Search

Looming stimuli evoke behavioral responses in most animals, yet the mechanisms of looming detection in vertebrates are poorly understood. Here we hypothesize that looming detection in the tectum may rely on spontaneous emergence of synfire chains: groups of neurons connected to each other in the same sequence in which they are activated during a loom. We then test some specific consequences of this hypothesis. First, we use high-speed calcium imaging to reconstruct functional connectivity of small networks within the tectum of Xenopus tadpoles. We report that reconstructed directed graphs are clustered and hierarchical, that their modularity increases in development, and that looming-selective cells tend to collect activation within these graphs. Second, we describe spontaneous emergence of looming selectivity in a computational developmental model of the tectum, governed by both synaptic and intrinsic plasticity, and driven by structured visual inputs. We show that synfire chains contribute to looming detection in the model; that structured inputs are critical for the emergence of selectivity, and that biological tectal networks follow most, but not all predictions of the model. Finally, we propose a conceptual scheme for understanding the emergence and fine-tuning of collision detection in developing aquatic animals.

Homeostatic intrinsic plasticity is often described as an adjustment of neuronal excitability to maintain stable spiking output. Here we report that intrinsic plasticity in the tectum of Xenopus tadpoles also supports temporal tuning, wherein neurons independently adjust spiking responses to fast and slow patterns of synaptic activation. Using the dynamic clamp technique, and five different types of visual, acoustic, and multisensory conditioning, we show that in tadpoles exposed to light flashes, tectal neurons became selective for fast synaptic inputs, while neurons exposed to looming and multisensory stimuli remained responsive to longer inputs. We also report a homeostatic co-tuning between synaptic and intrinsic temporal properties in tectal cells, as neurons that naturally received fast synaptic inputs tended to be most responsive to long-lasting synaptic conductances, and the other way around. These results expand our understanding of plasticity in the brain, and inform future work on the mechanisms of sensorimotor transformation.Significance statementWith the recent explosion of work in neural connectivity reconstruction and biologically inspired deep learning, most researchers concentrate on the topology of connections between neurons, rather than on differences in neuronal tuning. Here we show that in a sensory network in Xenopus tadpoles, different neurons are tuned, and respond stronger, to either short or long synaptic inputs. This tuning tended to be opposite to the actual dynamics of synaptic inputs each cell received, such that neurons that normally receive shorter inputs generated stronger spiking in response to longer testing currents, and the other way around. This observation shows that even in networks that don’t generate oscillations, neurons reshape their temporal selectivity, to optimize their impact on distributed calculations.

Escape behaviors, orienting reflexes, and social behaviors in Xenopus laevis tadpoles have been well-documented in the literature. Schooling behavior experiments allow for the observation of tadpole social interactions and in the past have been used to characterize behavioral deficits in models of neurodevelopmental disorders. Unlike other species of frogs, Xenopus tadpoles show polarized schooling. Not only do tadpoles aggregate, they also swim in the same direction. Quantifying both aggregation and relative swim angle can give us an important measure of social behavior and sensory integration. Past iterations of these experiments have required the continued presence of an experimenter throughout the duration of each trial and relied on expensive software for subsequent data analysis. The instrument configuration and analysis protocol outlined here provide an automated method to assess schooling by delivering a series of timed vibratory stimuli to a group of tadpoles to induce swimming behavior and then controlling a camera to document their positions via still images. Both stimulus delivery and image acquisition are automated using the Python programming language. Analysis is done using ImageJ and custom Python scripts, which are provided in this protocol. The specific equipment configuration and scripts shown here provide one solution, but other equipment and custom scripts can be substituted.

In teaching, the best exam questions are those that seem simple at first but can lead to deep and nuanced conversations. Similarly, to probe brain development, we should look for behaviors that are easy to evoke and quantify, but that are demanding, malleable, and inherently variable. Visual collision avoidance is an example of such a behavior; it is ecologically relevant, robust, and easy to record, but also nuanced and shaped by the sensory history of the animal. Here we describe how to set up a visual avoidance assay and how to use it to test sensory processing and sensorimotor transformations in the vertebrate brain.