The design of modern scientific experiments requires the control and monitoring of many different data streams. However, the serial execution of programming instructions in a computer makes it a challenge to develop software that can deal with the asynchronous, parallel nature of scientific data. Here we present Bonsai, a modular, high-performance, open-source visual programming framework for the acquisition and online processing of data streams. We describe Bonsai's core principles and architecture and demonstrate how it allows for the rapid and flexible prototyping of integrated experimental designs in neuroscience. We specifically highlight some applications that require the combination of many different hardware and software components, including video tracking of behavior, electrophysiology and closed-loop control of stimulation.

Adult zebrafish are robustly social animals whereas larvae are not. We designed an assay to determine at what stage of development zebrafish begin to interact with and prefer other fish. One week old zebrafish show no social preference whereas most three week old zebrafish strongly prefer to remain in a compartment where they can view conspecifics. However, for some individuals, the presence of conspecifics drives avoidance instead of attraction. Social preference is dependent on vision and requires viewing fish of a similar age/size. In addition, over the same one to three week period larval zebrafish increasingly tend to coordinate their movements, a simple form of social interaction. Finally, social preference and coupled interactions are differentially modified by an NMDAR antagonist and acute exposure to ethanol, both of which are known to alter social behaviour in adult zebrafish.

Food ingestion is one of the defining behaviours of all animals, but its quantification and analysis remain challenging. This is especially the case for feeding behaviour in small, genetically tractable animals such as Drosophila melanogaster. Here, we present a method based on capacitive measurements, which allows the detailed, automated and high-throughput quantification of feeding behaviour. Using this method, we were able to measure the volume ingested in single sips of an individual, and monitor the absorption of food with high temporal resolution. We demonstrate that flies ingest food by rhythmically extending their proboscis with a frequency that is not modulated by the internal state of the animal. Instead, hunger and satiety homeostatically modulate the microstructure of feeding. These results highlight similarities of food intake regulation between insects, rodents, and humans, pointing to a common strategy in how the nervous systems of different animals control food intake.

The ability to control a behavioral task or stimulate neural activity based on animal behavior in real-time is an important tool for experimental neuroscientists. Ideally, such tools are noninvasive, low-latency, and provide interfaces to trigger external hardware based on posture. Recent advances in pose estimation with deep learning allows researchers to train deep neural networks to accurately quantify a wide variety of animal behaviors. Here, we provide a new DeepLabCut-Live! package that achieves low-latency real-time pose estimation (within 15 ms, >100 FPS), with an additional forward-prediction module that achieves zero-latency feedback, and a dynamic-cropping mode that allows for higher inference speeds. We also provide three options for using this tool with ease: (1) a stand-alone GUI (called DLC-Live! GUI), and integration into (2) Bonsai, and (3) AutoPilot. Lastly, we benchmarked performance on a wide range of systems so that experimentalists can easily decide what hardware is required for their needs.

Cross-validating new methods for recording neural activity is necessary to accurately interpret and compare the signals they measure. Here we describe a procedure for precisely aligning two probes for in vivo “paired-recordings” such that the spiking activity of a single neuron is monitored with both a dense extracellular silicon polytrode and a juxtacellular micropipette. Our new method allows for efficient, reliable, and automated guidance of both probes to the same neural structure with micrometer resolution. We also describe a new dataset of paired-recordings, which is available online. We propose that our novel targeting system, and ever expanding cross-validation dataset, will be vital to the development of new algorithms for automatically detecting/sorting single-units, characterizing new electrode materials/designs, and resolving nagging questions regarding the origin and nature of extracellular neural signals.

The ability to dynamically control a behavioral task based on real-time animal behavior is an important feature for experimental neuroscientists. However, designing automated boxes for behavioral studies requires a coordinated combination of mechanical, electronic, and software design skills which can challenge even the best engineers, and for that reason used to be out of reach for the majority of experimental neurobiology and behavioral pharmacology researchers. Due to parallel advances in open-source hardware and software developed for neuroscience researchers, by neuroscience researchers, the landscape has now changed significantly. Here, we discuss powerful approaches to the study of behavior using examples and tutorials in the Bonsai visual programming language, towards designing simple neuroscience experiments that can help researchers immediately get started. This language makes it easy for researchers, even without programming experience, to combine the operation of several open-source devices in parallel and design their own integrated custom solutions, enabling unique and flexible approaches to the study of behavior, including video tracking of behavior and closed-loop electrophysiology.