Biology subjects

Zlatic, M.

Publications and source records attributed to Zlatic, M..

At least 19 recordsLinked to original sources

bioRxiv2026

The female larval fly CNS connectome: first draft

We release an enhanced FIB-SEM volume and associated neuron, mitochondria, synapse and glia automated segmentations for the complete, gap- and artifact-free central nervous system of the first instar female larval Drosophila melanogaster. Segmentation data is largely unproofread besides select neurons in the premotor and proprioceptive system and select subvolumes where synapses have been manually annotated, serving as training and evaluation data. The dataset represents the first complete central nervous system of the larval fly, and now awaits substantial improvement of machine vision approaches that effectively leverage domain knowledge for the automated yet data-guided proofreading, or substantial and costly human proofreading. All the data and methods are publicly available for further improvement and analysis.

neuroscience↗

bioRxiv2025

Descending neurons integrate learnt information from mushroom body with context to promote escape behaviour

To behave adaptively in the environment and select appropriate responses to sensory stimuli, animals integrate learnt information about the valences of stimuli with contextual information. While significant progress has been made in understanding how animals learn which stimuli predict positive and negative outcomes, how the outputs of associative learning circuits flexibly promote different behaviours, depending on context, is not well understood. Addressing this question requires mapping the pathways from learning circuits to nerve cord command neurons that promote specific actions, and understanding where and how contextual information converges with these pathways to modulate their activity. These are daunting tasks in larger brains where synaptic-resolution connectivity maps are not available. We, therefore, addressed this question in the tractable Drosophila larva using a combination of connectomic analyses, imaging and manipulation of neural activity and behavioural analysis. We characterised, with synaptic-resolution, the pathways from the higher-order learning circuit (the mushroom body, MB) all the way to a specific cord command neuron (Goro) in the nerve cord that promotes the most vigorous escape response, rolling. Rolling is the fastest, but also the most energetically costly escape response and is activated by multisensory cues in the context of predator attack. We identify a pair of brain descending neurons, Ipsigoro, that integrate learnt information (via input from MBONs) and nociceptive context (via input from ascending neurons) to facilitate rolling via direct inputs to Goro. Our study reveals the circuit mechanism by which context and learnt information are integrated by brain descending neurons to activate specific nerve cord command neurons and promote specific actions.

neuroscience↗

bioRxiv2025

The central complex of the larval fruit fly brain

In holometabolous insects such as the fruit fly Drosophila melanogaster, the brain central complex (CX) develops during metamorphosis and serves the adult stage. Whether a form of the CX exists in the brain of the evolutionarily novel larval stages is not known. Here, we analyzed the connectome of the larval brain and, on the basis of neuronal lineages, synaptic connectivity patterns, and anatomy, identified a putative larval CX, comprising 4 key neuropils: the protocerebral bridge (PB), the ellipsoid body (EB), the fan-shaped body (FB) and the noduli (NO). Consistent with our interpretation, we found in the larval brain synaptic connectivity patterns characteristic of the adult, including (i) visual input into the PB and EB; (ii) modulation of CX neuropil inputs by the mushroom body (MB); (iii) reciprocal connectivity between CX neuropils and select MB compartments; and (iv) strong connectivity between CX neuropils. While some neuronal lineages contributing to the larval CX do not contribute to the adult CX, many others are conserved. The characterization of a larval CX brings structure to largely unexamined larval brain circuits, linking with a vast body of literature, and will inform the design of experiments to probe larval brain function.

neuroscience↗

bioRxiv2025

Beyond Agreement: Standardizing Crowdsourced Synapse Annotations through Proofreading in EM Connectomics

AO_SCPLOWBSTRACTC_SCPLOWReliable synapse identification in volumetric EM is hampered by subtle, 3D cues that yield variable human judgments. We present a standardized proofreading protocol that pairs explicit, operational criteria with machine-learning candidate generation and a two-stage calibration of annotators. In two larval Drosophila melanogaster volumes imaged at 8x8x8 nm, five raters (expert + 4 calibrated annotators) reviewed model-proposed candidates using efficient node-based labels. Multi-rater judgments were aggregated with a probabilistic Dawid-Skene (DS) model to produce consensus labels with calibrated uncertainty. Post-calibration, individual annotator accuracy versus the expert improved (McNemar p < 0.05 for all raters), DS-expert agreement increased, and DS posterior entropy decreased for true positives/negatives, indicating more decisive consensus; gains were modest and dataset-dependent in chance-corrected agreement (Krippendorffs ). By making uncertainty explicit, this protocol converts noisy judgments into auditable supervision suitable for training and evaluation, while honestly communicating residual ambiguity essential for reliable and robust connectomics at scale.

