By Karen Eng
Greg Gage is a reliable source of both shock and awe at TED. Onstage over the years, this TED Fellow has demonstrated his low-cost DIY teaching kits by amputating a cockroach leg to show how neurons fire, remote-controlling a cyborg cockroach to demonstrate how electrical stimulation guides behavior, and taking away an audience member’s free will to show how one person’s brain can control the arm movements of another.
Gage, a latecomer to science himself, is passionate about revolutionizing neuroscience education. His goal is to make research equipment previously only accessible in university labs available to teachers and home enthusiasts. Here, he tells the TED Blog about the evolution of Backyard Brains, his plan to create a class of independent neuroscience researchers and his home for retired cockroaches.
Tell us about your scientific background. Were you always passionate about neuroscience?
I was actually an engineer for many years. I made circuit boards. I had a nice job with a technology company and lived in Europe, where I was in charge of engineering for Europe, the Middle East, Africa and South Asia Pacific. I always enjoyed science — I read Scientific American and science books, but I thought science was stuff that you learned in school. That it was made up of already-accumulated facts. I never realized that science was a career you could do — that you can actually get paid to make experiments, and understand how nature works.
That all changed when I attended an evening talk on astronomy at Leiden University in the Netherlands. After the lecture, I talked to the graduate students who presented, and found out this was their full-time job: making experiments, collecting data and writing results and papers. I realized: this could be me! I could actually become a scientist! So I quit my job and went back to grad school in Michigan. Meanwhile, everyone told me I was crazy to leave my well-paid and comfortable job.
What did you study?
I was a basal ganglia guy, studying deep brain structures. I recorded from the motor cortex, the nucleus accumbens, the striatum. I trained rats to do decision-making tasks. When they heard a tone, they’d go one way, and when I played another tone, they’d go another way. At some points, I’d play both tones at the same time to confuse them, and I’d observe which direction they chose. I’d use that data to look at brain cells. I recorded the spike trains from the cells, looking at the exact moment in which they were making the decision — then determine what cells were firing when, to get an idea of what the microarchitecture and microcircuitry of our brains are like.
We made some nice discoveries: we found that certain interneurons — which are missing in people with schizophrenia — fire really, really fast at the moment you’re making a decision. These cells seem to be suppressing unwanted decisions, and only allowing the ones that are the strongest to escape and be chosen. It actually fits pretty well with schizophrenia. This research was published in the journal Neuron, and was a fairly high-impact paper.
How did you veer into creating neuroscience kits?
Given my experience, I felt it was important to explain what scientific careers are, so young people wouldn’t miss their calling like I almost did. While I was in grad school, I did outreach. I’d visit schools with my neuroscience labmate Tim Marzullo to teach kids about how the brain works. We’d explain that there are really cool opportunities out there to study neuroscience. I used to tell kids, “If you like Sudoku, solving puzzles in general or building things, you’ll love being a scientist.”
Tim and I would enter these little competitions called “Brains Rule!” where we’d try to create better demos to get kids interested. That led to trying to bring what we did in our graduate research into the classroom, so kids could see more than just demos consisting of ping-pong balls as transmitters. Many science exhibits make science too “fun” so that it just becomes a game — and then kids leave without a real understanding of what the brain or neurons actually do.
In 2008, we identified this need to make it real. But we couldn’t bring in our equipment from the lab because it cost $40,000. We couldn’t bring our animals in, because that’s illegal. If someone wants to study the brain, they typically have to go to grad school — which is silly. This isn’t the case in other areas of science. You can study the planets or stars with a cheap telescope — you don’t have to get a PhD in astrophysics.
So we set about building what we called the “$100 spike” — inspired by Nicholas Negroponte’s $100 laptop. Could we build neuroscience equipment rugged enough that students could use it, and cheap enough that schools could afford it? Six months later, we revealed our first prototypes at the Society for Neuroscience conference. We got some publicity. People started writing us, wanting to buy one. People loved the idea of making neuroscience equipment available to classrooms.
I was still writing my dissertation, recording data, training rats. I’d had a couple of high-profile publications, but never really received much feedback. On this little $100 spike project, Tim and I were getting emails all the time. I was learning in my research how the basal ganglia uses dopamine to change the probability of future behaviors. So it made sense that as we kept getting positive attention for the project — and none for our graduate school work — we decided to focus on this venture. We named it Backyard Brains.
