Biology subjects

Ramdya, P.

Publications and source records attributed to Ramdya, P..

17 recordsLinked to original sources

bioRxiv2026

Object manipulation and affordance learning in Drosophila

Animals must manipulate objects to perform tasks like pushing away debris while navigating over complex, natural terrain. For previously unseen objects, efficient manipulation requires that their affordances--the possible actions one can perform upon them--first be learned through experience. However, the behavioral and neural mechanisms underlying the learning of object affordances remain largely unknown. To address this gap, we show that adult flies can learn to push small spherical objects without being given any explicit reward. To do this, flies appear to learn the balls pushability affordance: pushing is delayed when animals are first exposed to an immobile ball, and manipulating one ball accelerates pushing of a second one in a new context. Behavioral quantification of a largescale neural silencing screen reveals that specific visual projection neurons and olfactory sensory neurons regulate initial reactions to the object while dopaminergic neurons and circuits in the mushroom bodies, a center for associative learning and memory in insects, are critical for learning object affordances. These findings open the door to a mechanistic understanding of object manipulation and affordance learning.

neuroscience↗

bioRxiv2026

Precise kinematic and muscle recording in freely behaving flies enabled by closed-loop tracking and annotation-free pose estimation

Understanding the neuromuscular basis for behavior requires measuring both kinematic and physiological data at high resolution in unconstrained conditions: a technically challenging goal. Here we present an integrated experimental-computational pipeline for measuring and quantifying body part kinematics and muscle activity in freely behaving Drosophila melanogaster. We first present Spotlight, a closed-loop videography system that performs real-time tracking to record untethered flies at high resolution (6 {micro}m/pixel) and high frame rate (330 Hz) while also enabling optical recordings of limb muscle activity via a fluorescent calcium reporter. To analyze these massive datasets without manual image annotation, we introduce PoseForge, a synthetic-data-driven framework that exploits morphologically accurate biomechanical simulations to generate synthetic data, and contrastive self-supervised learning to infer 3D keypoints and dense body-part segmentation from a single camera view. Using resulting 3D kinematic data, we can replay recorded behaviors in a biomechanical digital twin, NeuroMechFly, to infer forces generated and experienced by the flys limbs. Finally, we illustrate the capability of our system to optically record muscle activity. We show how the legs long-tendon muscles activate upon mechanical vibration, possibly to activate gripping and to maintain a stable posture. Taken together, this workflow enables scalable, high-resolution measurement and modeling of unconstrained, natural behavior.

neuroscience↗

bioRxiv2024

Centralized brain networks underlie body part coordination during grooming

Animals must coordinate multiple body parts to perform important tasks such as grooming, or locomotion. How this movement synchronization is achieved by the nervous system remains largely unknown. Here, we uncover the neural basis of body part coordination during goal-directed antennal grooming in the fly, Drosophila melanogaster. We find that unilateral or bilateral grooming of one or both antenna, respectively, arises from synchronized movements of the head, antennae, and forelegs. Simulated replay of these body part kinematics in a biomechanical model shows that this coordination makes grooming more efficient by permitting unobstructed, forceful collisions between the foreleg tibiae and antennae. Movements of one body part do not require proprioceptive sensory feedback from the others: neither amputation of the forelegs or antennae, nor immobilization of the head prevented movements of the other unperturbed body parts. By constructing a comprehensive antennal grooming network from the fly brain connectome, we find that centralized interneurons and shared premotor neurons interconnect and thus likely synchronize neck, antennal, and foreleg motor networks. A simulated activation screen of neurons in this network reveals cell classes required for the coordination of antennal movements during unilateral grooming. These cells form two coupled circuit motifs that enable robust body part synchronization: a recurrent excitatory subnetwork that promotes contralateral antennal pitch and broadcast inhibition that suppresses ipsilateral antennal pitch. Similarly centralized controllers may enable the flexible co-recruitment of multiple body parts to subserve a variety of behaviors.

