Memory guides behavior across widely varying environments and must therefore be both sufficiently specific and general. A memory too specific will be useless in even a slightly different environment, while an overly general memory may lead to suboptimal choices. Animals successfully learn to both distinguish between very similar stimuli and generalize across cues. Rather than forming memories that strike a balance between specificity and generality, Drosophila can flexibly categorize a given stimulus into different groups depending on the options available. We asked how this flexibility manifests itself in the well-characterized learning and memory pathways of the fruit fly. We show that flexible categorization in neuronal activity as well as behavior depends on the order and identity of the perceived stimuli. Our results identify the neural correlates of flexible stimulus-categorization in the fruit fly.

Multicellular organisms generate tissues of diverse shapes and functions from cells and extracellular matrices. Their adhesion molecules mediate cell-cell and cell-matrix interactions, which not only play crucial roles in maintaining tissue integrity but also serve as key regulators of tissue morphogenesis. Cells constantly probe their environment to make decisions: They integrate chemical and mechanical information from the environment via diffusible ligand- or adhesion-based signaling to decide whether to release specific signaling molecules or enzymes, to divide or differentiate, to move away or stay, or even whether to live or die. These decisions in turn modify their environment, including the chemical nature and mechanical properties of the extracellular matrix. Tissue morphology is the physical manifestation of the remodeling of cells and matrices by their historical biochemical and biophysical landscapes. We review our understanding of matrix and adhesion molecules in tissue morphogenesis, with an emphasis on key physical interactions that drive morphogenesis. Expected final online publication date for the , Volume 39 is October 2023. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Activity related to movement is found throughout sensory and motor regions of the brain. However, it remains unclear how movement-related activity is distributed across the brain and whether systematic differences exist between brain areas. Here, we analyzed movement related activity in brain-wide recordings containing more than 50,000 neurons in mice performing a decision-making task. Using multiple techniques, from markers to deep neural networks, we find that movement-related signals were pervasive across the brain, but systematically differed across areas. Movement-related activity was stronger in areas closer to the motor or sensory periphery. Delineating activity in terms of sensory- and motor-related components revealed finer scale structures of their encodings within brain areas. We further identified activity modulation that correlates with decision-making and uninstructed movement. Our work charts out a largescale map of movement encoding and provides a roadmap for dissecting different forms of movement and decision-making related encoding across multi-regional neural circuits.

Natural behaviors are a coordinated symphony of motor acts which drive self-induced or reafferent sensory activation. Single sensors only signal presence and magnitude of a sensory cue; they cannot disambiguate exafferent (externally-induced) from reafferent sources. Nevertheless, animals readily differentiate between these sources of sensory signals to make appropriate decisions and initiate adaptive behavioral outcomes. This is mediated by predictive motor signaling mechanisms, which emanate from motor control pathways to sensory processing pathways, but how predictive motor signaling circuits function at the cellular and synaptic level is poorly understood. We use a variety of techniques, including connectomics from both male and female electron microscopy volumes, transcriptomics, neuroanatomical, physiological and behavioral approaches to resolve the network architecture of two pairs of ascending histaminergic neurons (AHNs), which putatively provide predictive motor signals to several sensory and motor neuropil. Both AHN pairs receive input primarily from an overlapping population of descending neurons, many of which drive wing motor output. The two AHN pairs target almost exclusively non-overlapping downstream neural networks including those that process visual, auditory and mechanosensory information as well as networks coordinating wing, haltere, and leg motor output. These results support the conclusion that the AHN pairs multi-task, integrating a large amount of common input, then tile their output in the brain, providing predictive motor signals to non-overlapping sensory networks affecting motor control both directly and indirectly.

To interpret the sensory environment, the brain combines ambiguous sensory measurements with knowledge that reflects context-specific prior experience. But environmental contexts can change abruptly and unpredictably, resulting in uncertainty about the current context. Here we address two questions: how should context-specific prior knowledge optimally guide the interpretation of sensory stimuli in changing environments, and do human decision-making strategies resemble this optimum? We probe these questions with a task in which subjects report the orientation of ambiguous visual stimuli that were drawn from three dynamically switching distributions, representing different environmental contexts. We derive predictions for an ideal Bayesian observer that leverages knowledge about the statistical structure of the task to maximize decision accuracy, including knowledge about the dynamics of the environment. We show that its decisions are biased by the dynamically changing task context. The magnitude of this decision bias depends on the observer's continually evolving belief about the current context. The model therefore not only predicts that decision bias will grow as the context is indicated more reliably, but also as the stability of the environment increases, and as the number of trials since the last context switch grows. Analysis of human choice data validates all three predictions, suggesting that the brain leverages knowledge of the statistical structure of environmental change when interpreting ambiguous sensory signals.

