AI summary · arXiv: 2610.33881v1 · q-bio.NC
Sensorimotor behaviour from an untrained brain-and-nerve-cord connectome
The essentials
An untrained fly connectome supports simulated odour navigation through fixed sensorimotor rules, while associative learned avoidance requires correcting odour-dependent activity and using a signed output readout.
Question
How much sensorimotor function follows from the measured wiring of an adult fly’s brain and nerve cord without task-specific parameter fitting? Can that wiring support navigation and learned behavioural reversal inside a simulated body?
Method
The authors built a leaky integrate-and-fire network from the female BANC connectome: 172,433 units, 11.75 million connections and 35.7 million synapses, using unchanged parameters from a published brain model. They tested 15 known pathways and coupled the network to NeuroMechFly through fixed sensory and motor rules, comparing intact, degree-and-sign-preserving rewired and mushroom-body-lesioned networks across nine arenas containing 32 sources, with 20 simulation seeds per condition. Learning experiments imposed dopamine-gated depression on 8,377 Kenyon-cell-to-output-neuron connections and compared paired reinforcement, reinforcement alone and no plasticity, including a variant with cholinergic local-neuron output scaled to 0.15.
Findings
-
Result
In the identified-target tests, pathways supplying at least 2% of their target’s input conducted at some stimulation rate, whereas seven pathways supplying 0.03% or less did not.
See the passage in the paper
Every pathway tested onto an identified neuron conducts at some drive if it supplies at least 2% of that neuron’s input, and the seven supplying 0.03% or less do not (Fig. 2f).
-
Result
Across the nine arenas, the intact network reached 95% of odour sources, compared with 18% after rewiring that preserved neuronal degrees and signs.
See the passage in the paper
In the NeuroMechFly body[38], coupled through two fixed rules, the network reaches 95% of 32 odour sources across nine arenas, where a network rewired to keep every neuron’s degrees and signs reaches 18%.
-
Result
Replaying recorded motor commands without sensory feedback reduced source acquisition to 2%, compared with 94% for closed-loop runs from the same rotated starting orientation.
See the passage in the paper
The replays reach 0.07 sources per run against 3.29 for the closed loop from the same rotated start (2.0% against 94.0% of the sources, per fly).
-
Result
The as-read olfactory network produced dense, persistent Kenyon-cell activity, while scaling one excitatory synapse class to 0.15 brought the tested odour code into the fly’s measured range.
See the passage in the paper
This closes a loop between projection and local neurons that sustains itself once started, and two thirds of the Kenyon cells then fire to every odour. Scaling that one class of synapse to 0.15 of its face value brings the Kenyon-cell code into the fly’s range.
-
Result
In the scaled variant, paired reinforcement produced avoidance in all 20 flies, whereas reinforcement alone preserved approach in all 20, unlike the non-associative reversal in the as-read network.
See the passage in the paper
Pairing takes the weight to 0.195 mV and the valence to $-0.53$, and none of 20 trained flies reaches the source (median closest approach 21.1 mm). Reinforcement alone leaves the weight at 0.508 mV and the valence at 1.38, and all 20 flies reach the source, as they do without plasticity (paired against reinforcement alone, $P=6.8\times 10^{-8}$, $\delta=+1.00$) (Fig. 6h, i).
Limitations
- Odour-directed turning is partly supplied by a hand-written interface, so the rescue experiments identify a required output signal rather than demonstrate that the connectome computes the turn.
See the passage in the paper
The rescue locates the deficit of both manipulations at the mushroom-body output. It does not show that the connectome computes the turn.
- The model uses one animal’s wiring and simplified neuronal dynamics, omitting electrical synapses, spontaneous activity and endogenous neuromodulation.
See the passage in the paper
The model makes as few assumptions as a spiking network can: point neurons with one sign each, chemical synapses only, no neuromodulation, no spontaneous activity, one animal.
- Learned avoidance depends on the chosen signed output readout and restriction of reinforcement to aversive compartments.
See the passage in the paper
The trained fly turns away only when the readout is signed and the reinforcement is confined to the aversive compartments, where 41 of the 50 MBONs drive approach.
Why it matters
The authors present an unfitted baseline for separating contributions of measured wiring from those of the body interface and sensory feedback. Its successes and identifiable failures provide fixed tests for adding missing physiology and motivate experimental predictions about pathway input share, activity-dependent steering and associative learning.