What do the products do and teach?
We have a number of inventions. The first one was the Neuron SpikerBox, a kit that allows you to record the living brain cells of insects. The idea was to bring electrophysiology into the classroom. We demonstrated the process at TEDYouth: first you anesthetize the insect in ice water — which is the recommended way to do it, according to Vincent Wigglesworth, who published the standard paper on insect pain in 1980.
When the cockroach is anesthetized, we remove one of its legs, and let the leg warm back up so the neurons fire. We put the pins in the leg, which pick up the small electrical discharge from the spikes and amplify it — so you can hear it. Then you can plug it into your smartphone and can actually see, record and do data analysis on it.
With that prep, you can then carry out a dozen or so experiments. You can look at somatotopy — what do different parts of the legs encode, and what is the location of these neurons? This is very much like the different parts of our brain; our somatosensory cortex is laid out in certain ways, so that our fingers and hands are represented by large areas of the brain while the backs of our thighs are represented by a very small area of the brain. You can do that experiment in the cockroach leg, and actually figure out some of the representations of neurons in certain areas of the leg.
So can you tease out each one? Say, this neuron is about movement, and this one is about pain, and this one is about being touched?
It’s more about density and location. What’s important to the cockroach is in the tarsus and tibia areas — the hands. You can see a lot more dense representation of neurons there, while in the upper arm area you don’t see as much.
You can also do neuropharmacology — which is looking at how neurons respond to an increase in neurotransmitters. You can look at functional electrical stimulation, which is what’s used for treating people with stroke — basically, stimulating muscles using an electrical discharge. That’s what we do when we make the leg dance with the music from an iPod.
These are really advanced neuroscience experiments we’re allowing people to do at an amateur and high-school level. We have another line of products called the Muscle SpikerBox, which allows you to record from the output of the human brain — the muscles. You’re recording from the motor cortex on down, so it records from the arms and potential actions of the hand. It records the individual motor units from the lower motor neuron in the spinal cord — so you can actually see a little pulse as that neuron in the brain is telling the muscle to move.
What about the Roboroach, the kit that lets you remote-control a live cockroach. You billed it as the world’s first commercially available cyborg.
Roboroach is an interesting invention, because it allows us to study behavioral effects of the brain. You surgically fit an electronic backpack onto the roach, and it sends an electrical current directly into the antenna nerves. When you use the app to send the current, the roach responds with a turning behavior.
You can then ask, “Why is that cockroach doing that?” It’s the nature of the roach — you touch its antenna, and it turns in the other direction. It’s called a wall-following behavior. With the Roboroach kit, we’re talking to the same neurons using small pulses of electricity. We’re making the roach think it’s touching something.
Behavior is what’s really interesting about neuroscience. Neurons are the things that we’re firing when we do and think anything. They drive behavior. The more you can see neurons and behavior working together, the more interesting it gets.
The fact that the SpikerBox and Roboroach require removing parts of the cockroach caused quite a bit of controversy among animal activists. What do you say about that?
I think it’s partly perception of what the Roboroach, for example, does. Some people thought that Roboroach is a permanently remote-controlled insect — and that it was a slippery slope to something more macabre. But that’s not what’s happening. The reality is that this works on the cockroach because the insect naturally follows simple rules. But very soon, it adapts to a unnatural stimulus. After 15 minutes, the roach ignores the backpack. It retains its free will.
The other issue is about pain or damage to the cockroach. To install the micro-stimulator, you put the cockroach under ice water, you remove a portion of its antenna, and you place some stimulating wires inside. Afterwards, you remove the backpack and put the cockroach back in its cage. We’ve looked carefully at behavior before and after surgery, and the cockroach appears, in every way, shape and form, to function with the smaller antennae.
We have a retirement community called “Shady Acres” for roaches that have given their service to Backyard Brains. When we put food in the cage, their antennas move and they walk over. They appear to be functioning well. They also live just as long as other cockroaches do — the death rate of roaches used in experiments is equal to that of controls. In nature, you’ll see cockroaches in the wild missing antennae and even limbs. Their ability to adapt easily to damage is different from that of humans.