neuroscience↗

bioRxiv2024

Conspecific sociability is regulated by associative learning circuits

Sociability, an animals ease and propensity to interact with members of its own species, is a prerequisite for many important social interactions including courtship, mating, brood rearing, and collective behavior. Despite its fundamental nature, the manner through which sociability arises--whether it is innate or acquired and its neural mechanisms-- remains unknown. Here, we found that the fly, Drosophila melanogaster, produces a constellation of fearful reactions when it encounters another fly for the first time. However, these animals become sociable following several hours of exposure to conspecific (but not heterospecific) odors. Two large-scale neural silencing screens of 188 brain cell types revealed that both initial fearful reactions and learned sociability depend upon overlapping networks including circuits in the mushroom body, the principal center for associative learning and memory in the insect brain. Functional recordings of key mushroom body output neurons (MBONs) from this network over two hours of social interactions support a mechanistic model whereby fly odors modulate social valence by rebalancing MBON population activity, biasing action selection, and thereby driving sociable rather than fearful responses towards other flies. Thus, a center for learning and memory plays a fundamental role in establishing the basis for most social interactions.

neuroscience↗

bioRxiv2024

Modeling conditional distributions of neural and behavioral data with masked variational autoencoders

Extracting the relationship between high-dimensional recordings of neural activity and complex behav- ior is a ubiquitous problem in systems neuroscience. Toward this goal, encoding and decoding models attempt to infer the conditional distribution of neural activity given behavior and vice versa, while dimensionality reduc- tion techniques aim to extract interpretable low-dimensional representations. Variational autoencoders (VAEs) are flexible deep-learning models commonly used to infer low-dimensional embeddings of neural or behavioral data. However, it is challenging for VAEs to accurately model arbitrary conditional distributions, such as those encountered in neural encoding and decoding, and even more so simultaneously. Here, we present a VAE-based approach for accurately calculating such conditional distributions. We validate our approach on a task with known ground truth and demonstrate the applicability to high-dimensional behavioral time series by retrieving the condi- tional distributions over masked body parts of walking flies. Finally, we probabilistically decode motor trajectories from neural population activity in a monkey reach task and query the same VAE for the encoding distribution of neural activity given behavior. Our approach provides a unifying perspective on joint dimensionality reduction and learning conditional distributions of neural and behavioral data, which will allow for scaling common analyses in neuroscience to todays high-dimensional multi-modal datasets.

neuroscience↗

bioRxiv2023

NeuroMechFly 2.0, a framework for simulating embodied sensorimotor control in adult Drosophila

Discovering principles underlying the control of animal behavior requires a tight dialogue between experiments and neuromechanical models. Until now, such models, including NeuroMechFly for the adult fly, Drosophila melanogaster, have primarily been used to investigate motor control. Far less studied with realistic body models is how the brain and motor systems work together to perform hierarchical sensorimotor control. Here we present NeuroMechFly v2, a framework that expands Drosophila neuromechanical modeling by enabling visual and olfactory sensing, ascending motor feedback, and complex terrains that can be navigated using leg adhesion. We illustrate its capabilities by first constructing biologically inspired locomotor controllers that use ascending motor feedback to perform path integration and head stabilization. Then, we add visual and olfactory sensing to this controller and train it using reinforcement learning to perform a multimodal navigation task in closed loop. Finally, we illustrate more biorealistic modeling in two ways: our model navigates a complex odor plume using a Drosophila odor taxis strategy, and it uses a connectome-constrained visual system network to follow another simulated fly. With this framework, NeuroMechFly can be used to accelerate the discovery of explanatory models of the nervous system and to develop machine learning-based controllers for autonomous artificial agents and robots.

neuroscience↗

bioRxiv2023

Networks of descending neurons transform command-like signals into population-based behavioral control

To transform 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 behaviors--the circuit mechanisms for which remain unclear. Here, we show that command-like DNs in Drosophila directly recruit networks of additional DNs to orchestrate flexible behaviors. Specifically, we found that optogenetic activation of command-like DNs previously thought to drive behaviors alone in fact co-activate larger populations of DNs. Connectome analysis revealed that this functional recruitment can be explained by direct excitatory connections between command-like DNs and networks of interconnected DNs in the brain. The size of downstream DN networks is predictive of whether descending population recruitment is necessary to generate a complete behavior: DNs with many downstream descending partners require network recruitment to drive flexible behaviors, while neurons with fewer partners can alone drive stereotyped behaviors and simple movements. Finally, DN networks reside within behavior-specific clusters that inhibit one another. These results support a mechanism for command-like descending control whereby a continuum of stereotyped to flexible behaviors are generated through the recruitment of increasingly large DN networks which likely construct a complete behavior by combining multiple motor subroutines.

