Search

SummaryMany animals rely on vision to navigate through their environment. The pattern of changes in the visual scene induced by self-motion is theoptic flow1, which is first estimated in local patches by directionally selective (DS) neurons2–4. But how should the arrays of DS neurons, each responsive to motion in a preferred direction at a specific retinal position, be organized to support robust decoding of optic flow by downstream circuits? Understanding this global organization is challenging because it requires mapping fine, local features of neurons across the animal’s field of view3. InDrosophila, the asymmetric dendrites of the T4 and T5 DS neurons establish their preferred direction, making it possible to predict DS responses from anatomy4,5. Here we report that the preferred directions of fly DS neurons vary at different retinal positions and show that this spatial variation is established by the anatomy of the compound eye. To estimate the preferred directions across the visual field, we reconstructed hundreds of T4 neurons in a full brain EM volume6and discovered unexpectedly stereotypical dendritic arborizations that are independent of location. We then used whole-head μCT scans to map the viewing directions of all compound eye facets and found a non-uniform sampling of visual space that explains the spatial variation in preferred directions. Our findings show that the organization of preferred directions in the fly is largely determined by the compound eye, exposing an intimate and unexpected connection between the peripheral structure of the eye, functional properties of neurons deep in the brain, and the control of body movements.

We develop an automatic method for synaptic partner identification in insect brains and use it to predict synaptic partners in a whole-brain electron microscopy dataset of the fruit fly. The predictions can be used to infer a connectivity graph with high accuracy, thus allowing fast identification of neural pathways. To facilitate circuit reconstruction using our results, we develop CIRCUITMAP, a user interface add-on for the circuit annotation tool CATMAID. A deep-learning-based approach enables automatic identification of synaptically connected neurons in electron microscopy datasets of the fly brain.

Gustatory sensory neurons detect caloric and harmful compounds in potential food and convey this information to the brain to inform feeding decisions. To examine the signals that gustatory neurons transmit and receive, we reconstructed gustatory axons and their synaptic sites in the adult Drosophila melanogaster brain, utilizing a whole-brain electron microscopy volume. We reconstructed 87 gustatory projections from the proboscis labellum in the right hemisphere and 57 from the left, representing the majority of labellar gustatory axons. Gustatory neurons contain a nearly equal number of interspersed pre- and postsynaptic sites, with extensive synaptic connectivity among gustatory axons. Morphology- and connectivity-based clustering revealed six distinct groups, likely representing neurons recognizing different taste modalities. The vast majority of synaptic connections are between neurons of the same group. This study resolves the anatomy of labellar gustatory projections, reveals that gustatory projections are segregated based on taste modality, and uncovers synaptic connections that may alter the transmission of gustatory signals.

Gustatory sensory neurons detect caloric and harmful compounds in potential food and convey this information to the brain to inform feeding decisions. To examine the signals that gustatory neurons transmit and receive, we reconstructed gustatory axons and their synaptic sites in the adult Drosophila melanogaster brain, utilizing a whole-brain electron microscopy volume. We reconstructed 87 gustatory projections from the proboscis labellum in the right hemisphere and 57 from the left, representing the majority of labellar gustatory axons. Gustatory neurons contain a nearly equal number of interspersed pre-and post-synaptic sites, with extensive synaptic connectivity among gustatory axons. Morphology- and connectivity-based clustering revealed six distinct groups, likely representing neurons recognizing different taste modalities. The vast majority of synaptic connections are between neurons of the same group. This study resolves the anatomy of labellar gustatory projections, reveals that gustatory projections are segregated based on taste modality, and uncovers synaptic connections that may alter the transmission of gustatory signals.

