Descending networks transform command signals into population motor control.

To convert intentions into actions, movement instructions must pass from the brain to downstream motor circuits through descending neurons (DNs). These include small sets of command-like neurons that are sufficient to drive behaviours1-the circuit mechanisms for which remain unclear. Here we show that command-like DNs in Drosophila directly recruit networks of additional DNs to orchestrate behaviours that require the active control of numerous body parts. Specifically, we found that command-like DNs previously thought to drive behaviours alone2-4 in fact co-activate larger populations of DNs. Connectome analyses and experimental manipulations revealed that this functional recruitment can be explained by direct excitatory connections between command-like DNs and networks of interconnected DNs in the brain. Descending population recruitment is necessary for behavioural control: DNs with many downstream descending partners require network co-activation to drive complete behaviours and drive only simple stereotyped movements in their absence. These DN networks reside within behaviour-specific clusters that inhibit one another. These results support a mechanism for command-like descending control in which behaviours are generated through the recruitment of increasingly large DN networks that compose behaviours by combining multiple motor subroutines.

The neuromechanics of animal locomotion: From biology to robotics and back.

Robotics and neuroscience are sister disciplines that both aim to understand how agile, efficient, and robust locomotion can be achieved in autonomous agents. Robotics has already benefitted from neuromechanical principles discovered by investigating animals. These include the use of high-level commands to control low-level central pattern generator-like controllers, which, in turn, are informed by sensory feedback. Reciprocally, neuroscience has benefited from tools and intuitions in robotics to reveal how embodiment, physical interactions with the environment, and sensory feedback help sculpt animal behavior. We illustrate and discuss exemplar studies of this dialog between robotics and neuroscience. We also reveal how the increasing biorealism of simulations and robots is driving these two disciplines together, forging an integrative science of autonomous behavioral control with many exciting future opportunities.

Ascending neurons convey behavioral state to integrative sensory and action selection brain regions.

Knowing one's own behavioral state has long been theorized as critical for contextualizing dynamic sensory cues and identifying appropriate future behaviors. Ascending neurons (ANs) in the motor system that project to the brain are well positioned to provide such behavioral state signals. However, what ANs encode and where they convey these signals remains largely unknown. Here, through large-scale functional imaging in behaving animals and morphological quantification, we report the behavioral encoding and brain targeting of hundreds of genetically identifiable ANs in the adult fly, Drosophila melanogaster. We reveal that ANs encode behavioral states, specifically conveying self-motion to the anterior ventrolateral protocerebrum, an integrative sensory hub, as well as discrete actions to the gnathal ganglia, a locus for action selection. Additionally, AN projection patterns within the motor system are predictive of their encoding. Thus, ascending populations are well poised to inform distinct brain hubs of self-motion and ongoing behaviors and may provide an important substrate for computations that are required for adaptive behavior.

Descending neuron population dynamics during odor-evoked and spontaneous limb-dependent behaviors.

Deciphering how the brain regulates motor circuits to control complex behaviors is an important, long-standing challenge in neuroscience. In the fly, Drosophila melanogaster, this is coordinated by a population of ~ 1100 descending neurons (DNs). Activating only a few DNs is known to be sufficient to drive complex behaviors like walking and grooming. However, what additional role the larger population of DNs plays during natural behaviors remains largely unknown. For example, they may modulate core behavioral commands or comprise parallel pathways that are engaged depending on sensory context. We evaluated these possibilities by recording populations of nearly 100 DNs in individual tethered flies while they generated limb-dependent behaviors, including walking and grooming. We found that the largest fraction of recorded DNs encode walking while fewer are active during head grooming and resting. A large fraction of walk-encoding DNs encode turning and far fewer weakly encode speed. Although odor context does not determine which behavior-encoding DNs are recruited, a few DNs encode odors rather than behaviors. Lastly, we illustrate how one can identify individual neurons from DN population recordings by using their spatial, functional, and morphological properties. These results set the stage for a comprehensive, population-level understanding of how the brain's descending signals regulate complex motor actions.

