Variance adaptation in navigational decision making

CRASH2p: Closed-loop Two Photon Imaging in a Freely Moving Animal

Direct measurement of neural activity in freely moving animals is essential for understanding how the brain controls and represents behaviors. Genetically encoded calcium indicators report neural activity as changes in fluorescence intensity, but brain motion confounds quantitative measurement of fluorescence. Translation, rotation, and deformation of the brain and the movements of intervening scattering or auto-fluorescent tissue all alter the amount of fluorescent light captured by a microscope. Compared to single-photon approaches, two photon microscopy is less sensitive to scattering and off-target fluorescence, but more sensitive to motion, and two photon imaging has always required anchoring the microscope to the brain. We developed a closed-loop resonant axial-scanning high-speed two photon (CRASH2p) microscope for real-time 3D motion correction in unrestrained animals, without implantation of reference markers. We complemented CRASH2p with a novel scanning strategy and a multi-stage registration pipeline. We performed volumetric ratiometrically corrected functional imaging in the CNS of freely moving Drosophila larvae and discovered previously unknown neural correlates of behavior.

Tracking the Navigation Behavior of Drosophila Larvae in Real and Virtual Odor Gradients by Using the Raspberry Pi Virtual Reality (PiVR) System

In a closed-loop experimental paradigm, an animal experiences a modulation of its sensory input as a function of its own behavior. Tools enabling closed-loop experiments are crucial for delineating causal relationships between the activity of genetically labeled neurons and specific behavioral responses. We have recently developed an experimental platform known as "Raspberry Pi Virtual Reality" (PiVR) that is used to perform closed-loop optogenetic stimulation of neurons in unrestrained animals. PiVR is a system that operates at high temporal resolution (>30-Hz) and with low latencies. Larvae of the fruit fly Drosophila melanogaster are ideal to study the role of individual neurons in modulating behavior to aid the understanding of the neural pathways underlying various guided behaviors. Here, we introduce larval chemotaxis as an example of a navigational behavior in which an animal seeks to locate a target-in this case, the attractive source of an odor-by tracking a concentration gradient. The methodologies that we describe here combine the use of PiVR with the study of larval chemotaxis in real and virtual odor gradients, but these can also be readily adapted to other sensory modalities.

Perception of visual variance is mediated by subcortical mechanisms

Variance characterizes the structure of the environment. This statistical concept plays a critical role in evaluating the reliability of evidence for human decision-making. The present study examined the involvement of subcortical structures in the processing of visual variance. To this end, we used a stereoscope to sequentially present two circle arrays in a dichoptic or monocular fashion while participants compared the perceived variance of the two arrays. In Experiment 1, two arrays were presented monocularly to the same eye, dichopticly to different eyes, or binocularly to both eyes. The variance judgment was less accurate in different-eye condition than the other conditions. In Experiment 2, the first circle array was split into a large-variance and a small-variance set, with either the large-variance or small-variance set preceding the presentation of the second circle array in the same eye. The variance of the first array was judged larger when the second array was preceded by the large-variance set in the same eye, showing that the perception of variance was modulated by the visual variance processed in the same eye. Taken together, these findings provide evidence for monocular processing of visual variance, suggesting that subcortical structures capture the statistical structure of the visual world.

Linking neural circuits to the mechanics of animal behavior in Drosophila larval locomotion

The motions that make up animal behavior arise from the interplay between neural circuits and the mechanical parts of the body. Therefore, in order to comprehend the operational mechanisms governing behavior, it is essential to examine not only the underlying neural network but also the mechanical characteristics of the animal's body. The locomotor system of fly larvae serves as an ideal model for pursuing this integrative approach. By virtue of diverse investigation methods encompassing connectomics analysis and quantification of locomotion kinematics, research on larval locomotion has shed light on the underlying mechanisms of animal behavior. These studies have elucidated the roles of interneurons in coordinating muscle activities within and between segments, as well as the neural circuits responsible for exploration. This review aims to provide an overview of recent research on the neuromechanics of animal locomotion in fly larvae. We also briefly review interspecific diversity in fly larval locomotion and explore the latest advancements in soft robots inspired by larval locomotion. The integrative analysis of animal behavior using fly larvae could establish a practical framework for scrutinizing the behavior of other animal species.

Multi-neuronal recording in unrestrained animals with all acousto-optic random-access line-scanning two-photon microscopy

To understand how neural activity encodes and coordinates behavior, it is desirable to record multi-neuronal activity in freely behaving animals. Imaging in unrestrained animals is challenging, especially for those, like larval Drosophila melanogaster, whose brains are deformed by body motion. A previously demonstrated two-photon tracking microscope recorded from individual neurons in freely crawling Drosophila larvae but faced limits in multi-neuronal recording. Here we demonstrate a new tracking microscope using acousto-optic deflectors (AODs) and an acoustic GRIN lens (TAG lens) to achieve axially resonant 2D random access scanning, sampling along arbitrarily located axial lines at a line rate of 70 kHz. With a tracking latency of 0.1 ms, this microscope recorded activities of various neurons in moving larval Drosophila CNS and VNC including premotor neurons, bilateral visual interneurons, and descending command neurons. This technique can be applied to the existing two-photon microscope to allow for fast 3D tracking and scanning.