Associative memory formation and recall in the adult fruit flyDrosophila melanogasteris subserved by the mushroom body (MB). Upon arrival in the MB, sensory information undergoes a profound transformation. Olfactory projection neurons (PNs), the main MB input, exhibit broadly tuned, sustained, and stereotyped responses to odorants; in contrast, their postsynaptic targets in the MB, the Kenyon cells (KCs), are nonstereotyped, narrowly tuned, and only briefly responsive to odorants. Theory and experiment have suggested that this transformation is implemented by random connectivity between KCs and PNs. However, this hypothesis has been challenging to test, given the difficulty of mapping synaptic connections between large numbers of neurons to achieve a unified view of neuronal network structure. Here we used a recent whole-brain electron microscopy (EM) volume of the adult fruit fly to map large numbers of PN- to-KC connections at synaptic resolution. Comparison of the observed connectome to precisely defined null models revealed unexpected network structure, in which a subset of food-responsive PN types converge on individual downstream KCs more frequently than expected. The connectivity bias is consistent with the neurogeometry: axons of the overconvergent PNs tend to arborize near one another in the MB main calyx, making local KC dendrites more likely to receive input from those types. Computational modeling of the observed PN-to-KC network showed that input from the overconvergent PN types is better discriminated than input from other types. These results suggest an ‘associative fovea’ for olfaction, in that the MB is wired to better discriminate more frequently occurring and ethologically relevant combinations of food-related odors.

Color and polarization provide complementary information about the world and are detected by specialized photoreceptors. However, the downstream neural circuits that process these distinct modalities are incompletely understood in any animal. Using electron microscopy, we have systematically reconstructed the synaptic targets of the photoreceptors specialized to detect color and skylight polarization in Drosophila, and we have used light microscopy to confirm many of our findings. We identified known and novel downstream targets that are selective for different wavelengths or polarized light, and followed their projections to other areas in the optic lobes and the central brain. Our results revealed many synapses along the photoreceptor axons between brain regions, new pathways in the optic lobes, and spatially segregated projections to central brain regions. Strikingly, photoreceptors in the polarization-sensitive dorsal rim area target fewer cell types, and lack strong connections to the lobula, a neuropil involved in color processing. Our reconstruction identifies shared wiring and modality-specific specializations for color and polarization vision, and provides a comprehensive view of the first steps of the pathways processing color and polarized light inputs.

Diverse mechanosensory neurons detect different mechanical forces that can impact animal behavior. Yet our understanding of the anatomical and physiological diversity of these neurons and the behaviors that they influence is limited. We previously discovered that grooming of theDrosophila melanogasterantennae is elicited by an antennal mechanosensory chordotonal organ, the Johnston’s organ (JO) (Hampel et al., 2015). Here, we describe anatomically and physiologically distinct JO mechanosensory neuron subpopulations that each elicit antennal grooming. We show that the subpopulations project to different, discrete zones in the brain and differ in their responses to mechanical stimulation of the antennae. Although activation of each subpopulation elicits antennal grooming, distinct subpopulations also elicit the additional behaviors of wing flapping or backward locomotion. Our results provide a comprehensive description of the diversity of mechanosensory neurons in the JO, and reveal that distinct JO subpopulations can elicit both common and distinct behavioral responses.

Reconstruction of neural circuitry at single-synapse resolution is a key target for improving understanding of the nervous system in health and disease. Serial section transmission electron microscopy (ssTEM) is among the most prolific imaging methods employed in pursuit of such reconstructions. We demonstrate how Flood-Filling Networks (FFNs) can be used to computationally segment a forty-teravoxel whole-brain Drosophila ssTEM volume. To compensate for data irregularities and imperfect global alignment, FFNs were combined with procedures that locally re-align serial sections as well as dynamically adjust and synthesize image content. The proposed approach produced a largely merger-free segmentation of the entire ssTEM Drosophila brain, which we make freely available. As compared to manual tracing using an efficient skeletonization strategy, the segmentation enabled circuit reconstruction and analysis workflows that were an order of magnitude faster.

Large scientific projects in genomics and astronomy are influential not because they answer any single question but because they enable investigation of continuously arising new questions from the same data-rich sources. Advances in automated mapping of the brain's synaptic connections (connectomics) suggest that the complicated circuits underlying brain function are ripe for analysis. We discuss benefits of mapping a mouse brain at the level of synapses.