neuroscience↗

bioRxiv2025

A multisensory, bidirectional, valence encoder guides behavioral decisions

A key function of the brain is to categorize sensory cues as repulsive or attractive and respond accordingly. While we have some understanding of how sensory information is processed in the sensory periphery, the classification of cues according to valence in central brain circuits is less well understood. Here, we addressed this question in the Drosophila larva, where we could leverage the synaptic resolution connectome to determine where innate and learnt information from distinct aversive and appetitive sensory modalities converges; and combine this with imaging and manipulation of neural activity to determine how valence information is encoded and used to guide navigation. We found that information from multiple innately aversive and attractive sensory modalities converges onto a common output neuron of the learning circuit, specifically the Mushroom Body output neuron, MBON-m1. We discovered that this neuron is required for navigating both attractive odor and aversive temperature cues and is activated by attractive cues, such as food odour and sugar, and inhibited by different aversive cues, such as cooling, salt, or non-food odors. Together, our study reveals a neuron that bi-directionally encodes valence and controls actions.

neuroscience↗

bioRxiv2025

Combining brainwide activity imaging and electron microscopy reveals novel nociceptive circuits

To understand how brains work, it is necessary to connect neural activity to synaptic-resolution circuit architecture. Recent advances in light-sheet microscopy (LSM) enable whole-brain, cellular-resolution imaging of activity of all neuronal cell bodies, however, most neurons from such datasets cannot be identified. In most organisms, neurons are identifiable based on their projections (and not based on their cell body position) which, when densely labelled, cannot be resolved using LSM. Here, we present a novel methodology to overcome this by combining whole-brain activity imaging with subsequent volume electron microscopy imaging of the same brain to visualise neuronal projections and identify neurons with interesting activity. We used this approach to identify brain neurons that process input from multisensory (nociceptive and mechanosensory) Basin interneurons that trigger vigorous escape in Drosophila larvae in response to threatening somatosensory stimuli. After whole-brain imaging of neuronal activity during Basin activation, we imaged the same brain with an enhanced focused ion-beam electron microscope (eFIB-SEM). We registered the functional and anatomical volumes and reconstructed (in the eFIB-SEM volume) the projections of neurons that responded to Basin activation to determine their developmental lineage and identity. This revealed a distributed network for processing threatening somatosensory stimuli that trigger vigorous escape spanning 25 distinct lineages and many distinct brain areas, and included direct brain targets of Basin neurons that integrate somatosensory information with other modalities, as well as brain output neurons (descending neurons [DN]) that likely contribute to action-selection. Our workflow provides a powerful framework for mapping neuronal activity onto structure across an entire brain, yielding novel insights into the distributed central processing of noxious stimuli.

neuroscience↗

bioRxiv2025

Cross-species comparative connectomics reveals the evolution of an olfactory circuit

Animal behavioural diversity ultimately stems from variation in neural circuitry, yet how central neural circuits evolve remains poorly understood. Studies of neural circuit evolution often focus on a few elements within a network. However, addressing fundamental questions in evolutionary neuroscience, such as whether some elements are more evolvable than others, requires a more global and unbiased approach. Here, we used synapse-level comparative connectomics to examine how an entire olfactory circuit evolves. We compared the full antennal lobe connectome of the larvae of two closely related Drosophila species, D. melanogaster and D. erecta, which differ in their ecological niches and odour-driven behaviours. We found that evolutionary change is unevenly distributed across the network. Some features, including neuron types, neuron numbers and interneuron-to-interneuron connectivity, are highly conserved. These conserved elements delineate a core circuit blueprint presumably required for fundamental olfactory processing. Superimposed on this scaffold, we find rewiring changes that mirror each species ecologies, including a systematic shift in the excitation-to-inhibition balance in the feedforward pathways. We further show that some neurons have changed more than others, and that even within individual neurons some synaptic elements remain conserved while others display major species-specific changes, suggesting evolutionary hot-spots within the circuit. Our findings reveal constrained and adaptable elements within olfactory networks, and establish a framework for identifying general principles in the evolution of neural circuits underlying behaviour.

neuroscience↗

bioRxiv2024

Why do we have so many excitatory neurons?

