Why would an aerospace engineer study the fruit fly? The fly may be an impressive flying machine, but what attracts Janelia Farm fellow Vivek Jayaraman to the insect is its brain.
"I took the scenic route to neuroscience," is how this engineer-turned-scientist puts it. As an undergraduate at the Indian Institute of Technology in Bombay, Vivek found himself drawn to aerospace engineering by the dramatic images of the space shuttle and space travel, "although it was more of a boyhood dream rather than something carefully thought out," he says. He stayed in the field long enough to earn a master's degree at the University of Florida, where his research focused on computational simulations of fluid flow. Along the way, Vivek became captivated by a different field, artificial intelligence (AI), and began thinking about applying his quantitative skills to problems in that area.
"At first, I was curious about why AI couldn't simply mimic the algorithms and coding principles that the brain used," he recalls. "But it quickly became clear that part of the reason was that we didn't know much about what the brain actually did."
Vivek's interest in that profound mystery eventually led him to Caltech to study neuroscience. There, in Gilles Laurent's laboratory, he realized that insects were excellent models for studying how neural circuits process sensory information. One of his studies on locusts with colleague Mark Stopfer explored how the insect's olfactory neural circuitry represented the same odor at different intensities.
"Responses of individual olfactory neurons in the locust brain can change dramatically depending on the intensity of the odor. So, looking at the responses of a single neuron, a faint whiff of coffee might seem as different from a strong blast of coffee as, say, the scent of cherry from that of banana. The interesting question is how your brain nonetheless identifies the aroma as coffee, whether it's coming from the table across the room or the cup you're drinking from," he says.
Vivek used computational techniques to analyze the dynamics of ensembles of olfactory neurons. The analysis, which allowed him to "decode" the ensemble responses much as the locust brain itself does, revealed consistent patterns not detectable when studying single neurons. Thus, for the system as a whole, different responses to different intensities of an odor are merely variations on a theme. "It's an encoding scheme that allows the next stage of the neural pathway, and presumably the locust itself, to reliably recognize and differentiate odors across intensities," he says.
Vivek continues to look at questions of neural coding at Janelia Farm, but he has switched to the fruit fly as his model system. His entry into fruit fly research was sparked by the recent development of a method to electrically eavesdrop on individual neurons in the exposed brain of an intact fly—a remarkable technical achievement by Caltech colleagues Rachel Wilson and Glenn Turner.
Vivek adapted this method to record from neurons electrically while optically monitoring their activity using fluorescent proteins under a specialized microscope. This is the first step toward performing such recordings while simultaneously observing the fly's behavior, he says.
At Janelia Farm, he uses this combination of techniques to help turn the fruit fly into a premier model for understanding how neural circuitry processes sensory information and determines the choices that the fly makes. "We know something about how flies behave in a multisensory environment, and have indications of which brain regions are involved in sensory processing," he says. "But what kinds of computations do neurons in those regions perform to make the fly pick, for example, a ripe banana over a raw one? This may not be the same process by which one might choose, say, a career, but understanding the intricacies of decision-making in simple systems can nonetheless provide us with insights into the workings of more complex brains."
One key factor that makes the fruit fly a good neurophysiological and behavioral model is the ability to use genetic mutation to precisely alter the activity of selected neurons. At Janelia Farm, he draws on the expertise in fly genetics of group leader Julie Simpson and in fly behavior of fellow Michael Reiser.
However, says Vivek, fully harnessing the potential of the fruit fly as a model for neural information processing requires the ability to record from populations of neurons with high resolution, something that new imaging techniques may soon make possible. "Today, you can express sensors of neuronal activity in specific subsets of neurons and see them light up. But these sensors aren't yet precise enough to tell you how many spikes the neurons fired, and they don't say anything about the timing of individual spikes. Spikes are the main currency of communication within the nervous system, so these are things you want to know." Vivek's current approach is to use imaging for the population-level perspective and electrical recordings for a higher resolution investigation of selected neurons.
Janelia Farm offers a great opportunity to pioneer a fly model linking the dynamics of brain circuitry with behavior, says Vivek. "The appeal of Janelia Farm for me is that I can work with people with expertise in fly genetics and behavior, imaging, and protein design. This is not a safe project that will produce an immediate result," he says. "But at Janelia, there's more opportunity and freedom to think long-term, and to take such risks. The people here are thinking less about small-scale projects and quicker payoffs."