SummaryThe capacity of utilizing past experience to guide future action is a fundamental and conserved function of the nervous system. Associative memory formation initiated by the coincident detection of a conditioned stimulus (CS, e.g. odour) and an unconditioned stimulus (US, e.g. sugar reward) can lead to a short-lived memory trace (STM) within distinct circuits [1-5]. Memories can be consolidated into long-term memories (LTM) through processes that are not fully understood, but depend on de-novo protein synthesis [6, 7], require structural modifications within the involved neuronal circuits and might lead to the recruitment of additional ones [8-17]. Compared to modulation of existing connections, the reorganization of circuits affords the unique possibility of sampling for potential new partners [18-20]. Nonetheless, only few examples of rewiring associated with learning have been established thus far [14, 21-24]. Here, we report that memory consolidation is associated with the structural and functional reorganization of an identified circuit in the adult fly brain. The formation and retrieval of olfactory associative memories in Drosophila requires the mushroom body (MB) [25]. We identified the individual synapses of olfactory projection neurons (PNs) that deliver a conditioned odour to the MB and reconstructed the complexity of the microcircuit they form. Combining behavioural experiments with high-resolution microscopy and functional imaging, we demonstrated that the consolidation of appetitive olfactory memories closely correlates with an increase in the number of synaptic complexes formed by the PNs that deliver the conditioned stimulus and their postsynaptic partners. These structural changes result in additional functional synaptic connections.

Summary Animals exhibit innate and learned preferences for temperature and humidity—conditions critical for their survival and reproduction. Leveraging a whole-brain electron microscopy volume, we studied the adult Drosophila melanogaster circuitry associated with antennal thermo- and hygrosensory neurons. We have identified two new target glomeruli in the antennal lobe, in addition to the five known ones, and the ventroposterior projection neurons (VP PNs) that relay thermo- and hygrosensory information to higher brain centers, including the mushroom body and lateral horn, seats of learned and innate behavior. We present the first connectome of a thermo- and hygrosensory neuropil, the lateral accessory calyx (lACA), by reconstructing neurons downstream of heating- and cooling-responsive VP PNs. A few mushroom body-intrinsic neurons solely receive thermosensory input from the lACA, while most receive additional olfactory and thermo- and/or hygrosensory PN inputs. Furthermore, several classes of lACA-associated neurons form a local network with outputs to other brain neuropils, suggesting that the lACA serves as a hub for thermo- and hygrosensory circuitry. For example, DN1a neurons link thermosensory PNs in the lACA to the circadian clock via the accessory medulla. Finally, we survey strongly connected downstream partners of VP PNs across the protocerebrum; these include a descending neuron targeted by dry-responsive VP PNs, meaning that just two synapses might separate hygrosensory inputs from motor circuits. These data provide a comprehensive first- and second-order layer analysis of Drosophila thermo- and hygrosensory systems and an initial survey of third-order neurons that could directly modulate behavior., Graphical Abstract, Highlights • Two novel thermo- and/or hygrosensory glomeruli in the fly antennal lobe • First complete set of thermosensory and hygrosensory projection neurons • First connectome for a thermo- and hygrosensory neuropil • Third-order thermo- and hygrosensory neurons, including link to circadian clock, Marin et al. use connectomics and genetics for comprehensive identification of temperature and humidity sensory neurons in the Drosophila brain. They reconstruct all projections to higher brain areas and select higher-order targets, including the mushroom body lateral accessory calyx, linking thermosensation to memory and the circadian clock.

Diverse subpopulations of mechanosensory neurons detect different mechanical forces and influence behavior. How these subpopulations connect with central circuits to influence behavior remains an important area of study. We previously discovered a neural circuit that elicits grooming of theDrosophila melanogasterantennae that is activated by an antennal mechanosensory chordotonal organ, the Johnston’s organ (JO) (Hampel et al., 2015). Here, we describe anatomically and physiologically distinct JO mechanosensory neuron subpopulations and define how they interface with the circuit that elicits antennal grooming. We show that the subpopulations project to distinct zones in the brain and differ in their responses to mechanical stimulation of the antennae. Each subpopulation elicits grooming through direct synaptic connections with a single interneuron in the circuit, the dendrites of which span the different mechanosensory afferent projection zones. Thus, distinct JO subpopulations converge onto the same neural circuit to elicit a common behavioral response.