But the real question is: what is the human benefit and does it outweigh the cost to the cockroach? This is a question you have to ask every time you do an animal experiment. The benefit is the ability to demonstrate neurotechnology to a group of students who may be interested in pursuing a career in science. Students are able to study neural systems and behavior, and learn how the most advanced neurological treatments work in humans. The Roboroach is deep brain stimulation — the same technology used to treat diseases like Parkinson’s. About 20% of the world is diagnosed with a neurological disorder that doesn’t have a cure. So I think the benefits to humanity make it our moral responsibility to teach about the brain using insects.
Can you use Backyard Brains to study cognition in humans?
We have a kit that can measure how much time it takes for your eyes to see something and for you to react. We can record the delay between a green LED coming on and your muscles moving in response. It takes about 350 milliseconds.
We can make the experiment more complex. Instead of just a green light, we can use an additional red light to distract you. The task is the same: react as fast as you can to the green light. But reaction time is longer because you need to be sure the color is correct first. You can record how long it takes your brain to deal with this extra cognitive step.
You can also use a tone, so you know how long information takes to go from your ear to your muscle. It turns out the eye processes information 100 milliseconds faster, because we have more neurons in our visual pathway, as we’re more visual creatures.
What are some of your latest inventions?
We’ve been moving into neural interfaces — connecting machines to the brain via electrical signals we can detect. We have devices that snap onto an Arduino board, and students build their own computer interfaces from their muscles, heart or brain. Our talk at this year’s TED used a Muscle SpikerBox paired with an Arduino to control a muscle stimulator. We call it the “human-to-human interface.”
We focus on the latest technology that labs are using, and make DIY versions of it. Right now we’re developing an OptoStimmer that will soon make affordable optogenetics classroom tools. Optogenetics is a breakthrough technology that allows you to turn on and off specific neurons in the brain by genetic targeting. Specific neurons enable a gene that grows little channels that make the cell communicate when you shine a light. Normally, light doesn’t affect neurons, so this technique causes only targeted neurons to fire spikes. You can pulse this light, and all the other neurons ignore it except for the ones that you want to control. You now have this amazing ability to turn on or off any neuron, any time.
I can’t overstate how important this technique is to our field. This has been the holy grail, so you can figure out what neurons are actually doing. It’s given answers to long-standing debates. For example, no one really knew how deep-brain stimulation for treating Parkinson’s works. There were theories and models, but now it can be tested. Scientists carefully targeted possible neurons in mice and other animals using optogenetics, pulsed the light and looked to see which neurons actually had therapeutic effects. It turns out it was not even where they were stimulating in the deep brain — it was another section way far away, the motor cortex, that mattered.
We’d like to make this tool available for high school students to do experiments using fruit flies. They’ll be able to do some recordings on optogenetic flies under a microscope. Then pulse a little bright red LED, which shines through the skin of the fruit fly. When a targeted neuron sees the red light, it will start firing. Depending on which neuron you are targeting, you can drive vastly different behaviors: from thinking they tasted something sweet to moonwalking like Michael Jackson. You’ll be able to see how specific neurons are affecting behavior.
Are there amateur neuroscientists outside the classroom using Backyard Brains products to do research?
Yes. One of our goals is to have a peer-reviewed paper that comes from an amateur with an institutional address that is their home address. It’s happening already in mathematics and astronomy, but not in neuroscience. We want to change that. We want real discoveries to happen at home, using our gear.
What I like about Backyard Brains is that we not only push out products, we push out experiments. We want our experiments to be novel and educational. We want to develop new tools and techniques that we can publish in academic peer-reviewed journals. We train undergraduates on how to do experiments, and we write those up and publish them. Our first article about the SpikerBox was published in 2012 in PLOS ONE, and we’ve been publishing every year since.
Our work is independent from any university and is financed with the money that we’re generating from grants and sales at Backyard Brains. It feels great to finally be an independent scientist. Our goal is to see scientific neuroscience papers published by amateur scientists. The neuro-revolution is coming.
Above: Backyard Brains shows exactly what happens when you play hip-hop music into the light-reflecting nerve cells of a squid.