Written by AI from the paper. Passages are checked word for word against the text we analysed; the rest may contain errors.
Ask questions about this paper to our AI assistant
How to chat with this paper? Illustrated guide
You can also chat with multiple papers at once here.
Keep this conversation — a free account saves your questions and answers, your collections of papers and your alerts.
Create a free accountTools for this paper
Advanced tools Premium
For researchers who follow a field closely: citation alerts, journal-club slides, version comparison and Zotero export. Discover Premium
Papers around this one
See them on a mapRanked by how close their title and abstract are to this paper’s. Select some to read them together.
Closest papers
-
Distributed control circuits across a brain-and-cord connectomeA. Bates, Jasper S. Phelps, Minsu Kim et al. · 2026 · Nature · cited 69 times · Not summarized yet
Distributed control circuits across a brain-and-cord connectome provides the adult fly brain-and-nerve-cord wiring used to build the simulated network.
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 worms1, 2–3, sea squirts4 and comb jellies5 (103–104 synapses). By contrast, the fruit fly is more complex (108 synaptic connections), with a brain that supports learning and spatial memory6,7 and an intricate ventral nerve cord analogous to the vertebrate spinal cord8, 9, 10, 11–12. 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, endoc…
-
Connectomic reconstruction of a female Drosophila ventral nerve cordAnthony W. Azevedo, Ellen Lesser, Jasper S. Phelps et al. · 2024 · Nature · cited 88 times · Not summarized yet
Connectomic reconstruction of a female Drosophila ventral nerve cord maps the female fly nerve-cord wiring incorporated into the brain-and-cord model.
Abstract
A deep understanding of how the brain controls behaviour requires mapping neural circuits down to the muscles that they control. Here, we apply automated tools to segment neurons and identify synapses in an electron microscopy dataset of an adult female Drosophila melanogaster ventral nerve cord (VNC)1, which functions like the vertebrate spinal cord to sense and control the body. We find that the fly VNC contains roughly 45 million synapses and 14,600 neuronal cell bodies. To interpret the output of the connectome, we mapped the muscle targets of leg and wing motor neurons using genetic driver lines2 and X-ray holographic nanotomography3. With this motor neuron atlas, we identified neural c…
-
Whole-Brain Connectomic Graph Model Enables Whole-Body Locomotion Control in Fruit FlyZe-Hao Jin, Yao Zhu, Chen Zhang et al. · 2026 · arXiv.org · preprint, not peer reviewed · cited 10 times · Not summarized yet
Whole-Brain Connectomic Graph Model uses an adult fly connectome to control a simulated body, but trains its controller with deep reinforcement learning.
Abstract
Animals perform coordinated whole-body movements under the control of neural systems shaped by brain-wide connectivity. The mapping of the whole-brain neural connections, or the connectomes, provides a natural graph for modeling sensorimotor information flow, yet its potential as a neural controller for embodied agents remains largely unexplored. Here, we introduce the Fly-connectomic Graph Model, which directly instantiates the whole-brain connectome of an adult Drosophila as a graph-structured neural controller for movements of a simulated biomechanical fruit fly via deep reinforcement learning. We achieve stable performance across diverse locomotion tasks, as well as better sample efficie…
-
The connectome of an insect brainMichael Winding, B. Pedigo, Christopher L. Barnes et al. · 2022 · bioRxiv · cited 383 times · Not summarized yet
The connectome of an insect brain maps the larval fly brain and its brain–nerve-cord pathways, providing earlier connectomic context.
Abstract
Brains contain networks of interconnected neurons, so knowing the network architecture is essential for understanding brain function. We therefore mapped the synaptic-resolution connectome of an insect brain (Drosophila larva) with rich behavior, including learning, value-computation, and action-selection, comprising 3,013 neurons and 544,000 synapses. We characterized neuron-types, hubs, feedforward and feedback pathways, and cross-hemisphere and brain-nerve cord interactions. We found pervasive multisensory and interhemispheric integration, highly recurrent architecture, abundant feedback from descending neurons, and multiple novel circuit motifs. The brain’s most recurrent circuits compri…
-
What an atlas-fitted connectome can and cannot do: evolutionary search over interneuron stimulation in a whole-body C. elegans modelTaehee Lee · 2026 · bioRxiv · preprint, not peer reviewed · Not summarized yet
What an atlas-fitted connectome can and cannot do tests which behaviours connectome wiring supports in a whole-body simulation, using a C. elegans model.