Visual systems can exploit spatial correlations in the visual scene by using retinotopy, the organizing principle by which neighboring cells encode neighboring spatial locations. However, retinotopy is often lost, such as when visual pathways are integrated with other sensory modalities. How is spatial information processed in the absence of retinotopy? Here, we focused on visual looming responsive LC6 cells in Drosophila, a population whose dendrites collectively tile the visual field, but whose axons form a single glomerulus—a structure lacking retinotopic organization—in the central brain. We identified multiple glomerulus neurons and found that they respond to looming in different portions of the visual field, unexpectedly preserving spatial information. Through EM reconstruction of all LC6 synaptic inputs to the glomerulus, we found that LC6 and downstream cell types form circuits within the glomerulus that establish spatial readout of visual features and contralateral suppression—mechanisms that transform visual information for behavioral control.

Visual systems can exploit spatial correlations in the visual scene by using retinotopy, the organizing principle by which neighboring cells encode neighboring spatial locations. However, retinotopy is often lost, such as when visual pathways are integrated with other sensory modalities. How is spatial information processed outside of strictly visual brain areas? Here, we focused on visual looming responsive LC6 cells in Drosophila, a population whose dendrites collectively cover the visual field, but whose axons form a single glomerulus—a structure without obvious retinotopic organization—in the central brain. We identified multiple cell types downstream of LC6 in the glomerulus and found that they more strongly respond to looming in different portions of the visual field, unexpectedly preserving spatial information. Through EM reconstruction of all LC6 synaptic inputs to the glomerulus, we found that LC6 and downstream cell types form circuits within the glomerulus that enable spatial readout of visual features and contralateral suppression—mechanisms that transform visual information for behavioral control.

Serotonergic neurons project widely throughout the brain to modulate diverse physiological and behavioral processes. However, a single-cell resolution understanding of the connectivity of serotonergic neurons is currently lacking. Using a whole-brain EM dataset of a female Drosophila, we comprehensively determine the wiring logic of a broadly projecting serotonergic neuron (the CSDn) that spans several olfactory regions. Within the antennal lobe, the CSDn differentially innervates each glomerulus, yet surprisingly, this variability reflects a diverse set of presynaptic partners, rather than glomerulus-specific differences in synaptic output, which is predominately to local interneurons. Moreover, the CSDn has distinct connectivity relationships with specific local interneuron subtypes, suggesting that the CSDn influences distinct aspects of local network processing. Across olfactory regions, the CSDn has different patterns of connectivity, even having different connectivity with individual projection neurons that also span these regions. Whereas the CSDn targets inhibitory local neurons in the antennal lobe, the CSDn has more distributed connectivity in the LH, preferentially synapsing with principal neuron types based on transmitter content. Last, we identify individual novel synaptic partners associated with other sensory domains that provide strong, top-down input to the CSDn. Together, our study reveals the complex connectivity of serotonergic neurons, which combine the integration of local and extrinsic synaptic input in a nuanced, region-specific manner. SIGNIFICANCE STATEMENT All sensory systems receive serotonergic modulatory input. However, a comprehensive understanding of the synaptic connectivity of individual serotonergic neurons is lacking. In this study, we use a whole-brain EM microscopy dataset to comprehensively determine the wiring logic of a broadly projecting serotonergic neuron in the olfactory system of Drosophila. Collectively, our study demonstrates, at a single-cell level, the complex connectivity of serotonergic neurons within their target networks, identifies specific cell classes heavily targeted for serotonergic modulation in the olfactory system, and reveals novel extrinsic neurons that provide strong input to this serotonergic system outside of the context of olfaction. Elucidating the connectivity logic of individual modulatory neurons provides a ground plan for the seemingly heterogeneous effects of modulatory systems.

Animals exhibit innate behaviours to a variety of sensory stimuli including olfactory cues. In Drosophila, one higher olfactory centre, the lateral horn (LH), is implicated in innate behaviour. However, our structural and functional understanding of the LH is scant, in large part due to a lack of sparse neurogenetic tools for this region. We generate a collection of split-GAL4 driver lines providing genetic access to 82 LH cell types. We use these to create an anatomical and neurotransmitter map of the LH and link this to EM connectomics data. We find ~30% of LH projections converge with outputs from the mushroom body, site of olfactory learning and memory. Using optogenetic activation, we identify LH cell types that drive changes in valence behavior or specific locomotor programs. In summary, we have generated a resource for manipulating and mapping LH neurons, providing new insights into the circuit basis of innate and learned olfactory behavior.