The emerging electron microscopy connectome datasets provides connectivity maps of the brains at single cell resolution, enabling us to estimate various network statistics, such as connectedness. We desire the ability to assess how the functional complexity of these networks depends on these network statistics. To this end, we developed an analysis pipeline and a statistic, XORness, which quantifies the functional complexity of these networks with varying network statistics. We illustrate that actual connectomes have high XORness, as do generated connectomes with the same network statistics, suggesting a normative role for functional complexity in guiding the evolution of connectomes, and providing clues to guide the development of artificial neural networks.

neuroscience↗

bioRxiv2024

Statistical signature of subtle behavioural changes inlarge-scale behavioural assays

The central nervous system can generate various behaviours, including motor responses, which we can observe through video recordings. Recent advancements in genetics, automated behavioural acquisition at scale, and machine learning enable us to link behaviours to their underlying neural mechanisms causally. Moreover, in some animals, such as the Drosophila larva, this mapping is possible at unprecedented scales of millions of animals and single neurons, allowing us to identify the neural circuits generating particular behaviours. These high-throughput screening efforts are invaluable, linking the activation or suppression of specific neurons to behavioural patterns in millions of animals. This provides a rich dataset to explore how diverse nervous system responses can be to the same stimuli. However, challenges remain in identifying subtle behaviours from these large datasets, including immediate and delayed responses to neural activation or suppression, and understanding these behaviours on a large scale. We introduce several statistically robust methods for analyzing behavioural data in response to these challenges: 1) A generative physical model that regularizes the inference of larval shapes across the entire dataset. 2) An unsupervised kernel-based method for statistical testing in learned behavioural spaces aimed at detecting subtle deviations in behaviour. 3) A generative model for larval behavioural sequences, providing a benchmark for identifying complex behavioural changes. 4) A comprehensive analysis technique using suffix trees to categorize genetic lines into clusters based on common action sequences. We showcase these methodologies through a behavioural screen focused on responses to an air puff, analyzing data from 280,716 larvae across 568 genetic lines. Author SummaryThere is a significant gap in understanding between the architecture of neural circuits and the mechanisms of action selection and behaviour generation.Drosophila larvae have emerged as an ideal platform for simultaneously probing behaviour and the underlying neuronal computation [1]. Modern genetic tools allow efficient activation or silencing of individual and small groups of neurons. Combining these techniques with standardized stimuli over thousands of individuals makes it possible to relate neurons to behaviour causally. However, extracting these relationships from massive and noisy recordings requires the development of new statistically robust approaches. We introduce a suite of statistical methods that utilize individual behavioural data and the overarching structure of the behavioural screen to deduce subtle behavioural changes from raw data. Given our studys extensive number of larvae, addressing and preempting potential challenges in body shape recognition is critical for enhancing behaviour detection. To this end, we have adopted a physics-informed inference model. Our first group of techniques enables robust statistical analysis within a learned continuous behaviour latent space, facilitating the detection of subtle behavioural shifts relative to reference genetic lines. A second array of methods probes for subtle variations in action sequences by comparing them to a bespoke generative model. Together, these strategies have enabled us to construct representations of behavioural patterns specific to a lineage and identify a roster of "hit" neurons with the potential to influence behaviour subtly.

neuroscience↗

bioRxiv2024

LarvaTagger: Manual and automatic tagging of Drosophila larval behaviour

MotivationAs more behavioural assays are carried out in large-scale experiments on Drosophila larvae, the definitions of the archetypal actions of a larva are regularly refined. In addition, video recording and tracking technologies constantly evolve. Consequently, automatic tagging tools for Drosophila larval behaviour must be retrained to learn new representations from new data. However, existing tools cannot transfer knowledge from large amounts of previously accumulated data. We introduce LarvaTagger, a piece of software that combines a pre-trained deep neural network, providing a continuous latent representation of larva actions for stereotypical behaviour identification, with a graphical user interface to manually tag the behaviour and train new automatic taggers with the updated ground truth. ResultsWe reproduced results from an automatic tagger with high accuracy, and we demonstrated that pre-training on large databases accelerates the training of a new tagger, achieving similar prediction accuracy using less data. AvailabilityAll the code is free and open source. Docker images are also available. See git-lab.pasteur.fr/nyx/LarvaTagger.jl.

bioinformatics↗

bioRxiv2022

The connectome of an insect brain

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 brains most recurrent circuits comprised the input and output neurons of the learning center. Some structural features, including multilayer shortcuts and nested recurrent loops, resembled powerful machine learning architectures. The identified brain architecture provides a basis for future experimental and theoretical studies of neural circuits. One-Sentence SummaryWe generated a synaptic-resolution brain connectome and characterized its connection types, neuron types, and circuit motifs.

