Neurochemistry Explains the High of Getting It Right


Birds sing in tune thanks to neurochemical feedback, according to a paper published today in Science. Zebra finches, in particular, feature built in circuitry for delivering dopaminergic rewards for getting notes right—there exists a hardwiring for matching sensory feedback with expected, optimal outcomes. Singing well makes birds feel good, and this has some interesting implications for our general understanding of neural reward systems. We are our own worst critics, neurologically speaking.

The brain wants simple things: food, social interaction, sex. Tick one of these boxes, and an individual can be expected to receive a neurochemical reward. It’s built in.

These are primary rewards; they correspond to things necessary for survival and-or reproduction. The brain has good reason to come equipped with dedicated positive feedback circuitry for consuming calories and making babies. But there are other reward structures for things like money, which we don’t come with by default and must instead learn through experience and conditioning. These are extrinsic rewards.

Other rewards are less well understood. For example, there are what are known as performance rewards. One’s brain has an interest in providing rewards for “getting it right,” even if the consequences of getting it right (or wrong) aren’t so clear in terms of raw nature (food, sex). Performance rewards are what keep birds in tune.

We’ve known for several decades that dopamine is part of the primary reward systems in the brain. Get some good food, and the brain delivers a dopamine pat on the back. Take a thirsty monkey, give it some juice, and you’ll see characteristic patterns in the part of the brain associated with dopamine rewards. Condition the monkey to expect some juice in the presence of a bell (a la Pavlov’s dog), but withhold the juice, and the result is a suppression of dopamine in this part of the brain.

“Surprisingly good outcomes cause dopamine activation and reinforcement of behavior and disappointingly bad outcomes cause suppression of dopamine activity and extinguishing of behavior,” Jesse Goldberg, a neurobiologist at Cornell University and co-author of the new study, explains. “And at the extremes of this if you have addiction. If you take a drug that activates dopamine receptors, it’s basically hijacking your brain’s way of saying, that was great, do that again. And at the other extreme, if you lose dopamine neurons, that results in Parkinson’s disease which is basically when animals don’t even move anymore. Dopamine suppression is like saying, don’t do that again, and when you lose all your dopamine, you don’t do anything anymore.”

Image: Goldberg et al

The question then behind the new study is how this system of reinforcement translates to relatively arbitrary behaviors, such as those related to performance.

As it turns out, songbirds make for an ideal test case because so much of their lives revolve around the relatively arbitrary behavior of singing.

“Zebra finches when they’re very young hear a tutor song, usually from their father,” Goldberg explains. “And then they memorize it. And then you can take a bird and put it alone in a cage and it will just spend the next three months of its life babbling and singing. Eventually they wind up making an imitation of that song from earlier in their life. They basically work through a gradual process of trial and error.”

To see what dopamine has to do with said process, the researchers needed some way of tricking the birds into thinking that they had sung a wrong note. They were able to do this using very fast software that could take a birdsong as input, distort it, and then play it back to the bird in almost real time. It was sort of like the birds were unknowingly singing through an effects pedal, but only at certain random intervals.

The brain signals they observed in the birds almost perfectly matched those seen in mammals during reinforcement learning (bells and juice). Bird brains track sensory input and dole out chemical rewards based on song quality.

In other words, the researchers’ hypothesis was correct.

“What was crazy was that the activation looks exactly like the activation you would see in a thirsty monkey at the moment it gets juice,” Goldberg says. “And the suppression that we saw was basically exactly like the suppression in a monkey that was expecting juice at the moment he realized he didn’t get it.”

“The animal is just trying to match an internal goal,” he explains. “And so I think it opens up the possibility that when you or I are practicing a sport or learning an instrument or when a baby is learning to speak, maybe it’s the case that dopamine neurons in our brains also do this just like the songbirds.”

Get six of our favorite Motherboard stories every day by signing up for our newsletter.

Powered by WPeMatico