Mating and egg laying are tightly cooordinated events in the reproductive life of all oviparous females. Oviposition is typically rare in virgin females but is initiated after copulation. Here we identify the neural circuitry that links egg laying to mating status in Drosophila melanogaster. Activation of female-specific oviposition descending neurons (oviDNs) is necessary and sufficient for egg laying, and is equally potent in virgin and mated females. After mating, sex peptide—a protein from the male seminal fluid—triggers many behavioural and physiological changes in the female, including the onset of egg laying1. Sex peptide is detected by sensory neurons in the uterus2–4, and silences these neurons and their postsynaptic ascending neurons in the abdominal ganglion5. We show that these abdominal ganglion neurons directly activate the female-specific pC1 neurons. GABAergic (γ-aminobutyric-acid-releasing) oviposition inhibitory neurons (oviINs) mediate feed-forward inhibition from pC1 neurons to both oviDNs and their major excitatory input, the oviposition excitatory neurons (oviENs). By attenuating the abdominal ganglion inputs to pC1 neurons and oviINs, sex peptide disinhibits oviDNs to enable egg laying after mating. This circuitry thus coordinates the two key events in female reproduction: mating and egg laying. Neuron-tracing and labelling experiments in Drosophila females reveal the neural circuitry that coordinates mating and egg laying, and the role of sex peptide from male seminal fluid in triggering these neurons.

Reconstruction of neural circuitry at single-synapse resolution is a key target for improving understanding of the nervous system in health and disease. Serial section transmission electron microscopy (ssTEM) is among the most prolific imaging methods employed in pursuit of such reconstructions. We demonstrate how Flood-Filling Networks (FFNs) can be used to computationally segment a forty-teravoxel whole-brain Drosophila ssTEM volume. To compensate for data irregularities and imperfect global alignment, FFNs were combined with procedures that locally re-align serial sections as well as dynamically adjust and synthesize image content. The proposed approach produced a largely merger-free segmentation of the entire ssTEM Drosophila brain, which we make freely available. As compared to manual tracing using an efficient skeletonization strategy, the segmentation enabled circuit reconstruction and analysis workflows that were an order of magnitude faster.

Summary Nervous systems contain sensory neurons, local neurons, projection neurons, and motor neurons. To understand how these building blocks form whole circuits, we must distil these broad classes into neuronal cell types and describe their network connectivity. Using an electron micrograph dataset for an entire Drosophila melanogaster brain, we reconstruct the first complete inventory of olfactory projections connecting the antennal lobe, the insect analog of the mammalian olfactory bulb, to higher-order brain regions in an adult animal brain. We then connect this inventory to extant data in the literature, providing synaptic-resolution “holotypes” both for heavily investigated and previously unknown cell types. Projection neurons are approximately twice as numerous as reported by light level studies; cell types are stereotyped, but not identical, in cell and synapse numbers between brain hemispheres. The lateral horn, the insect analog of the mammalian cortical amygdala, is the main target for this olfactory information and has been shown to guide innate behavior. Here, we find new connectivity motifs, including axo-axonic connectivity between projection neurons, feedback, and lateral inhibition of these axons by a large population of neurons, and the convergence of different inputs, including non-olfactory inputs and memory-related feedback onto third-order olfactory neurons. These features are less prominent in the mushroom body calyx, the insect analog of the mammalian piriform cortex and a center for associative memory. Our work provides a complete neuroanatomical platform for future studies of the adult Drosophila olfactory system., Graphical Abstract, Highlights • First complete parts list for second-order neurons of an adult olfactory system • Quantification of left-right stereotypy in cell and synapse number • Axo-axonic connections form hierarchical communities in the lateral horn • Local and memory-related feedback target projection neuron axons, Bates, Schlegel et al. report the first complete part list of second-order neurons in an adult arthropod olfactory system. They show that these neurons are stereotyped across hemispheres and reveal their hierarchical interconnectivity and how they connect to third-order targets and memory-related neurons.

