthew Kaufman, PhD
Assistant Professor of Organismal Biology and Anatomy
Assistant Professor of Neuroscience Institute
Committee on Neurobiology
In 2012, Kaufman was part of a team that introduced the idea that activity in the motor cortex was “dynamical,” following mathematical rules during reaching movements to generate the control signals for the arm. Specifically, when you mapped out the activity of motor cortical neurons, they “rotated”—or oscillated—in a sine wave, up and down in a fixed pattern. This was an essential start, but later research showed that such rotational dynamics explained only about 25% of the activity in the motor cortex during movement, and they couldn’t link the dynamics directly to behavior.
In the new study, Kaufman and David Sabatini, a former undergraduate student in his lab, reanalyzed data from previous experiments with monkeys, which recorded activity in the motor cortex while they were reaching for targets. The researchers discovered that while the previous attempts were on the right track, they had missed a critical aspect of how the system works.
The initial concept of rotational dynamics describes a system like a record player: the record always spins at the same speed in the same orientation. Now, Kaufman and Sabatini found that instead, the system acts like a record player in a bowl: the record always spins at the same speed, but depending on where the player is placed in the bowl, the orientation of the record will vary. Each coordinate inside the bowl represents the activity of a set of neurons. Tilt the record 45 degrees to play the song where the hand reaches up and to the left; turn it 30 degrees to spin the tune for reaching down and to the right. They call this new, geometrically oriented model Location-Dependent Rotations (LDR).
“If you tilt your rotations, that determines how much activity makes it into each coordinate, and therefore into each neuron. So, if your record is tilted one way, you'll get a lot of the 2 Hz sine wave in the neurons that drive the biceps. If you tilt it another way, you might get no activity in those same neurons but drive the deltoid instead. This lets you control how much of each of these sine waves goes to each muscle,” Kaufman said.