Multisensory correlation computations in the human brain identified by a time-resolved encoding model

Neural mechanisms that arbitrate between integrating and segregating multisensory information are essential for complex scene analysis and for the resolution of the multisensory correspondence problem. However, these mechanisms and their dynamics remain largely unknown, partly because classical models of multisensory integration are static. Here, we used the Multisensory Correlation Detector, a model that provides a good explanatory power for human behavior while incorporating dynamic computations. Participants judged whether sequences of auditory and visual signals originated from the same source (causal inference) or whether one modality was leading the other (temporal order), while being recorded with magnetoencephalography. First, we confirm that the Multisensory Correlation Detector explains causal inference and temporal order behavioral judgments well. Second, we found strong fits of brain activity to the two outputs of the Multisensory Correlation Detector in temporo-parietal cortices. Finally, we report an asymmetry in the goodness of the fits, which were more reliable during the causal inference task than during the temporal order judgment task. Overall, our results suggest the existence of multisensory correlation detectors in the human brain, which explain why and how causal inference is strongly driven by the temporal correlation of multisensory signals.

Circuits for integrating learned and innate valences in the insect brain

Animal behavior is shaped both by evolution and by individual experience. Parallel brain pathways encode innate and learned valences of cues, but the way in which they are integrated during action-selection is not well understood. We used electron microscopy to comprehensively map with synaptic resolution all neurons downstream of all mushroom body (MB) output neurons (encoding learned valences) and characterized their patterns of interaction with lateral horn (LH) neurons (encoding innate valences) in Drosophila larva. The connectome revealed multiple convergence neuron types that receive convergent MB and LH inputs. A subset of these receives excitatory input from positive-valence MB and LH pathways and inhibitory input from negative-valence MB pathways. We confirmed functional connectivity from LH and MB pathways and behavioral roles of two of these neurons. These neurons encode integrated odor value and bidirectionally regulate turning. Based on this, we speculate that learning could potentially skew the balance of excitation and inhibition onto these neurons and thereby modulate turning. Together, our study provides insights into the circuits that integrate learned and innate valences to modify behavior.

Switch-like and persistent memory formation in individual Drosophila larvae

Associative learning allows animals to use past experience to predict future events. The circuits underlying memory formation support immediate and sustained changes in function, often in response to a single example. Larval Drosophila is a genetic model for memory formation that can be accessed at molecular, synaptic, cellular, and circuit levels, often simultaneously, but existing behavioral assays for larval learning and memory do not address individual animals, and it has been difficult to form long-lasting memories, especially those requiring synaptic reorganization. We demonstrate a new assay for learning and memory capable of tracking the changing preferences of individual larvae. We use this assay to explore how activation of a pair of reward neurons changes the response to the innately aversive gas carbon dioxide (CO2). We confirm that when coupled to CO2 presentation in appropriate temporal sequence, optogenetic reward reduces avoidance of CO2. We find that learning is switch-like: all-or-none and quantized in two states. Memories can be extinguished by repeated unrewarded exposure to CO2 but are stabilized against extinction by repeated training or overnight consolidation. Finally, we demonstrate long-lasting protein synthesis dependent and independent memory formation.

Neural activity temporal pattern dictates long-range propagation targets

Brain possesses a complex spatiotemporal architecture for efficient information processing and computing. However, it remains unknown how neural signal propagates to its intended targets brain-wide. Using optogenetics and functional MRI, we arbitrarily initiated various discrete neural activity pulse trains with different temporal patterns and revealed their distinct long-range propagation targets within the well-defined, topographically organized somatosensory thalamo-cortical circuit. We further observed that such neural activity propagation over long range could modulate brain-wide sensory functions. Electrophysiological analysis indicated that distinct propagation pathways arose from system level neural adaptation and facilitation in response to the neural activity temporal characteristics. Together, our findings provide fundamental insights into the long-range information transfer and processing. They directly support that temporal coding underpins the whole brain functional architecture in presence of the vast and relatively static anatomical architecture.

Useful road maps: studying Drosophila larva's central nervous system with the help of connectomics

The larva of Drosophila melanogaster is emerging as a powerful model system for comprehensive brain-wide understanding of the circuit implementation of neural computations. With an unprecedented amount of tools in hand, including synaptic-resolution connectomics, whole-brain imaging, and genetic tools for selective targeting of single neuron types, it is possible to dissect which circuits and computations are at work behind behaviors that have an interesting level of complexity. Here we present some of the recent advances regarding multisensory integration, learning, and action selection in Drosophila larva.

Mini-brain computations converting dynamic olfactory inputs into orientation behavior

The neural logic underlying the conversion of non-stationary (dynamic) olfactory inputs into odor-search behaviors has been difficult to crack due to the distributed nature of the olfactory code - food odors typically co-activate multiple olfactory sensory neurons. In the Drosophila larva, the activity of a single olfactory sensory neuron is sufficient to direct accurate reorientation maneuvers in odor gradients (chemotaxis). In this reduced sensory system, a descending pathway essential for larval chemotaxis has been delineated from the peripheral olfactory system down to the premotor system. Here, I review how anatomical and functional inspections of this pathway have advanced our understanding of the neural mechanisms that convert behaviorally relevant sensory signals into orientation responses.