neuroscience↗

bioRxiv2021

Serotonergic neurons mediate operant conditioning in Drosophila larvae

Observed across species, operant conditioning facilitates learned associations between behaviours and outcomes, biasing future action selection to maximise reward and avoid punishment. To elucidate the underlying neural mechanisms, we built a high-throughput tracker for Drosophila melanogaster larvae, combining real-time behaviour detection with closed-loop optogenetic and thermogenetic stimulation capabilities. We demonstrate operant conditioning in Drosophila larvae by inducing a bend direction preference through optogenetic activation of reward-encoding serotonergic neurons. Specifically, we establish that the ventral nerve cord is necessary for this memory formation. Our results extend the role of serotonergic neurons for learning in insects as well as the existence of learning circuits outside the mushroom body. This work supports future studies on the function of serotonin and the mechanisms underlying operant conditioning at both circuit and cellular levels.

neuroscience↗

bioRxiv2020

Circuits for integrating learnt and innate valences in the fly brain

Animal behavior is shaped both by evolution and by individual experience. In many species parallel brain pathways are thought to encode innate and learnt behavior drives and as a result may link the same sensory cue to different actions if innate and learnt drives are in opposition. How these opposing drives are integrated into a single coherent action is not well understood. In insects, the Mushroom Body Output Neurons (MBONs) and the Lateral Horn Neurons (LHNs) are thought to provide the learnt and innate drives, respectively. However their patterns of convergence and the mechanisms by which their outputs are used to select actions are not well understood. We used electron microscopy reconstruction to comprehensively map the downstream targets of all MBONs in Drosophila larva and characterise their patterns of convergence with LHNs. We discovered convergence neurons that receive direct input from MBONs and LHNs and compare opposite behaviour drives. Functional imaging and optogenetic manipulation suggest these convergence neurons compute the overall predicted value of approaching or avoiding an odor and mediate action selection. Our study describes the circuit mechanisms allowing integration of opposing drives from parallel olfactory pathways.

neuroscience↗

bioRxiv2019

Comparative single-cell transcriptomics of complete insect nervous systems

Molecular profiles of neurons influence information processing, but bridging the gap between genes, circuits, and behavior has been very difficult. Furthermore, the behavioral state of an animal continuously changes across development and as a result of sensory experience. How behavioral state influences molecular cell state is poorly understood. Here we present a complete atlas of the Drosophila larval central nervous system composed of over 200,000 single cells across four developmental stages. We develop polyseq, a python package, to perform cell-type analyses. We use single-molecule RNA-FISH to validate our scRNAseq findings. To investigate how internal state affects cell state, we optogentically altered internal state with high-throughput behavior protocols designed to mimic wasp sting and over activation of the memory system. We found nervous system-wide and neuron-specific gene expression changes. This resource is valuable for developmental biology and neuroscience, and it advances our understanding of how genes, neurons, and circuits generate behavior.

neuroscience↗

bioRxiv2019

Synaptic specificity is collectively determined by partner identity, location and activity

Our nervous system is organized into circuits with specifically matched and tuned cell-to-cell connections that are essential for proper function. The mechanisms by which presynaptic axon terminals and postsynaptic dendrites recognize each other and establish the correct number of connections are still incompletely understood. Sperrys chemoaffinity hypothesis proposes that pre- and postsynaptic partners express specific combinations of molecules that enable them to recognize each other. Alternatively, Peters rule proposes that presynaptic axons and postsynaptic dendrites use non-partner-derived global positional cues to independently reach their target area, and once there they randomly connect with any available neuron. These connections can then be further refined by additional mechanisms based on synaptic activity. We used the tractable genetic model system, the Drosophila embryo and larva, to test these hypotheses and elucidate the roles of 1) global positional cues, 2) partner-derived cues and 3) synaptic activity in the establishment of selective connections in the developing nerve cord. We altered the position or activity of presynaptic partners and analyzed the effect of these manipulations on the number of synapses with specific postsynaptic partners, strength of functional connections, and behavior controlled by these neurons. For this purpose, we combined developmental live imaging, electron microscopy reconstruction of circuits, functional imaging of neuronal activity, and behavioral experiments in wildtype and experimental animals. We found that postsynaptic dendrites are able to find, recognize, and connect to their presynaptic partners even when these have been shifted to ectopic locations through the overexpression of receptors for midline guidance cues. This suggests that neurons use partner-derived cues that allow them to identify and connect to each other. However, while partner-derived cues are sufficient for recognition between specific partners and establishment of connections;; without orderly positioning of axon terminals by positional cues and without synaptic activity during embryonic development, the numbers of functional connections are altered with significant consequences for behavior. Thus, multiple mechanisms including global positional cues, partner-derived cues, and synaptic activity contribute to proper circuit assembly in the developing Drosophila nerve cord.

neuroscience↗

bioRxiv2019

Multilevel feedback architecture for adaptive regulation of learning in the insect brain