Summary Drosophila melanogaster has a rich repertoire of innate and learned behaviors. Its 100,000-neuron brain is a large but tractable target for comprehensive neural circuit mapping. Only electron microscopy (EM) enables complete, unbiased mapping of synaptic connectivity; however, the fly brain is too large for conventional EM. We developed a custom high-throughput EM platform and imaged the entire brain of an adult female fly at synaptic resolution. To validate the dataset, we traced brain-spanning circuitry involving the mushroom body (MB), which has been extensively studied for its role in learning. All inputs to Kenyon cells (KCs), the intrinsic neurons of the MB, were mapped, revealing a previously unknown cell type, postsynaptic partners of KC dendrites, and unexpected clustering of olfactory projection neurons. These reconstructions show that this freely available EM volume supports mapping of brain-spanning circuits, which will significantly accelerate Drosophila neuroscience. Video Abstract, Graphical Abstract, Highlights • A complete adult Drosophila brain was imaged with EM and has been made publicly available • The imaged volume enables brain-spanning mapping of circuits at synaptic resolution • All mushroom body (MB) calyx inputs were mapped, revealing a new cell type, MB-CP2 • Previously unidentified synaptic partners form recurrent microcircuits in MB calyx, Electron microscopy imaging of the entire adult fruit fly brain at synapse resolution reveals circuitry spanning multiple regions and connectivity between known and previously unknown cell types.

Most sensory systems are organized into parallel neuronal pathways that process distinct aspects of incoming stimuli. For example, in insects, second order olfactory projection neurons target both the mushroom body, which is required for learning, and the lateral horn (LH), which has been proposed to mediate innate olfactory behavior. Mushroom body neurons encode odors in a sparse population code, which does not appear stereotyped across animals. In contrast the functional principles of odor coding in the LH remain poorly understood. We have carried out a comprehensive anatomical analysis of the Drosophila LH, counting ~1400 neurons; combining genetic driver lines, anatomical and functional criteria, we identify 165 LHN cell types. We then show that genetically labeled LHNs have stereotyped odor responses across animals for 33 of these cell types. LHN tuning can be ultra-sparse (1/40 odors tested), but on average single LHNs respond to three times more odors than single projection neurons. This difference can be rationalized by our observation that LHNs are better odor categorizers, likely due to pooling of input projection neurons responding to different odors of the same category. Our results reveal some of the principles by which a higher sensory processing area can extract innate behavioral significance from sensory stimuli.

Accurately predicting an outcome requires that animals learn supporting and conflicting evidence from sequential experience. In mammals and invertebrates, learned fear responses can be suppressed by experiencing predictive cues without punishment, a process called memory extinction. Here, we show that extinction of aversive memories in Drosophila requires specific dopaminergic neurons, which indicate that omission of punishment is remembered as a positive experience. Functional imaging revealed co-existence of intracellular calcium traces in different places in the mushroom body output neuron network for both the original aversive memory and a new appetitive extinction memory. Light and ultrastructural anatomy are consistent with parallel competing memories being combined within mushroom body output neurons that direct avoidance. Indeed, extinction-evoked plasticity in a pair of these neurons neutralizes the potentiated odor response imposed in the network by aversive learning. Therefore, flies track the accuracy of learned expectations by accumulating and integrating memories of conflicting events.

In this contribution we demonstrate how a Multicut-based segmentation pipeline can be scaled up to datasets of hundreds of Gigabytes in size. Such datasets are prevalent in connectomics, where neuron segmentation needs to be performed across very large electron microscopy image volumes. We show the advantages of a hierarchical block-wise scheme over local stitching strategies and evaluate the performance of different Multicut solvers for the segmentation of the blocks in the hierarchy. We validate the accuracy of our algorithm on a small fully annotated dataset (5×5×5 μm) and demonstrate no significant loss in segmentation quality compared to solving the Multicut problem globally. We evaluate the scalability of the algorithm on a 95 x 60 x 60 μm image volume and show that solving the Multicut problem is no longer the bottleneck of the segmentation pipeline.