Microengineered devices enable long-term imaging of the ventral nerve cord in behaving adult Drosophila.

The dynamics and connectivity of neural circuits continuously change on timescales ranging from milliseconds to an animal's lifetime. Therefore, to understand biological networks, minimally invasive methods are required to repeatedly record them in behaving animals. Here we describe a suite of devices that enable long-term optical recordings of the adult Drosophila melanogaster ventral nerve cord (VNC). These consist of transparent, numbered windows to replace thoracic exoskeleton, compliant implants to displace internal organs, a precision arm to assist implantation, and a hinged stage to repeatedly tether flies. To validate and illustrate our toolkit we (i) show minimal impact on animal behavior and survival, (ii) follow the degradation of chordotonal organ mechanosensory nerve terminals over weeks after leg amputation, and (iii) uncover waves of neural activity caffeine ingestion. Thus, our long-term imaging toolkit opens up the investigation of premotor and motor circuit adaptations in response to injury, drug ingestion, aging, learning, and disease.

NeuroMechFly, a neuromechanical model of adult Drosophila melanogaster.

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 measurements during walking and grooming. Then, we show how, by replaying these behaviors in the simulator, one can predict otherwise unmeasured torques and contact forces. Finally, we leverage NeuroMechFly's full neuromechanical capacity to discover neural networks and muscle parameters that drive locomotor gaits optimized for speed and stability. Thus, NeuroMechFly can increase our understanding of how behaviors emerge from interactions between complex neuromechanical systems and their physical surroundings.

Connecting the dots in ethology: applying network theory to understand neural and animal collectives.

A major goal shared by neuroscience and collective behavior is to understand how dynamic interactions between individual elements give rise to behaviors in populations of neurons and animals, respectively. This goal has recently become within reach, thanks to techniques providing access to the connectivity and activity of neuronal ensembles as well as to behaviors among animal collectives. The next challenge using these datasets is to unravel network mechanisms generating population behaviors. This is aided by network theory, a field that studies structure-function relationships in interconnected systems. Here we review studies that have taken a network view on modern datasets to provide unique insights into individual and collective animal behaviors. Specifically, we focus on how analyzing signal propagation, controllability, symmetry, and geometry of networks can tame the complexity of collective system dynamics. These studies illustrate the potential of network theory to accelerate our understanding of behavior across ethological scales.

LiftPose3D, a deep learning-based approach for transforming two-dimensional to three-dimensional poses in laboratory animals.

Markerless three-dimensional (3D) pose estimation has become an indispensable tool for kinematic studies of laboratory animals. Most current methods recover 3D poses by multi-view triangulation of deep network-based two-dimensional (2D) pose estimates. However, triangulation requires multiple synchronized cameras and elaborate calibration protocols that hinder its widespread adoption in laboratory studies. Here we describe LiftPose3D, a deep network-based method that overcomes these barriers by reconstructing 3D poses from a single 2D camera view. We illustrate LiftPose3D's versatility by applying it to multiple experimental systems using flies, mice, rats and macaques, and in circumstances where 3D triangulation is impractical or impossible. Our framework achieves accurate lifting for stereotypical and nonstereotypical behaviors from different camera angles. Thus, LiftPose3D permits high-quality 3D pose estimation in the absence of complex camera arrays and tedious calibration procedures and despite occluded body parts in freely behaving animals.

Extensive and diverse patterns of cell death sculpt neural networks in insects.

