Local circuit loops within body control fly behavior, new ‘embodied’ connectome reveals

As powerful as it is, the nervous system can’t act alone—it needs a body to sense the environment and carry out behaviors. Yet existing maps of the brain’s connections omit how it interacts with the rest of the body.

“The previous connectomes, which were wonderful and important, were kind of disembodied—they were large groups of cells connected with each other, but it was kind of a brain in a vat,” says Rachel Wilson, professor of neurobiology at Harvard University.

A new connectome, led by an international group of researchers, including Wilson, adds that layer of body connections to our understanding of the Drosophila melanogaster nervous system. The team described the new map in a preprint posted on bioRxiv last month and made the data available publicly online.

“A big advantage of this connectome is they put a lot of focus on how the nervous system is connected to the body,” says Anita Devineni, assistant professor of biology at Emory University, who was not involved with the effort. “It’s about it being embodied.”

For researchers, it “allows you to analyze the flow of information into the nervous system, through the nervous system and out of the nervous system at a level of both precision and completeness that is really unprecedented,” Devineni adds.

Most effector neurons, or those that control body parts such as organs, muscle or viscera, receive the strongest connections from sensory neurons in that same body part, which then get information back from the region, forming local feedback loops, the team found. Weaker long-range connections between the brain and the ventral nerve cord link these local loops together, which may underlie the control of different behaviors.

“It seems like those local loops are the building block of behavior, and then you link them together to get an actual coordinated behavior,” Devineni says. The findings challenge the idea that behavioral control is centralized; rather, it is local and distributed, she adds. “Organs are basically controlling themselves with some input and gentle nudges from longer-range signals.”

T

o create the new resource, the team performed serial-section electron microscopy on tissue slices across the entire brain and ventral nerve cord of an adult female fruit fly, categorized all of the cells and identified which areas of the body, such as muscles or organs, those cells connected to. Then they estimated the degree of “influence” one neuron has on another by simulating how signals travel through the network.

Sensory neurons within a particular organ or body part have the most influence over effector neurons in that same area, the analysis found. For example, pharynx sensory cells have the greatest influence over pharynx motor neurons. “Cognitive regions of the brain are not probably the immediate cause of most actions—most actions are probably driven by these local loops,” Wilson says.

This type of setup, similar to reflex circuits, “would make the behavioral execution a little bit faster,” says Martha Bhattacharya, associate professor of neuroscience at the University of Arizona, who was not involved in the study.

Some of the ventral nerve cord’s ascending and descending neurons, it turns out, cluster into groups based on their synaptic connectivity and their influence on sensory and effector neurons. Many of these clusters contain neurons that control specific behaviors, such as walking, feeding and grooming.

One cluster, for example, contains visual neurons, mechanoreceptors, endocrine neurons and other cells previously linked to escape behavior, so it is likely in charge of responses to threats, the team predicted. This cluster might integrate visual and tactile inputs to recruit energy stores and help the fly flee, Devineni says. The findings give a “bigger picture of what the circuits actually look like.” Clusters also influence one another; the threat-response cluster, for instance, strongly influences the walking cluster.

Central brain regions, on the other hand, weakly influence these local loops and clusters, the team found, so they may be acting as “supervisors” of the lower-level modules, Wilson says. “It’s kind of a mystery how the cognitive regions of the brain actually have such a big influence on behavior, given that they make these sort of weak, distributed anatomical connections over the body parts that actually control movement.”

The next step is to experimentally test the functions of the identified loops and clusters, because there might be situations in which the anatomy does not accurately predict the exact connectivity or behavior, Bhattacharya says. Mismatches between the anatomical and functional connectomes have already been documented for fruit flies and roundworms.

And if some of the fly’s behavior control really is decentralized, it raises questions about whether similar distributed networks exist in other organisms, including humans, or if the cognitive and centralized regions are something that evolved, Wilson says. “If centralized decision-making exists in humans—and it may or may not—where in evolution does that arise, and where does it reside?”