Summary The behavioral response to a sensory stimulus may depend on both learned and innate neuronal representations. How these circuits interact to produce appropriate behavior is unknown. In Drosophila, the lateral horn (LH) and mushroom body (MB) are thought to mediate innate and learned olfactory behavior, respectively, although LH function has not been tested directly. Here we identify two LH cell types (PD2a1 and PD2b1) that receive input from an MB output neuron required for recall of aversive olfactory memories. These neurons are required for aversive memory retrieval and modulated by training. Connectomics data demonstrate that PD2a1 and PD2b1 neurons also receive direct input from food odor-encoding neurons. Consistent with this, PD2a1 and PD2b1 are also necessary for unlearned attraction to some odors, indicating that these neurons have a dual behavioral role. This provides a circuit mechanism by which learned and innate olfactory information can interact in identified neurons to produce appropriate behavior. Video Abstract, Highlights • Specific Drosophila lateral horn neurons mediate innate attraction to food odors • The same neurons receive plastic odor information from the mushroom body • Recall after associative learning depends on reduced drive to lateral horn neurons • Connectomics circuit for integration of learned and innate odor representations, Sensory stimuli can engage both learned and innate behaviors. Dolan et al. identify neurons in Drosophila that directly integrate unlearned and plastic odor representations; they are required for innate approach to food odors but also learned aversive recall.

Reference EPFL-ARTICLE-226946doi:10.1038/nmeth.4151View record in Web of Science Record created on 2017-03-27, modified on 2017-07-13

Summary Identified neuron classes in vertebrate cortical [ 1 , 2 , 3 , 4 ] and subcortical [ 5 , 6 , 7 , 8 ] areas and invertebrate peripheral [ 9 , 10 , 11 ] and central [ 12 , 13 , 14 ] brain neuropils encode specific visual features of a panorama. How downstream neurons integrate these features to control vital behaviors, like escape, is unclear [ 15 ]. In Drosophila, the timing of a single spike in the giant fiber (GF) descending neuron [ 16 , 17 , 18 ] determines whether a fly uses a short or long takeoff when escaping a looming predator [ 13 ]. We previously proposed that GF spike timing results from summation of two visual features whose detection is highly conserved across animals [ 19 ]: an object’s subtended angular size and its angular velocity [ 5 , 6 , 7 , 8 , 11 , 20 , 21 ]. We attributed velocity encoding to input from lobula columnar type 4 (LC4) visual projection neurons, but the size-encoding source remained unknown. Here, we show that lobula plate/lobula columnar, type 2 (LPLC2) visual projection neurons anatomically specialized to detect looming [ 22 ] provide the entire GF size component. We find LPLC2 neurons to be necessary for GF-mediated escape and show that LPLC2 and LC4 synapse directly onto the GF via reconstruction in a fly brain electron microscopy (EM) volume [ 23 ]. LPLC2 silencing eliminates the size component of the GF looming response in patch-clamp recordings, leaving only the velocity component. A model summing a linear function of angular velocity (provided by LC4) and a Gaussian function of angular size (provided by LPLC2) replicates GF looming response dynamics and predicts the peak response time. We thus present an identified circuit in which information from looming feature-detecting neurons is combined by a common post-synaptic target to determine behavioral output.

We describe a scalable database cluster for the spatial analysis and annotation of high-throughput brain imaging data, initially for 3-d electron microscopy image stacks, but for time-series and multi-channel data as well. The system was designed primarily for workloads that build connectomes---neural connectivity maps of the brain---using the parallel execution of computer vision algorithms on high-performance compute clusters. These services and open-science data sets are publicly available at http://openconnecto.me. The system design inherits much from NoSQL scale-out and data-intensive computing architectures. We distribute data to cluster nodes by partitioning a spatial index. We direct I/O to different systems---reads to parallel disk arrays and writes to solid-state storage---to avoid I/O interference and maximize throughput. All programming interfaces are RESTful Web services, which are simple and stateless, improving scalability and usability. We include a performance evaluation of the production system, highlighting the effectiveness of spatial data organization., 11 pages, 13 figures

The last decade has seen a rapid increase in the number of tools to acquire volume electron microscopy (EM) data. Several new scanning EM (SEM) imaging methods have emerged, and classical transmission EM (TEM) methods are being scaled up and automated. Here we summarize the new methods for acquiring large EM volumes, and discuss the tradeoffs in terms of resolution, acquisition speed, and reliability. We then assess each method's applicability to the problem of reconstructing anatomical connectivity between neurons, considering both the current capabilities and future prospects of the method. Finally, we argue that neuronal ‘wiring diagrams’ are likely necessary, but not sufficient, to understand the operation of most neuronal circuits: volume EM imaging will likely find its best application in combination with other methods in neuroscience, such as molecular biology, optogenetics, and physiology.

