Social behaviors involving the interaction of multiple individuals are complex and frequently crucial for an animal's survival. These interactions, ranging across sensory modalities, length scales, and time scales, are often subtle and difficult to quantify. Contextual effects on the frequency of behaviors become even more difficult to quantify when physical interaction between animals interferes with conventional data analysis, e.g. due to visual occlusion. We introduce a method for quantifying behavior in courting fruit flies that combines high-throughput video acquisition and tracking of individuals with recent unsupervised methods for capturing an animal's entire behavioral repertoire. We find behavioral differences in paired and solitary flies of both sexes, identifying specific behaviors that are affected by social and spatial context. Our pipeline allows for a comprehensive description of the interaction between multiple individuals using unsupervised machine learning methods, and will be used to answer questions about the depth of complexity and variance in fruit fly courtship.

DaCapo is a specialized deep learning library tailored to expedite the training and application of existing machine learning approaches on large, near-isotropic image data. In this correspondence, we introduce DaCapo's unique features optimized for this specific domain, highlighting its modular structure, efficient experiment management tools, and scalable deployment capabilities. We discuss its potential to improve access to large-scale, isotropic image segmentation and invite the community to explore and contribute to this open-source initiative.

There is renewed interest in modeling and understanding the nervous system of the nematode $\textit{Caenorhabditis elegans}$ ($\textit{C. elegans}$), as this small model system provides a path to bridge the gap between nervous system structure (connectivity) and function (physiology). However, existing physiology datasets, whether involving passive recording or stimulation, are in distinct formats, and connectome datasets require preprocessing before analysis can commence. Here we compile and homogenize datasets of neural activity and connectivity. Our neural activity dataset is derived from 12 $\textit{C. elegans}$ neuroimaging experiments, while our connectivity dataset is compiled from 9 connectome annotations based on 3 primary electron microscopy studies and 1 signal propagation study. Physiology datasets, collected under varying protocols, measure calcium fluorescence in labeled subsets of the worm's 300 neurons. Our preprocessing pipeline standardizes these datasets by consistently ordering labeled neurons and resampling traces to a common sampling rate, yielding recordings from approximately 900 worms and 250 uniquely labeled neurons. The connectome datasets, collected from electron microscopy reconstructions, represent the entire nervous system as a graph of connections. Our collection is accessible on HuggingFace, facilitating analysis of the structure-function relationship in biology using modern neural network architectures and enabling cross-lab and cross-animal comparisons.

The Importance Weighted Auto Encoder (IWAE) objective has been shown to improve the training of generative models over the standard Variational Auto Encoder (VAE) objective. Here, we derive importance weighted extensions to AVB and AAE. These latent variable models use implicitly defined inference networks whose approximate posterior density q_\phi(z|x) cannot be directly evaluated, an essential ingredient for importance weighting. We show improved training and inference in latent variable models with our adversarially trained importance weighting method, and derive new theoretical connections between adversarial generative model training criteria and marginal likelihood based methods. We apply these methods to the important problem of inferring spiking neural activity from calcium imaging data, a challenging posterior inference problem in neuroscience, and show that posterior samples from the adversarial methods outperform factorized posteriors used in VAEs.

Instance segmentation for unlabeled imaging modalities is a challenging but essential task as collecting expert annotation can be expensive and time-consuming. Existing works segment a new modality by either deploying a pre-trained model optimized on diverse training data or conducting domain translation and image segmentation as two independent steps. In this work, we propose a novel Cyclic Segmentation Generative Adversarial Network (CySGAN) that conducts image translation and instance segmentation jointly using a unified framework. Besides the CycleGAN losses for image translation and supervised losses for the annotated source domain, we introduce additional self-supervised and segmentation-based adversarial objectives to improve the model performance by leveraging unlabeled target domain images. We benchmark our approach on the task of 3D neuronal nuclei segmentation with annotated electron microscopy (EM) images and unlabeled expansion microscopy (ExM) data. Our CySGAN outperforms both pretrained generalist models and the baselines that sequentially conduct image translation and segmentation. Our implementation and the newly collected, densely annotated ExM nuclei dataset, named NucExM, are available at this https URL.

Geometrical distance is an important constraining factor underpinning the emergence of social and economic interactions of complex systems. Yet, agent-based studies supported by granular analysis of distances are limited. Here, we develop a complexity method that places the real physical world, represented by the actual geographical location of individual firms in Japan, at the epicentre of our research. By combining methods derived from network science (to evaluate the emerging properties of the agents) together with information theory measures (to capture the strength of interaction among these agents), we can systematically analyse a comprehensive dataset of Japanese inter-firm business transactions network and evaluate the effects of spatial features on the structural patterns of the economy. We find that the normalised probability distributions of distances between interacting firms show a power law like decay concomitant to the sizes of firms and regions, with slower decays in major cities. Furthermore, small firms would reach large distances to become a customer of large firms while trading between either only small firms, or only large firms, tends to be at smaller distances. However, a time evolution analysis suggests that a level of market optimisation occurs over time as a reduction in the overall average trading distances in last 20 years can be observed. Lastly, our analysis concerning the trading dynamics among prefectures indicate that the preference to trade with neighbouring prefectures tends to be more pronounced at rural regions as opposed to the larger central conurbations, leading to the formation of three distinct types of regional geographical clusters.