Persistent internal states are important for maintaining survival-promoting behaviors, such as aggression. In female Drosophila melanogaster, we have previously shown that individually activating either aIPg or pC1d cell types can induce aggression. Here we investigate further the individual roles of these cholinergic, sexually dimorphic cell types, and the reciprocal connections between them, in generating a persistent aggressive internal state. We find that a brief 30-second optogenetic stimulation of aIPg neurons was sufficient to promote an aggressive internal state lasting at least 10 minutes, whereas similar stimulation of pC1d neurons did not. While we previously showed that stimulation of pC1e alone does not evoke aggression, persistent behavior could be promoted through simultaneous stimulation of pC1d and pC1e, suggesting an unexpected synergy of these cell types in establishing a persistent aggressive state. Neither aIPg nor pC1d show persistent neuronal activity themselves, implying that the persistent internal state is maintained by other mechanisms. Moreover, inactivation of pC1d did not significantly reduce aIPg-evoked persistent aggression arguing that the aggressive state did not depend on pC1d-aIPg recurrent connectivity. Our results suggest the need for alternative models to explain persistent female aggression.

In most animals, a relatively small number of descending neurons (DNs) connect higher brain centers in the animal’s head to motor neurons (MNs) in the nerve cord of the animal’s body that effect movement of the limbs. To understand how brain signals generate behavior, it is critical to understand how these descending pathways are organized onto the body MNs. In the fly, Drosophila melanogaster, MNs controlling muscles in the leg, wing, and other motor systems reside in a ventral nerve cord (VNC), analogous to the mammalian spinal cord. In companion papers, we introduced a densely-reconstructed connectome of the Drosophila Male Adult Nerve Cord (MANC, Takemura et al., 2023), including cell type and developmental lineage annotation (Marin et al., 2023), which provides complete VNC connectivity at synaptic resolution. Here, we present a first look at the organization of the VNC networks connecting DNs to MNs based on this new connectome information. We proofread and curated all DNs and MNs to ensure accuracy and reliability, then systematically matched DN axon terminals and MN dendrites with light microscopy data to link their VNC morphology with their brain inputs or muscle targets. We report both broad organizational patterns of the entire network and fine-scale analysis of selected circuits of interest. We discover that direct DN-MN connections are infrequent and identify communities of intrinsic neurons linked to control of different motor systems, including putative ventral circuits for walking, dorsal circuits for flight steering and power generation, and intermediate circuits in the lower tectulum for coordinated action of wings and legs. Our analysis generates hypotheses for future functional experiments and, together with the MANC connectome, empowers others to investigate these and other circuits of the Drosophila ventral nerve cord in richer mechanistic detail.

Animal behavior is principally expressed through neural control of muscles. Therefore understanding how the brain controls behavior requires mapping neuronal circuits all the way to motor neurons. We have previously established technology to collect large-volume electron microscopy data sets of neural tissue and fully reconstruct the morphology of the neurons and their chemical synaptic connections throughout the volume. Using these tools we generated a dense wiring diagram, or connectome, for a large portion of the Drosophila central brain. However, in most animals, including the fly, the majority of motor neurons are located outside the brain in a neural center closer to the body, i.e. the mammalian spinal cord or insect ventral nerve cord (VNC). In this paper, we extend our effort to map full neural circuits for behavior by generating a connectome of the VNC of a male fly.

Our companion paper (Takemura et al., 2023) introduces the first completely proofread connectome of the nerve cord of an animal that can walk or fly. The base connectome consists of neuronal morphologies and the connections between them. However, in order to efficiently navigate and understand this connectome, it is crucial to have a system of annotations that systematically categorises and names neurons, linking them to the existing literature. In this paper we describe the comprehensive annotation of the VNC connectome, first by a system of hierarchical coarse annotations, then by grouping left-right and serially homologous neurons and eventually by defining systematic cell types for the intrinsic interneurons and sensory neurons of the VNC; descending and motor neurons are typed in (Cheong et al., 2023). We assign a sensory modality to over 5000 sensory neurons, cluster them by connectivity, and identify serially homologous cell types and a layered organisation likely corresponding to peripheral topography. We identify the developmental neuroblast of origin of the large majority of VNC neurons and confirm that (in most cases) all secondary neurons of each hemilineage express a single neurotransmitter. Neuroblast hemilineages are serially repeated along the segments of the nerve cord and generally exhibit consistent hemilineage-to-hemilineage connectivity across neuromeres, supporting the idea that hemilineages are a major organisational feature of the VNC. We also find that more than a third of individual neurons belong to serially homologous cell types, which were crucial for identifying motor neurons and sensory neurons across leg neuropils. Categorising interneurons by their neuropil innervation patterns provides an additional organisation axis. Over half of the intrinsic neurons of the VNC appear dedicated to the legs, with the majority restricted to single leg neuropils; in contrast, inhibitory interneurons connecting different leg neuropils, especially those crossing the midline, appear rarer than anticipated by standard models of locomotor circuitry. Our annotations are being released as part of the neuprint.janelia.org web application and also serve as the basis of programmatic analysis of the connectome through dedicated tools that we describe in this paper.

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.