Changes to the structure and function of neural networks are thought to underlie the evolutionary adaptation of animal behaviours. Among the many developmental phenomena that generate change programmed cell death (PCD) appears to play a key role. We show that cell death occurs continuously throughout insect neurogenesis and happens soon after neurons are born. Mimicking an evolutionary role for increasing cell numbers, we artificially block in the medial neuroblast lineage in Drosophila melanogaster, which results in the production of 'undead' neurons with complex arborisations and distinct neurotransmitter identities. Activation of these 'undead' neurons and recordings of neural activity in behaving animals demonstrate that they are functional. Focusing on two dipterans, which have lost flight during evolution, we reveal that reductions in populations of flight interneurons are likely caused by increased cell death during development. Our findings suggest that the evolutionary modulation of death-based patterning could generate novel network configurations.

Just like a sculptor chips away at a block of granite to make a statue, the nervous system reaches its mature state by eliminating neurons during development through a process known as programmed cell death. In vertebrates, this mechanism often involves newly born neurons shrivelling away and dying if they fail to connect with others during development. Most studies in insects have focused on the death of neurons that occurs at metamorphosis, during the transition between larva to adult, when cells which are no longer needed in the new life stage are eliminated. Pop et al. harnessed a newly designed genetic probe to point out that, in fruit flies, programmed cell death of neurons at metamorphosis is not the main mechanism through which cells die. Rather, the majority of cell death takes place as soon as neurons are born throughout all larval stages, when most of the adult nervous system is built. To gain further insight into the role of this 'early' cell death, the neurons were stopped from dying, showing that these cells were able to reach maturity and function. Together, these results suggest that early cell death may be a mechanism fine-tuned by evolution to shape the many and varied nervous systems of insects. To explore this, Pop et al. looked for hints of early cell death in relatives of fruit flies that are unable to fly: the swift lousefly and the bee lousefly. This analysis showed that early cell death is likely to occur in these two insects, but it follows different patterns than in the fruit fly, potentially targeting the neurons that would have controlled flight in these flies' ancestors. Brains are the product of evolution: learning how neurons change their connections and adapt could help us understand how the brain works in health and disease. This knowledge may also be relevant to work on artificial intelligence, a discipline that often bases the building blocks and connections in artificial 'brains' on how neurons communicate with one another.

Serotonergic Modulation of Walking in Drosophila.

To navigate complex environments, animals must generate highly robust, yet flexible, locomotor behaviors. For example, walking speed must be tailored to the needs of a particular environment. Not only must animals choose the correct speed and gait, they must also adapt to changing conditions and quickly respond to sudden and surprising new stimuli. Neuromodulators, particularly the small biogenic amine neurotransmitters, have the ability to rapidly alter the functional outputs of motor circuits. Here, we show that the serotonergic system in the vinegar fly, Drosophila melanogaster, can modulate walking speed in a variety of contexts and also change how flies respond to sudden changes in the environment. These multifaceted roles of serotonin in locomotion are differentially mediated by a family of serotonergic receptors with distinct activities and expression patterns.

DeepFly3D, a deep learning-based approach for 3D limb and appendage tracking in tethered, adult Drosophila.

Studying how neural circuits orchestrate limbed behaviors requires the precise measurement of the positions of each appendage in three-dimensional (3D) space. Deep neural networks can estimate two-dimensional (2D) pose in freely behaving and tethered animals. However, the unique challenges associated with transforming these 2D measurements into reliable and precise 3D poses have not been addressed for small animals including the fly, Drosophila melanogaster. Here, we present DeepFly3D, a software that infers the 3D pose of tethered, adult Drosophila using multiple camera images. DeepFly3D does not require manual calibration, uses pictorial structures to automatically detect and correct pose estimation errors, and uses active learning to iteratively improve performance. We demonstrate more accurate unsupervised behavioral embedding using 3D joint angles rather than commonly used 2D pose data. Thus, DeepFly3D enables the automated acquisition of Drosophila behavioral measurements at an unprecedented level of detail for a variety of biological applications.

Imaging neural activity in the ventral nerve cord of behaving adult Drosophila.

