Alexander Shakeel Bates, Jasper S Phelps, Minsu Kim, Helen H Yang, Arie Matsliah, Zaki Ajabi, Eric Perlman, Kevin M Delgado, Mohammed Abdal Monium Osman, Christopher K Salmon, Jay Gager, Benjamin Silverman, Sophia Renauld, Matthew F Collie, Jingxuan Fan, Diego A Pacheco, Yunzhi Zhao, Janki Patel, Wenyi Zhang, Laia Serratosa Capdevilla, Ruairí JV Roberts, Eva J Munnelly, Nina Griggs, Helen Langley, Borja Moya-Llamas, Ryan T Maloney, Szi-chieh Yu, Amy R Sterling, Marissa Sorek, Krzysztof Kruk, Nikitas Serafetinidis, Serene Dhawan, Tomke Stürner, Finja Klemm, Paul Brooks, Ellen Lesser, Jessica M Jones, Sara E Pierce-Lundgren
5 min
This study presents the first unified, synapse-resolution connectome of the entire central nervous system (CNS) of an adult fruit fly, Drosophila melanogaster, spanning both the brain and the ventral nerve cord (VNC). By mapping the connections between sensory inputs, interneurons, and effector outputs (motor neurons, endocrine cells, and visceral efferents), the researchers investigated the fundamental principles governing behavioral control in a complex, limbed organism.
To analyze this massive dataset, the authors developed a scalable influence metric based on linear dynamical modeling. This metric allows researchers to estimate the functional weight of connections between any two neurons across the entire CNS. By clustering these influences, the team identified behavior-centric modules—groups of neurons that coordinate specific actions like flight, feeding, or escape—and analyzed how these modules interact with one another and with supervisory brain regions like the mushroom body and central complex.
The connectome reveals that behavioral control is not centralized but highly distributed. Effector neurons are primarily influenced by local sensory feedback loops, which minimize latency and simplify control. These local loops are integrated by long-range ascending and descending neurons (ANs and DNs) that coordinate activity across different body parts. The researchers found that these ANs and DNs are organized into functional superclusters, which act as behavior-centric modules. These modules often exhibit a subsumption-like architecture, where higher-level circuits can recruit or suppress lower-level modules to prioritize specific behaviors, such as interrupting walking to initiate an escape response.
This work provides a foundational map for understanding how a complex nervous system manages embodied control. By demonstrating that behavioral control is parallelized and distributed, the study challenges top-down models of neural processing and provides a blueprint for how neural systems achieve flexible, robust behavior. The open-source nature of the BANC dataset and the associated influence metrics offer a powerful resource for future experimental testing of these circuit-level hypotheses.
Abstract Just as genomes revolutionized molecular genetics, connectomes (maps of neurons and synapses) are transforming neuroscience. To date, the only organisms with complete connectomes are worms 1–3 , sea squirts 4 and comb jellies 5 (10 3 –10 4 synapses). By contrast, the fruit fly is more complex (10 8 synaptic connections), with a brain that supports learning and spatial memory 6,7 and an intricate ventral nerve cord analogous to the vertebrate spinal cord 8–12 . Here we report a densely reconstructed adult fly connectome that unites the brain and ventral nerve cord, and we leverage this resource to investigate principles of neural control. We show that effector neurons (motor neurons, endocrine cells and efferent neurons targeting the viscera) are primarily influenced by sensory neurons in the same body part, forming local feedback loops. These local loops are linked by long-range circuits that involve ascending and descending neurons organized into behaviour-centric modules. Single ascending and descending neurons are often positioned to influence the voluntary movements of multiple body parts, together with the endocrine cells or visceral organs that support those movements. Brain regions involved in learning and navigation supervise these circuits. These results reveal an architecture that is distributed, parallelized and embodied, reminiscent of distributed control architectures in engineered systems 13,14 .
Sam: It is, and the authors are explicit about it. The model assumes activity propagates linearly and reaches equilibrium — which means it inherently ignores nonlinear dynamics, temporal coding, and neuromodulation. Those aren't minor caveats. If the fly is executing a rapid escape maneuver, it isn't operating anywhere near a steady state. Transient, high-gain signals are almost certainly overriding the baseline loops the metric is designed to characterize.
Alex: So the metric is really a map of structural favorability — which pathways are wired to carry influence — not a simulation of real-time spiking dynamics.
Sam: That's the right framing. It generates hypotheses about which circuits matter, ranked by structural potential. Whether those rankings actually predict which circuits are recruited during active behavior is an open question. The honest next step is integrating this connectome with whole-brain calcium imaging — seeing whether the adjusted influence scores predict observed recruitment patterns in behaving animals. The map is detailed; the traffic hasn't been measured yet.
Alex: And that gap is where a careful referee would push back hardest.
Sam: Probably, yes. The other place a referee might push is generalizability. This is one adult fly, one sex, one developmental stage. Connectomes vary across individuals, and the fly's nervous system is plastic enough that experience could shift some of these weights. So the topology here is a reference, not a universal ground truth.
Alex: Does any of this change how we should think about the brain as a concept?
Sam: I think it reframes the question. The temptation in neuroscience has always been to look for the place where behavior is decided — the locus of control. What this connectome suggests is that for a large class of behaviors, there isn't one. Control is distributed across the nervous system as a whole, with the brain as one layer in a parallelized network rather than the apex of a command hierarchy. That has real implications for how we model motor control, and probably for how we think about what cognition is doing relative to the rest of the system.
Alex: A useful reframing — and one that's now grounded in the most complete wiring diagram of an adult nervous system we've had. Thanks for walking through the mechanism, Sam. And thanks for listening to ResearchPod.