We present a novel algorithm to accelerate feature based registration, and demonstrate the utility of the algorithm for the alignment of large transmission electron microscopy (TEM) images to create 3-D images of neural ultrastructure. In contrast to the most similar algorithms, which achieve small computation times by truncated search, our algorithm uses a novel randomized projection to accelerate feature comparison and to enable global search. Further, we demonstrate robust estimation of nonrigid transformations with a novel probabilistic correspondence framework, that enables large TEM images to be rapidly brought into alignment, removing characteristic distortions of the tissue fixation and imaging process. We analyze the impact of randomized projections upon correspondence detection, and upon transformation accuracy, and demonstrate that accuracy is maintained. We provide experimental results that demonstrate significant reduction in computation time and successful alignment of TEM images.

In the cerebral cortex, local circuits consist of tens of thousands of neurons, each of which makes thousands of synaptic connections. Perhaps the biggest impediment to understanding these networks is that we have no wiring diagrams of their interconnections. Even if we had a partial or complete wiring diagram, however, understanding the network would also require information about each neuron's function. Here we show that the relationship between structure and function can be studied in the cortex with a combination of in vivo physiology and network anatomy. We used two-photon calcium imaging to characterize a functional property—the preferred stimulus orientation—of a group of neurons in the mouse primary visual cortex. We then used large-scale electron microscopy (EM) of serial thin sections to trace a portion of these neurons’ local network. Consistent with a prediction from recent physiological experiments, inhibitory interneurons received convergent anatomical input from nearby excitatory neurons with a broad range of preferred orientations, although weak biases could not be rejected.

In this paper we introduce a novel algorithm for alignment of Electron Microscopy images for 3D reconstruction. The algorithm extends the Expectation Maximization - Iterative Closest Points (EM-ICP) algorithm to go from point matching to patch matching. We utilize local patch characteristics to achieve improved registration. The method is applied to enable 3D reconstruction of Transmission Electron Microscopy (TEM) images. We demonstrate results on large TEM images and show the increased alignment accuracy of our approach.

3D reconstruction from serial 2D microscopy images depends on non-linear alignment of serial sections. For some structures, such as the neuronal circuitry of the brain, very large images at very high resolution are necessary to permit reconstruction. These very large images prevent the direct use of classical registration methods. We propose in this work a method to deal with the non-linear alignment of arbitrarily large 2D images using the finite support properties of cubic B-splines. After initial affine alignment, each large image is split into a grid of smaller overlapping sub-images, which are individually registered using cubic B-splines transformations. Inside the overlapping regions between neighboring sub-images, the coefficients of the knots controlling the B-splines deformations are blended, to create a virtual large grid of knots for the whole image. The sub-images are resampled individually, using the new coefficients, and assembled together into a final large aligned image. We evaluated the method on a series of large transmission electron microscopy images and our results indicate significant improvements compared to both manual and affine alignment.

We describe a method for fully automated detection of chemical synapses in serial electron microscopy images with highly anisotropic axial and lateral resolution, such as images taken on transmission electron microscopes. Our pipeline starts from classification of the pixels based on 3D pixel features, which is followed by segmentation with an Ising model MRF and another classification step, based on object-level features. Classifiers are learned on sparse user labels; a fully annotated data subvolume is not required for training. The algorithm was validated on a set of 238 synapses in 20 serial 7197×7351 pixel images (4.5×4.5×45 nm resolution) of mouse visual cortex, manually labeled by three independent human annotators and additionally re-verified by an expert neuroscientist. The error rate of the algorithm (12% false negative, 7% false positive detections) is better than state-of-the-art, even though, unlike the state-of-the-art method, our algorithm does not require a prior segmentation of the image volume into cells. The software is based on the ilastik learning and segmentation toolkit and the vigra image processing library and is freely available on our website, along with the test data and gold standard annotations (http://www.ilastik.org/synapse-detection/sstem).