Modulatory (e.g. dopaminergic) neurons provide \"teaching signals\" that drive associative learning across the animal kingdom, but the circuits that regulate their activity and compute teaching signals are still poorly understood. We provide the first synaptic-resolution connectome of the circuitry upstream of all modulatory neurons in a brain center for associative learning, the mushroom body (MB) of the Drosophila larva. We discovered afferent pathways from sensory neurons, as well as an unexpected large population of 61 feedback neuron pairs that provide one- and two-step feedback from MB output neurons. The majority of these feedback pathways link distinct memory systems (e.g. aversive and appetitive). We functionally confirmed some of the structural pathways and found that some modulatory neurons compare inhibitory input from their own compartment and excitatory input from compartments of opposite valence, enabling them to compute integrated common-currency predicted values across aversive and appetitive memory systems. This architecture suggests that the MB functions as an interconnected ensemble during learning and that distinct types of previously formed memories can regulate future learning about a stimulus. We developed a model of the circuit constrained by the connectome and by the functional data which revealed that the newly discovered architectural motifs, namely the multilevel feedback architecture and the extensive cross-compartment connections, increase the computational performance and flexibility on learning tasks. Together our study provides the most detailed view to date of a recurrent brain circuit that computes teaching signals and provides insights into the architectural motifs that support reinforcement learning in a biological system.

neuroscience↗

bioRxiv2018

Neural substrates of navigational decision-making in Drosophila larva anemotaxis

Small animals use sensory information to navigate their environment in order to reach more favorable conditions. in gradients of light, temperature, odors and CO2, Drosophila larvae alternate periods of runs and turns, regulating the frequency size and direction of turns, to move in a favorable direction. Whether larvae use the same strategies when navigating in response to somatosensory input is unknown. Further, while many of the sensory neurons that mediate navigation behaviors have been described, where and how these navigational strategies are implemented in the central nervous system and controlled by neuronal circuit elements is not well known. Here we characterize for the first time the navigational strategies of Drosophila larvae in gradients of air-current speeds using high-throughput behavioral assays and quantitative behavioral analysis. We find that larvae extend runs towards favorable directions and shorten runs in unfavorable directions, and that larvae regulate both the direction and amplitudes of turns. These results suggest similar central decision-making mechanisms underlie navigation behaviors in somatosensory and other sensory modalities. By silencing the activity of individual neurons and very sparse expression patterns (2 or 3 neuron types), we further identify the sensory neurons and circuit elements in the ventral nerve cord and brain of the larva required for navigational decisions during anemotaxis. The phenotypes of these central neurons are consistent with a mechanism where the increase of the turning rate in unfavorable conditions and decrease in turning rate in favorable conditions are independently controlled. In addition, we find phenotypes that suggest that the decisions of whether and which way to turn are controlled independently. Our study reveals that different neuronal modules in the nerve cord and brain mediate different aspects of navigational decision making. The neurons identified in our screen provide a basis for future detailed mechanistic understanding of the circuit principles of navigational decisionmaking.

neuroscience↗

bioRxiv2017

Mapping neurons and brain regions underlying sensorimotor decisions and sequences

A fundamental property of the nervous systems across the animal kingdom is the ability to select appropriate actions and sequences of actions in response to sensory cues. Many behaviors are physically mutually exclusive so commitment to one requires complete suppression of others. Further, because many behaviors are serially organized into sequences of actions, mechanisms must also exist that control transitions from one action to the next. The circuit mechanisms by which nervous systems achieve behavioral choice, stability and transitions are still incompletely understood. A key step in understanding these functions is to identify neurons and brain areas involved in controlling behavioral choice, stability and transitions. To do this, we developed an approach where we combined a large-scale neuronal inactivation screen with an automated action detection of sensorimotor decisions and sequences in response to a sensory cue. We characterized the response of wild-type larvae to a mechanosensory stimulus and found they can respond to the stimulus with a probabilistic sequence of 4 possible actions (head-cast, head-retraction, back-up and stop). We analyzed behaviors from around three hundred thousand larvae (N=2.9x105) where we selectively silenced small numbers of neuron types and individual neuron types systematically across the nervous system using a library of Drosophila GAL4 lines and determined the effect of these manipulations on larval mechanosensory responses. We identified neurons and brain-regions that when inactivated affected Drosophila larval sensorimotor decisions and sequence transitions between the different actions. Specifically, we identified 51 candidate lines for sensory processing and 24 candidate lines for competitive interactions. We also detected phenotype categories for sequence transitions consistent with a model of sequence generation where transitions and reversals are independently controlled. These findings provide the basis for understanding how sensorimotor decisions and sequence transition are controlled by the nervous system.

neuroscience↗