Jesse H. Goldberg, Neurobiology and Behavior, remembers when he first had the idea to turn songbirds into a model system for medical research. He read about a breakthrough in brain research conducted at the Massachusetts Institute of Technology. Through studying how baby songbirds learn to sing, researchers there had identified an area of the brain—part of the basal ganglia thalamo cortical loop—that generates the exploration necessary for learning. This variability generator is controlled by dopamine and norepinephrine, chemicals released in the brain that act as neurotransmitters, shuttling data between neurons. “That was astonishing,” Goldberg says. “I was compelled by the idea of studying in the songbird a variability generator that’s part of the same brain circuit implicated in many diseases in humans.”
Goldberg was interested in diseases that can affect how people control variability in their thoughts or actions. Schizophrenia, Huntington’s disease, and dystonia, all present symptoms of disorganized thinking and uncoordinated movements that could result from excess variability. Then, there are diseases that could spring from less variability such as obsessive-compulsive disorders, which cause patients to have the same thought over and over and to repetitively engage in the same behavior, such as washing their hands.
These types of diseases had always been seen as causing a loss of function in patients—something was missing in the way their brains worked. “But I thought, rather than thinking of these diseases as causing a deficit,” Goldberg says, “maybe what’s going on is that there’s a variability generator that can either be cranked up and not controlled, or underactive.”
Songbirds and Humans
Goldberg set out to learn more about the variability generator and the way songbirds learn to sing. In one project, his lab studies young adult zebra finches who are still perfecting their songs. The birds are similar to a human who is learning how to play a sport or a musical instrument. Trial and error are crucial for both the bird and the human if they are going to master the behavior they intend to learn. “Variability is one side of the coin, the trial side,” Goldberg says. “I had already made progress on understanding variability by identifying which precise parts of the basal ganglia pathway were necessary for babbling. The next big mystery in the field is the error side: What happens after the bird produces the notes? How does the brain evaluate whether they are the right ones? It has to relate to what the bird intended to do. It has to match his personal internal goal.”
Goldberg turned to his lab’s postdoc Vikram Gadagkar, PhD’13 Physics, to build a system for monitoring the finches as they sing in real time. Using that system, the researchers are able to fool a bird into thinking he has made a mistake by routing the song through speakers with a millisecond delay, occasionally swapping out a correct note for a different note the bird sang earlier so that he experiences a worse-than-expected outcome. At the same time, they record the activity of dopamine neurons in the finch’s midbrain.
Previous studies in mammals had shown that dopamine neurons are activated by positive reward outcomes such as food or juice, but it was unknown if dopamine activity also assesses internally evaluated behaviors. The discharge of dopamine when the birds hear their song without an error compared to when they are fooled into thinking they made an error, shows that their brains are monitoring the quality of the behavior against an internal goal. When they hit the right note, their dopamine neurons are activated in the exact way the same neurons are activated in a thirsty rat or monkey the moment it gets a juice reward.
“That’s a game changer,” Goldberg says. “I think the fact that this dopamine system can evaluate behavior relative to internal goals in the same way external goals are evaluated means the system is always on.”
Goldberg sees this as important to our motor system, which constantly makes little adjustments in how we move. If we pick up a mug, for example, we might grasp the handle wrong, then adjust our fingers without consciously thinking about it. “You have performance at every instant of your life that’s better or worse than you expected,” he says. “If your dopamine system is constantly evaluating all aspects of your motor behavior, that does a good job of explaining the motor problems of Parkinson’s because dopamine neurons are the ones that die when you have that disease.”
If our dopamine system is constantly evaluating our behavior, that can also explain something like trichotillomania, an obsessive-compulsive disorder where sufferers constantly pull out their hair. “I would predict if you recorded dopamine neurons in those patients, every time they pull out hair there’s a big burst of dopamine,” Goldberg says. “That’s the problem. It’s like they’re addicted to pulling out their hair.”
“I was interested in psychiatry, but I thought, how are you going to fix someone’s dysfunctional thought pattern if you can’t fix a tremor?
Where Is the Variability Generator in the Mammalian Brain?
While the location of the variability generator in bird brains has been identified, no one knows yet where it is in mammals, because mammalian brains are organized in a more complicated fashion than bird brains. Goldberg, however, is hoping to find it in mice. To do that, he turned to one of his graduate students, Tejapratap Bollu, MEng’12 Biomedical Engineering, to build a device that can track the exact manner in which mice move their limbs. Bollu created a precise monitoring system, featuring a special joystick that the mice must manipulate in order to get food pellets. The system helped the Goldberg lab win a New Innovator Award worth $2.4 million over five years from the National Institutes of Health. “We’re hoping to identify pathways for motor variability in the mammalian brain,” Goldberg says. “We’re looking at the cortex so far. We’ve found parts of the brain related to timing and velocity, but we haven’t found the variability generator yet. If we can find where the variability comes from in mice, we can translate it more easily into specific parts of the brain to target in humans.”
Goldberg has a medical degree, and originally he thought he would follow in his father’s footsteps and become a psychiatrist. “I was interested in psychiatry, but I thought, how are you going to fix someone’s dysfunctional thought pattern if you can’t fix a tremor? There’s a lot of evidence that the principles of motor control will extend into cognition. If we can identify the sources of variability in movement, they are likely to have the same origins in the brain as the ones that control cognition. I think there’s no such thing as a psychiatric disease or a neurological disease. There’s no differentiation. They’re all neuropsychiatric.”