neuroscience↗

bioRxiv2022

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 accomplished 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. 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 analyzing their spatial, functional, and morphological properties. These results set the stage for a comprehensive, population-level understanding of how the brains descending signals regulate complex motor behaviors.

neuroscience↗

bioRxiv2022

Ascending neurons convey behavioral state to integrative sensory and action selection centers in the brain

Knowledge of ones own behavioral state--whether one is walking, grooming, or resting--is critical for contextualizing sensory cues including interpreting visual motion and tracking odor sources. Additionally, awareness of ones own posture is important to avoid initiating destabilizing or physically impossible actions. Ascending neurons (ANs), interneurons in the vertebrate spinal cord or insect ventral nerve cord (VNC) that project to the brain, may provide such high-fidelity behavioral state signals. However, little is known about what ANs encode and where they convey signals in any brain. To address this gap, we performed a large-scale functional screen of AN movement encoding, brain targeting, and motor system patterning in the adult fly, Drosophila melanogaster. Using a new library of AN sparse driver lines, we measured the functional properties of 247 genetically-identifiable ANs by performing two-photon microscopy recordings of neural activity in tethered, behaving flies. Quantitative, deep network-based neural and behavioral analyses revealed that ANs nearly exclusively encode high-level behaviors--primarily walking as well as resting and grooming--rather than low-level joint or limb movements. ANs that convey self-motion--resting, walking, and responses to gust-like puff stimuli--project to the brains anterior ventrolateral protocerebrum (AVLP), a multimodal, integrative sensory hub, while those that encode discrete actions--eye grooming, turning, and proboscis extension--project to the brains gnathal ganglion (GNG), a locus for action selection. The structure and polarity of AN projections within the VNC are predictive of their functional encoding and imply that ANs participate in motor computations while also relaying state signals to the brain. Illustrative of this are ANs that temporally integrate proboscis extensions over tens-of-seconds, likely through recurrent interconnectivity. Thus, in line with long-held theoretical predictions, ascending populations convey high-level behavioral state signals almost exclusively to brain regions implicated in sensory feature contextualization and action selection.

neuroscience↗

bioRxiv2021

Long-term imaging of the ventral nerve cord in behaving adult Drosophila

The dynamics and connectivity of neural circuits continuously change during an animals lifetime on timescales ranging from milliseconds to days. Therefore, to investigate how biological networks accomplish remarkable cognitive and behavioral tasks, minimally invasive methods are needed to perform repeated measurements, or perturbations of neural circuits in behaving animals across time. Such tools have been developed to investigate the brain but similar approaches are lacking for comprehensively and repeatedly recording motor circuits in behaving animals. Here we describe a suite of microfabricated technologies that enable long-term, minimally invasive optical recordings of the adult Drosophila melanogaster ventral nerve cord (VNC)--neural tissues that are functionally equivalent to the vertebrate spinal cord. These tools consist of (i) a manipulator arm that permits the insertion of (ii) a compliant implant into the thorax to expose the imaging region of interest; (iii) a numbered, transparent polymer window that encloses and provides optical access to the inside of the thorax, and (iv) a hinged remounting stage that allows gentle and repeated tethering of an implanted animal for two-photon imaging. We validate and illustrate the utility of our toolkit in several ways. First, we show that the thoracic implant and window have minimal impact on animal behavior and survival while also enabling neural recordings from individual animals across at least one month. Second, we follow the degradation of chordotonal organ mechanosensory nerve terminals in the VNC over weeks after leg amputation. Third, because our tools allow recordings of the VNC with the gut intact, we discover waves of neural population activity following ingestion of a high-concentration caffeine solution. In summary, our microfabricated toolkit makes it possible to longitudinally monitor anatomical and functional changes in premotor and motor neural circuits, and more generally opens up the long-term investigation of thoracic tissues.

