Richard Mooney is the director of graduate studies for the Program in Neurobiology and has been running a lab at Duke since 1994. Mooney, the George Barth Geller Professor of Research and Neurobiology, is dedicated to understanding the biological basis of learning by studying the learning processes of songbirds. The Chronicle’s Julian Spector spoke with Mooney about his research.
The Chronicle: What is the long-term research goal of your lab?
Richard Mooney: The primary focus of the lab is on learning, memory and communication—the intersection between those three things. Where they come together in part is through the studies of songbirds, which learn to sing. They use their songs to communicate to one another, they produce and perceive these really complex vocalizations—sort of analogous to the way that we produce and perceive speech sounds—and, unlike most other animals, but like us, they learn how to do it; they learn how to sing. They rely on the auditory experience of an appropriate model, much like we learn to speak in part by emulating our parents.
TC: How did you decide on songbirds as the ideal test subject for this research?
RM: I’m not a bird fancier necessarily, it’s just a really powerful system to study learning in. They teach one generation to the next; they teach song. In a relatively short time, in a few months one can witness the whole learning process from memorizing the model to emulating that model accurately, copying that model accurately.
TC: Generally speaking, how do these birds learn to sing?
RM: What they sing is a product of the environment they grow up in, the songs they heard when they were young. Those memories of the songs are translated into their own songs through a process of vocal practice and rehearsal. To draw a broad analogy, its like the transformation that an infant undergoes—as they learn to speak they babble, they refine their speech sounds and shape their speech sounds through sensory motor integration. They listen to what they’re doing, they have a memory or a target that they’re trying to match—another speech sound, like the sound of their parent.
TC: How do you study the birds, exactly?
RM: We use a really wide variety of techniques to see what the brain is doing, figuratively and literally. Literally, we use an advanced kind of microscopy to look into the brains of birds as they learn to sing to actually see what changes occur to the structure of the brain as the juvenile bird is learning. It allows you to look fairly deeply into the brain, maybe half a millimeter or so, which in terms of optics is really a lot. It turns out a critical area for sensory actions is right on the surface of the brain. The figurative approach is we use electrical techniques to record the activity of single nerve cells in birds as they are singing and as they are listening to other birds sing to them. We use these microdrives. They’re these very tiny chassis—they weigh about a gram—that have little motors in them that allow us to move very fine microelectrodes, really really fine wires.
TC: Does the bird need to be subdued at all?
RM: The microdrive technology is really powerful because it allows you to monitor what individual nerve cells are doing as an animal engages in behavior. Not subdued, not anesthetized but freely behaving.... And it seems like female birds, when they see birds with the headgear, are stimulated by them—it’s kind of like a hat or something.
TC: What level of detail can you get in imaging the songbirds’ brains?
RM: You can resolve what’s going on at the single synapse level in the living animal. If you’re familiar with fMRI or PET scanning, this gives you resolution on the level of millions of neurons or many tens or hundreds of thousands, not at the single synapse level. The reason that distinction is important is because synapses are really the unit of fundamental organization in the nervous system. That’s the building block from which all the complex computational power of the brain is derived.
TC: What do you see when you look into a songbird’s brain?
RM: Birds, like humans and other vertebrates have a structure where they have the cell body and these branchlike structures called dendrites, these bushy structures emanating from the cell body. Those are the sites where other nerve cells make contacts, make synapses with that nerve cell and where nerve cells are able to signal one another. And there are specialized protrusions on the dendrites called spines..... So, a nerve cell with more dendritic spines has more synaptic partners, more cells talking to it. Similarly, a cell that loses spines loses synapses. The general dogma is that when we learn new behaviors and form new memories, synapses get built. But testing that in a natural learning paradigm has been really hard. The studies we did were really the first to do that.
TC: What did you observe about these dendritic spines during the process of learning song?
RM: We found dendritic spines in naïve birds can turn over at a really high rate. Turnover is the average of spines gained and lost, in this case in a relatively short interval. If you take those naïve birds and expose them to the tutor—the adult bird that provides the model that the bird will then copy— the birds with the higher turnover rate will learn more from the tutor. We monitor turnover in this area, we look at spines, we count how many come and go and develop a baseline measure. Then we take the bird and expose it to its tutor and we ask a couple months later how much did you learn from that tutor? The birds that had high turnover right at the time when they were exposed to the tutor learned a lot. The ones that had low turnover didn’t learn. We also found out that right before and after they hear the tutor song for the first time there’s this rapid stabilization of spines. Spines go from being dynamic to being really stable. In a 24-hour period things snap into a completely new state.
Get The Chronicle straight to your inbox
Signup for our weekly newsletter. Cancel at any time.