To understand neural circuits that control limbs, one must measure their activity during behavior. Until now this goal has been challenging, because limb premotor and motor circuits have been largely inaccessible for large-scale recordings in intact, moving animals-a constraint that is true for both vertebrate and invertebrate models. Here, we introduce a method for 2-photon functional imaging from the ventral nerve cord (VNC) of behaving adult Drosophila melanogaster. We use this method to reveal patterns of activity across nerve cord populations during grooming and walking and to uncover the functional encoding of moonwalker ascending neurons (MANs), moonwalker descending neurons (MDNs), and a previously uncharacterized class of locomotion-associated A1 descending neurons. Finally, we develop a genetic reagent to destroy the indirect flight muscles and to facilitate experimental access to the VNC. Taken together, these approaches enable the direct investigation of circuits associated with complex limb movements.

FlyLimbTracker: An active contour based approach for leg segment tracking in unmarked, freely behaving Drosophila.

Understanding the biological underpinnings of movement and action requires the development of tools for quantitative measurements of animal behavior. Drosophila melanogaster provides an ideal model for developing such tools: the fly has unparalleled genetic accessibility and depends on a relatively compact nervous system to generate sophisticated limbed behaviors including walking, reaching, grooming, courtship, and boxing. Here we describe a method that uses active contours to semi-automatically track body and leg segments from video image sequences of unmarked, freely behaving D. melanogaster. We show that this approach yields a more than 6-fold reduction in user intervention when compared with fully manual annotation and can be used to annotate videos with low spatial or temporal resolution for a variety of locomotor and grooming behaviors. FlyLimbTracker, the software implementation of this method, is open-source and our approach is generalizable. This opens up the possibility of tracking leg movements in other species by modifications of underlying active contour models.

Climbing favours the tripod gait over alternative faster insect gaits.

To escape danger or catch prey, running vertebrates rely on dynamic gaits with minimal ground contact. By contrast, most insects use a tripod gait that maintains at least three legs on the ground at any given time. One prevailing hypothesis for this difference in fast locomotor strategies is that tripod locomotion allows insects to rapidly navigate three-dimensional terrain. To test this, we computationally discovered fast locomotor gaits for a model based on Drosophila melanogaster. Indeed, the tripod gait emerges to the exclusion of many other possible gaits when optimizing fast upward climbing with leg adhesion. By contrast, novel two-legged bipod gaits are fastest on flat terrain without adhesion in the model and in a hexapod robot. Intriguingly, when adhesive leg structures in real Drosophila are covered, animals exhibit atypical bipod-like leg coordination. We propose that the requirement to climb vertical terrain may drive the prevalence of the tripod gait over faster alternative gaits with minimal ground contact.

The neurogenetics of group behavior in Drosophila melanogaster.

Organisms rarely act in isolation. Their decisions and movements are often heavily influenced by direct and indirect interactions with conspecifics. For example, we each represent a single node within a social network of family and friends, and an even larger network of strangers. This group membership can affect our opinions and actions. Similarly, when in a crowd, we often coordinate our movements with others like fish in a school, or birds in a flock. Contributions of the group to individual behaviors are observed across a wide variety of taxa but their biological mechanisms remain largely unknown. With the advent of powerful computational tools as well as the unparalleled genetic accessibility and surprisingly rich social life of Drosophila melanogaster, researchers now have a unique opportunity to investigate molecular and neuronal determinants of group behavior. Conserved mechanisms and/or selective pressures in D. melanogaster can likely inform a much wider phylogenetic scale. Here, we highlight two examples to illustrate how quantitative and genetic tools can be combined to uncover mechanisms of two group behaviors in D. melanogaster: social network formation and collective behavior. Lastly, we discuss future challenges towards a full understanding how coordinated brain activity across many individuals gives rise to the behavioral patterns of animal societies.

Fluctuation-Driven Neural Dynamics Reproduce Drosophila Locomotor Patterns.