Abstract
The C. elegans connectome is complete, but which behaviors its wiring can produce, and which it cannot, is unknown. We asked this question in a whole-body simulation whose only adaptive element is a learning agent outside the nervous system. The agent may inject current into a short whitelist of interneurons, never into motor neurons or muscles; the 302-neuron OpenWorm c302 network, with dynamics fitted once to a signal-propagation atlas, a literature-based motor layer and a viscoelastic body on agar do the rest. Every hypothesis was pre-registered with its judgment criterion. Stimulating five command and steering interneurons (AVB, AVA, SMDD, SMDV, RIV) was sufficient for chemotaxis to a so…
-
A Connectome of the Male Drosophila Ventral Nerve CordShin-ya Takemura, K. Hayworth, Gary B. Huang et al. · 2023 · bioRxiv · preprint, not peer reviewed · cited 106 times · Not summarized yet
A Connectome of the Male Drosophila Ventral Nerve Cord provides a male nerve-cord wiring map, whereas this paper models the female fly.
Abstract
Animal behavior is principally expressed through neural control of muscles. Therefore understanding how the brain controls behavior requires mapping neuronal circuits all the way to motor neurons. We have previously established technology to collect large-volume electron microscopy data sets of neural tissue and fully reconstruct the morphology of the neurons and their chemical synaptic connections throughout the volume. Using these tools we generated a dense wiring diagram, or connectome, for a large portion of the Drosophila central brain. However, in most animals, including the fly, the majority of motor neurons are located outside the brain in a neural center closer to the body, i.e. the…
-
Connectome-constrained networks predict neural activity across the fly visual systemJanne K. Lappalainen, Fabian Tschopp, Sridhama Prakhya et al. · 2024 · Nature · cited 143 times · Not summarized yet
Connectome-constrained networks predict visual-system activity from measured connectivity, testing neural computation rather than embodied navigation and learned reversal.
Abstract
We can now measure the connectivity of every neuron in a neural circuit1–9, but we cannot measure other biological details, including the dynamical characteristics of each neuron. The degree to which measurements of connectivity alone can inform the understanding of neural computation is an open question10. Here we show that with experimental measurements of only the connectivity of a biological neural network, we can predict the neural activity underlying a specified neural computation. We constructed a model neural network with the experimentally determined connectivity for 64 cell types in the motion pathways of the fruit fly optic lobe1–5 but with unknown parameters for the single-neuron…
-
The fly connectome reveals a path to the effectomeDean A. Pospisil, Max J. Aragon, S. Dorkenwald et al. · 2024 · Nature · cited 43 times · Not summarized yet
The fly connectome reveals a path to the effectome combines connectivity with optogenetic perturbations to estimate causal effects, rather than simulating behaviour from wiring.
Abstract
A goal of neuroscience is to obtain a causal model of the nervous system. The recently reported whole-brain fly connectome1–3 specifies the synaptic paths by which neurons can affect each other, but not how strongly they do affect each other in vivo. To overcome this limitation, we introduce a combined experimental and statistical strategy for efficiently learning a causal model of the fly brain, which we refer to as the ‘effectome’. Specifically, we propose an estimator for a linear dynamical model of the fly brain that uses stochastic optogenetic perturbation data to estimate causal effects and the connectome as a prior to greatly improve estimation efficiency. We validate our estimator in…
Prior work it builds on
-
Mushroom body output neurons encode valence and guide memory-based action selection in DrosophilaYoshinori Aso, Divya Sitaraman, Toshiharu Ichinose et al. · 2014 · eLife · cited 637 times · Not summarized yet
Mushroom body output neurons encode valence and guide memory-based action selection reports valence-related MBON roles underlying the simulated odour-valence reversal.
Abstract
Animals discriminate stimuli, learn their predictive value and use this knowledge to modify their behavior. In Drosophila, the mushroom body (MB) plays a key role in these processes. Sensory stimuli are sparsely represented by ∼2000 Kenyon cells, which converge onto 34 output neurons (MBONs) of 21 types. We studied the role of MBONs in several associative learning tasks and in sleep regulation, revealing the extent to which information flow is segregated into distinct channels and suggesting possible roles for the multi-layered MBON network. We also show that optogenetic activation of MBONs can, depending on cell type, induce repulsion or attraction in flies. The behavioral effects of MBON p…
-
NeuroMechFly, a neuromechanical model of adult Drosophila melanogasterVíctor Lobato-Ríos, S. T. Ramalingasetty, Pembe Gizem Özdil et al. · 2021 · Nature Methods · cited 84 times · Not summarized yet
NeuroMechFly provides the simulated fly body and environment coupled to the connectome network.