neuroscience↗

bioRxiv2021

NeuroMechFly, a neuromechanical model of adult Drosophila melanogaster

Animal behavior emerges from a seamless interaction between neural network dynamics, musculoskeletal properties, and the physical environment. Accessing and understanding the interplay between these intertwined elements requires the development of integrative and morphologically realistic neuromechanical simulations. Until now, there has been no such simulation framework for the widely studied model organism, Drosophila melanogaster. Here we present NeuroMech-Fly, a data-driven model of the adult female fly within a physics-based simulation environment. NeuroMechFly combines a series of independent computational modules including a biomechanical exoskeleton with articulating body parts-legs, halteres, wings, abdominal segments, head, proboscis, and antennae-muscle models, and neural network controllers. To enable illustrative use cases, we first define minimal leg degrees-of-freedom by analyzing real 3D kinematic measurements during real Drosophila walking and grooming. Then, we show how, by replaying these behaviors using NeuroMechFlys biomechanical exoskeleton in its physics-based simulation environment, one can predict otherwise unmeasured torques and contact reaction forces. Finally, we leverage NeuroMechFlys full neuromechanical capacity to discover neural networks and muscle parameters that enable locomotor gaits optimized for speed and stability. Thus, NeuroMechFly represents a powerful testbed for building an understanding of how behaviors emerge from interactions between complex neuromechanical systems and their physical surroundings.

neuroscience↗

bioRxiv2020

LiftPose3D, a deep learning-based approach for transforming 2D to 3D pose in laboratory animals

Markerless 3D pose estimation has become an indispensable tool for kinematic studies of laboratory animals. Most current methods recover 3D pose by multi-view triangulation of deep network-based 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 LiftPose3Ds versatility by applying it to multiple experimental systems using flies, mice, rats, and macaque monkeys and in circumstances where 3D triangulation is impractical or impossible. Our framework achieves accurate lifting for stereotyped and non-stereotyped behaviors from different camera angles. Thus, LiftPose3D permits high-quality 3D pose estimation in the absence of complex camera arrays, tedious calibration procedures, and despite occluded body parts in freely behaving animals.

animal behavior and cognition↗

bioRxiv2019

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 rapidly adapt to changing conditions, and respond to sudden and surprising new stimuli. Neuromodulators, particularly the small biogenic amine neurotransmitters, allow motor circuits to rapidly alter their output by changing their functional connectivity. Here we show that the serotonergic system in the vinegar fly, Drosophilamelanogaster, can modulate walking speed in a variety of contexts and in response 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.

neuroscience↗

bioRxiv2019

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 3-dimensional (3D) space. Deep neural networks can estimate 2-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--or other animals--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 behavioral measurements at an unprecedented level of resolution for a variety of biological applications.

animal behavior and cognition↗

bioRxiv2019

Patterned cell death sculpts functional and adaptive neural networks in flies

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 appears to play a key role. We show that cell death occurs continuously throughout insect neurogenesis and happens soon after neurons are born. 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. Mimicking an evolutionary role for increasing cell numbers, we artificially block programmed cell death 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. Our findings suggest that the evolutionary modulation of death-based patterning could generate novel network configurations.

neuroscience↗

bioRxiv2018

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 the portion of the nervous system that contains limb premotor and motor circuits is largely inaccessible to large-scale recording techniques 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 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 novel class of locomotion-associated descending neurons. This new approach enables the direct investigation of circuits associated with complex limb movements.

neuroscience↗

bioRxiv2016

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 precise, quantitative, and high-throughput 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 Drosophila. We show that this approach is robust to wide variations in video spatial and temporal resolution and that it can be used to measure leg segment motions during 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.\n\nAuthor SummaryIn terrestrial animals, including humans, fundamental actions like locomotion and grooming emerge from the displacement of multiple limbs through space. Therefore, precise measurements of limb movements are critical for investigating and, ultimately, understanding the neural basis for behavior. The vinegar fly, Drosophila melanogaster, is an attractive animal model for uncovering general principles about limb control since its genome and nervous system are easy to manipulate. However, existing methods for measuring leg movements in freely behaving Drosophila have significant drawbacks: they require complicated experimental setups and provide limited information about each leg. Here we report a new method - and provide its open-source software implementation, FlyLimbTracker - for tracking the body and leg segments of freely behaving flies using only computational image processing approaches. We illustrate the power of this method by tracking fly limbs during five distinct walking and grooming behaviors and from videos across a wide range of spatial and temporal resolutions. Our approach is generalizable, allowing researchers to use and customize our software for limb tracking in Drosophila and in other species.

bioinformatics↗