The neural mechanisms determining the timing of even simple actions, such as when to walk or rest, are largely mysterious. One intriguing, but untested, hypothesis posits a role for ongoing activity fluctuations in neurons of central action selection circuits that drive animal behavior from moment to moment. To examine how fluctuating activity can contribute to action timing, we paired high-resolution measurements of freely walking Drosophila melanogaster with data-driven neural network modeling and dynamical systems analysis. We generated fluctuation-driven network models whose outputs-locomotor bouts-matched those measured from sensory-deprived Drosophila. From these models, we identified those that could also reproduce a second, unrelated dataset: the complex time-course of odor-evoked walking for genetically diverse Drosophila strains. Dynamical models that best reproduced both Drosophila basal and odor-evoked locomotor patterns exhibited specific characteristics. First, ongoing fluctuations were required. In a stochastic resonance-like manner, these fluctuations allowed neural activity to escape stable equilibria and to exceed a threshold for locomotion. Second, odor-induced shifts of equilibria in these models caused a depression in locomotor frequency following olfactory stimulation. Our models predict that activity fluctuations in action selection circuits cause behavioral output to more closely match sensory drive and may therefore enhance navigation in complex sensory environments. Together these data reveal how simple neural dynamics, when coupled with activity fluctuations, can give rise to complex patterns of animal behavior.

Mechanosensory interactions drive collective behaviour in Drosophila.

Collective behaviour enhances environmental sensing and decision-making in groups of animals. Experimental and theoretical investigations of schooling fish, flocking birds and human crowds have demonstrated that simple interactions between individuals can explain emergent group dynamics. These findings indicate the existence of neural circuits that support distributed behaviours, but the molecular and cellular identities of relevant sensory pathways are unknown. Here we show that Drosophila melanogaster exhibits collective responses to an aversive odour: individual flies weakly avoid the stimulus, but groups show enhanced escape reactions. Using high-resolution behavioural tracking, computational simulations, genetic perturbations, neural silencing and optogenetic activation we demonstrate that this collective odour avoidance arises from cascades of appendage touch interactions between pairs of flies. Inter-fly touch sensing and collective behaviour require the activity of distal leg mechanosensory sensilla neurons and the mechanosensory channel NOMPC. Remarkably, through these inter-fly encounters, wild-type flies can elicit avoidance behaviour in mutant animals that cannot sense the odour--a basic form of communication. Our data highlight the unexpected importance of social context in the sensory responses of a solitary species and open the door to a neural-circuit-level understanding of collective behaviour in animal groups.

Fluorescence behavioral imaging (FBI) tracks identity in heterogeneous groups of Drosophila.

Distinguishing subpopulations in group behavioral experiments can reveal the impact of differences in genetic, pharmacological and life-histories on social interactions and decision-making. Here we describe Fluorescence Behavioral Imaging (FBI), a toolkit that uses transgenic fluorescence to discriminate subpopulations, imaging hardware that simultaneously records behavior and fluorescence expression, and open-source software for automated, high-accuracy determination of genetic identity. Using FBI, we measure courtship partner choice in genetically mixed groups of Drosophila.

Complementary function and integrated wiring of the evolutionarily distinct Drosophila olfactory subsystems.

To sense myriad environmental odors, animals have evolved multiple, large families of divergent olfactory receptors. How and why distinct receptor repertoires and their associated circuits are functionally and anatomically integrated is essentially unknown. We have addressed these questions through comprehensive comparative analysis of the Drosophila olfactory subsystems that express the ionotropic receptors (IRs) and odorant receptors (ORs). We identify ligands for most IR neuron classes, revealing their specificity for select amines and acids, which complements the broader tuning of ORs for esters and alcohols. IR and OR sensory neurons exhibit glomerular convergence in segregated, although interconnected, zones of the primary olfactory center, but these circuits are extensively interdigitated in higher brain regions. Consistently, behavioral responses to odors arise from an interplay between IR- and OR-dependent pathways. We integrate knowledge on the different phylogenetic and developmental properties of these receptors and circuits to propose models for the functional contributions and evolution of these distinct olfactory subsystems.