Fly brain 'atlas' opens door to linking human neurons to actions

By Sharon Begley NEW YORK (Reuters) - Research unveiled on Thursday identifies the brain neurons that cause every behavior fruit fly larvae execute, raising the tantalizing possibility that neuroscientists will be able to construct a similar "atlas" in people. The accomplishment is an advance toward linking the neurons that fire when people make specific movements, and possibly even when they feel certain emotions, visualize objects, hear particular melodies or think certain thoughts. Such a human brain atlas is one goal of the $100 million BRAIN Initiative that President Barack Obama announced a year ago, and the fly research raises hopes that it will be possible to deal with the data deluge that project will generate. "This study is part of a sea change in neurobiology where we'll have to deal with vast amounts of data, so it's encouraging to see what they could do," said physicist Aravinthan Samuel of Harvard University, who applies computational techniques to the brains of the simplest animals and was not involved in the new research. For the study, which was published online by the journal Science, scientists led by biologist Marta Zlatic of the Howard Hughes Medical Institute's Janelia Farm research center in Virginia first activated a few neurons at a time in the larvae of the fruit fly drosophila, using a technique called optogenetics, in which light causes particular neurons to fire. She and her team then compiled thousands of hours of video recordings of how 37,780 fly larvae behaved in response to each neuronal activation. Hunching and wiggling, backing up and turning continuously were all popular. Next, using specially developed software to analyze the terabytes of data in the recordings, mathematician Carey Priebe of Johns Hopkins University identified the 29 behaviors the larvae can manage: crawling forward or backward, turning right or left, and the like. The result was an atlas of neurons whose activation causes any movement a fly larva is capable of. Two neurons in the bottom back of the brain, for instance, made most larvae turn, turn, and turn again. The drosophila project was feasible not only because the larvae are capable of so few behaviors but also because their brain and nerve cord contain only 10,000 neurons. The human brain has about 86 billion to 100 billion neurons. "It's a proof of principle," said Zlatic. "It's possible to classify neurons in terms of what decisions and actions they cause." Eventually, such a neuron-behavior atlas might be paired with the wiring diagram of a brain. In the case of the fly larvae, Zlatic's team identified which neurons cause which behaviors, but not which neurons were activated in between or how these neurons are connected. Doing so will require what neuroscientists call a "connectome," or the complete wiring diagram of a brain. Neurobiologists are close to accomplishing that for the mouse brain. The Human Connectome Project, launched in 2009 by the National Institutes of Health, is attempting to do so for the 10,000 or so connections made by each of the billions of neurons in the human brain. By pairing the human connectome with a neuron-behavior atlas like the one produced for fly larvae, "one might be able to figure out the principles by which brains work," said Zlatic. One of many possible hurdles is that, even in fly brains, activating the same neurons in different animals did not always evoke the same behavior. In fact, activating the same neurons in the same larva did not always trigger the same behavior. It did, however, make only a few of the 29 behaviors more likely, such as forward progress; activating a particular neuron did not make one larva crawl forward and another turn in circles, for instance. Still, the fact that the relationship between neuron activation and behavior is probabilistic rather than deterministic suggests that it will be that way in humans, too, making a clean neuron-behavior atlas a challenge, Zlatic said. "But I don't think it's completely unimaginable that the same kinds of approaches" used to create a neuron-behavior atlas in flies "can be scaled to bigger animals," said Harvard's Samuel. (Reporting by Sharon Begley; Editing by Michele Gershberg and Douglas Royalty)