Researchers examine complex neuron activity in movement planning

Nov. 16, 2010, 2:01 a.m.

“Imagine you’re a golfer, set up to tee, getting ready to swing,” said electrical engineering researcher Mark Churchland. “Something goes on inside your brain. Getting ready to hit the ball, to be motionless, means something.”

Despite the observed motionless nature of your body, your brain is full of movement. Neurons are firing, taking you from a baseline state to a starting state. Now, a Nov. 4 publication by a group of Stanford researchers in “Neuron” has explored the activity of cortical neurons responsible for movement planning and execution.

Krishna Shenoy, associate professor of engineering, was the paper’s senior author, while Churchland was the lead author. A graduate student, Matthew Kaufman, and two other researchers, John Cunningham, and Stephen Ryu, are also authors of the paper.

The previously held belief postulated that neurons would “vote” for movements they preferred, according to Churchland. If you wanted to reach right, these right-reaching neurons would be very active during the physical rightward reach.

“Then what does it mean to be getting ready to move right?” Churchland said. “Maybe the right [reaching neurons] are a little bit active. Maybe they vote just a little bit.”

This theory relied on a correlation between neural activity before and during movement for a group of neurons assuming a preferential movement.

The study found that a neuron’s “preferred movement” could differ during preparation and movement. Neurons could be very active in a rightward movement but not active before. Conversely, a neuron could be very active in the planning and inactive in the movement.

The study drew on data the researchers collected seven years ago. Churchland said that at the time, examining the data for neural planning analysis was confusing.

“It’d be like an astronomer collecting data on the position of planets and stars in the sky before he knew they were planets,” he said. Just as an astronomer collecting the data is unaware of the “underlying pressure” — the planets moving around the sun — so were the researchers unaware of the mechanisms of neural activity during planning.

“We try to collect really good rich data sets such that when our thinking improves, we have good data sets to analyze,” Kaufman said.

The team originally conducted research in this manner, recording the activity of a monkey’s individual cortical neurons while preparing to “swat a fly” in a video game.

When the team returned to the data sets with some idea of what they were looking for, the “underlying pressures” became more apparent.

The paper concluded that while preparatory and movement activities are linked, cortical neural activity is mechanistic rather than representational. This means that the neural activity in planning works like a pendulum schematic as opposed to preferential voting or representation, as previously believed.

To swing a pendulum rightward, an initial leftward push is necessary. Just as the pendulum maintains a consistent dynamic movement despite shifting directions and times, the cortical neurons direct a specific movement despite the seemingly shifting neural activity, Churchland said.

“It’s hard not to be interested in the brain when we know so little about it,” Churchland said. “The two obvious places to start are where information enters the brain through the visual cortex and where it leaves the brain in the motor cortex.”

“Understanding how the brain creates movement can allow us to move one stage earlier,” Churchland added. “How do you decide to move?”

Churchland explained that research in cortical neural planning also means progress for the field of neural prosthetics. Neural prosthetics, currently being developed in monkeys, would be able to link functioning muscles and a motor cortex, bypassing an impaired spinal cord.

Login or create an account

JOIN THE STANFORD DAILY

application deadline
Friday, Oct. 11

Days
Hours
Minutes
Seconds