Abstract
Animal behavior emerges from an interaction between neural network dynamics, musculoskeletal properties and the physical environment. Accessing and understanding the interplay between these elements requires the development of integrative and morphologically realistic neuromechanical simulations. Here we present NeuroMechFly, a data-driven model of the widely studied organism, Drosophila melanogaster. NeuroMechFly combines four independent computational modules: a physics-based simulation environment, a biomechanical exoskeleton, muscle models and neural network controllers. To enable use cases, we first define the minimum degrees of freedom of the leg from real three-dimensional kinematic m…
-
Mapping model units to visual neurons reveals population code for social behaviourBenjamin R. Cowley, Adam J. Calhoun, Nivedita Rangarajan et al. · 2024 · Nature · cited 48 times · Not summarized yet
Mapping model units to visual neurons links neural-network units to fly visual neurons through perturbations and behavioural predictions, a related sensorimotor modelling approach.
Abstract
The rich variety of behaviours observed in animals arises through the interplay between sensory processing and motor control. To understand these sensorimotor transformations, it is useful to build models that predict not only neural responses to sensory input1–5 but also how each neuron causally contributes to behaviour6,7. Here we demonstrate a novel modelling approach to identify a one-to-one mapping between internal units in a deep neural network and real neurons by predicting the behavioural changes that arise from systematic perturbations of more than a dozen neuronal cell types. A key ingredient that we introduce is ‘knockout training’, which involves perturbing the network during tra…
-
The digital sphinx: Can a worm brain control a fly body?BW Brunton, Elliott T. T. Abe, L. Hu et al. · 2026 · bioRxiv · preprint, not peer reviewed · cited 8 times · Not summarized yet
The digital sphinx discusses embodied simulations that couple nervous systems to a fly body, including the challenge of unknown biological parameters.
Abstract
Animal intelligence is not purely a product of abstract computation in the brain, but emerges from dynamic interactions between the nervous system and the body. New connectome datasets and musculoskeletal models now enable integrated, closed-loop simulations of the neural and biomechanical systems of the fruit fly Drosophila, an ideal model organism to investigate embodied intelligence. However, many biological parameters of the nervous system and the body, as well as how they interface, remain unknown. To fill such gaps, researchers are turning to deep reinforcement learning (DRL), a data-driven optimization framework, to create virtual animals that imitate the behavior of real animals. Her…
-
Connectome simulations identify a central pattern generator circuit for fly walkingSarah M. Pugliese, Grant M. Chou, Elliott T. T. Abe et al. · 2026 · bioRxiv · preprint, not peer reviewed · cited 16 times · Not summarized yet
Connectome simulations identify a central pattern generator circuit for fly walking simulates fly connectomes to identify neural drivers of motor activity.
Abstract
Animal locomotion relies on rhythmic body movements driven by central pattern generators (CPGs): neural circuits that produce oscillating output without oscillating input. However, the circuit structure of a CPG for walking is not known in any animal. To identify the cells and synapses that underlie rhythmic leg movement in walking flies, we developed dynamic simulations of the Drosophila ventral nerve cord (VNC) connectomes. A computational activation screen of descending neurons from the central brain identified DNg100—a known command neuron for walking—as the top driver of rhythmic leg motor activity. Simulated network pruning isolated a minimal rhythm-generating circuit consisting of one…
Go further
How are these papers found?
We compare the title and abstract of this paper with those of every paper summarized on this site, and with the papers Semantic Scholar recommends, those it cites and those citing it (for a much-cited paper, its most recent citations and, from OpenAlex, the most cited ones). Among the prior work and the follow-ups, the papers that use each other in depth, a method it uses and the much-cited follow-ups come a little higher. The levels (very close, close, related) come from the similarity of the texts, not from a judgment of quality; * marks a paper without an abstract, compared by its title. The sentence under a paper is written by an AI from the two abstracts: check it in the papers. A paper marked not relevant by three readers leaves the list.
Updated on 2 November 2026.
Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.