Ansgar Büschges

Summary: Walking animals must continuously reorient to navigate their environment, yet how leg kinematics are reconfigured to produce directional changes of varying magnitude remains poorly understood. Here, we used high-resolution leg tracking to characterize steering behavior in freely walking Drosophila . We focus our analysis on two major strategies, arcs and turns. Although arc and turn trajectories are statistically separable, several parameters suggest an underlying spatial continuum spanning both behavioral categories, involving a change in the steering mechanism rather than a discontinuity in the kinematic space. Arcs, representing shallow curves, and were generated primarily through: (1) graded shifts in the posterior extreme position (PEPs) of the front and inside middle legs, and bilateral asymmetries in stance amplitude, lengthening on the outside (2) and shortening on the inside (3) while step timing and ipsilateral coordination remained at straight-walking baseline. Turns, in contrast, were associated with higher steering demands and relied on: (4) progressive displacement of the anterior extreme positions (AEPs) in front and middle legs, (5) extreme reduction of inside middle and hind leg amplitudes to anchor the turning pivot, (6) skipped steps of inside middle and hind legs, (7) lateral steps of inside middle legs, and (8) spontaneous appearance of synchronized backward-stepping in particular legs beyond a critical turning threshold. Finally, a common kinematic configuration emerges at maximal turning demand. These results suggest a hierarchical, bipartite organization of steering: PEP- and AEP-driven mechanisms operating across moderate and severe trajectory curvatures, and a shared saturated motor state at the behavioral extreme that links descending steering commands to segment-specific leg kinematics in a freely walking insect Graphical Highlights: Drosophila flies use several strategies to change direction, including arcs and turns Arcs have constant curvature, speed and constant angular velocity, while turns require a decrease in speed before changing direction and their angular velocity tends to be higher. We identified eight leg kinematic mechanisms that contribute to these strategies. Some kinematic parameters vary smoothly and in a continuous fashion between mechanisms.

Distinct behaviors require the nervous system to execute specialized motor programs, each characterized by unique patterns of body muscle coordination. Whether the execution and adaptation of these programs follow conserved principles across species and perturbations remains unclear. To compare motor programs across species, perturbations, and behaviors, we developed the Python toolbox Automated Gait Analysis (AutoGaitA). Using AutoGaitA and inferring from kinematics, we found that locomotor programs in flies, mice, and humans rely on diverse mechanisms to generate limb propulsive strength, but employ a similar distal-to-proximal gradient of joint movement velocities. In addition, we showed that aging induces a loss of propulsive strength in all species while preserving the velocity gradient. Furthermore, we observed that in mice, locomotor programs adapt as an integrated function of concomitant perturbations, namely aging and task difficulty. Taken together, using our newly developed versatile quantitative framework AutoGaitA, we began to reveal the conserved and divergent mechanisms underlying the execution of locomotor programs in physiological and perturbed states.

Summary: With the newly uncovered access to a genetic line labeling all campaniform sensilla in Drosophila melanogaster and the many insights gathered from larger insects, we analyzed the distribution and patterning of this class of proprioceptors throughout the fly nervous system and studied CS function in motor control. We demonstrate via two-photon calcium imaging microscopy that campaniform sensilla activation induces activity in many leg muscles, showing that these proprioceptive stimuli can influence motor neuron activity. We then dissected how the lack of these proprioceptive stimuli influence walking behavior, leg kinematics, and interleg coordination in freely-walking flies using transient optogenetic inhibition and video tracking with high spatiotemporal resolution. We show that campaniform sensilla inhibition robustly affects Drosophila melanogaster ’s ability to reach their typical walking speeds. Detailed analysis of leg kinematic suggests that shorter stance amplitudes characterized by their long-lasting duration were behind the speed deficits. Dissecting this phenotype also revealed the inability to maintain coordinated interleg stepping patterns as well as postural control in absence of proprioceptive force and load feedback.

Load proprioceptive feedback measured by campaniform sensilla (CS) substantially contributes to the generation of leg stepping during walking. In stick insects, the effect of load feedback on the antagonistic protractor and retractor motor neurons (MNs) reverses depending on the walking direction to support the functional stance MN pool in both stepping directions. However, this was shown in highly reduced preparations in which the stimulated leg did not perform actual stepping movements, but the effect was inferred from MN activity profiles. It is still largely unknown whether load feedback similarly shows a direction dependent reversal in its effects on protractor and retractor activity during actual stepping of the stimulated leg. Here, we implemented load stimulation in a single-middle leg preparation and recorded protractor and retractor MN activity extracellularly during fictive forward and backward stepping. We show that CS stimulation during the stance phase prolonged stance (retractor) activity and induced stance (retractor) activity when applied during the swing phase in forward stepping. During backward stepping, when protractor and retractor switch phases, the same stimulation again prolonged stance (protractor) activity. However, it rarely induced stance (protractor) activity when applied during the swing phase. We provide evidence that this is due to the absence of signals about the stepping direction of the neighboring legs, thus highlighting the importance of intersegmental signals for task-specific processing of load proprioceptive feedback.

Summary: Animals need to fine-control the speed and direction of locomotion to navigate complex and dynamic environments. To achieve this, they integrate multimodal sensory inputs with their internal drive to constantly adjust their motor output. This integration involves the interplay of neuronal populations across different hierarchical levels along the sensorimotor axis – from sensory, central, and modulatory neurons in the brain to descending neurons and motor networks in the nerve cord. Here, we characterize two populations of neurons that control distinct aspects of walking on different hierarchical levels in Drosophila . First, we use in-vivo electrophysiological recordings to demonstrate that moonwalker descending neurons (MDN) integrate antennal touch to drive changes in walking direction from forward to backward. Second, we establish DopaMeander as an important component in the control of forward walking through a combination of optogenetic activation, silencing, connectomics, and in-vivo recordings. These dopaminergic modulatory neurons drive forward walking with increased turning, and the activity of individual neurons is correlated with ipsiversive turning. Hence, MDN and DopaMeander control opposite regimes of walking on different hierarchical levels. Computational models reveal that their activity predicts key parameters of spontaneous walking. Moreover, we find that both MDN and DopaMeander are gated out during flight. This suggests that neuronal populations across levels of control are modulated by the behavioral state to minimize cross-talk between motor programs.

This review discusses how the nervous system controls the complex body movements keeping animals up and running. In particular, we revisit how research in insects has shed light on motor control principles that govern movements across the animal kingdom. Starting with the organization and evolution of the insect nervous system, we discuss insights into the neuronal control of behaviors varying in complexity, including escape, flight, crawling, walking, grooming, and courtship. These behaviors share specific control features. For instance, central pattern-generating circuits (CPG), which reside in proximity to the motor neurons and muscles, support the generation of rhythmic motor activity. The number of CPGs involved depends on the complexity of the motor apparatus controlled, such as wing pairs for flight or six pairs of multisegmented legs for walking. The different control architectures are introduced with respect to their organization, topology, and operation. Sensory feedback plays a pivotal role in shaping CPG activity into a functional, well-coordinated motor output. The activity of motor circuits is orchestrated by descending neurons connecting the brain to the ventral nerve cord or spinal cord, which initiate, maintain, modulate, and terminate different actions. We therefore discuss the current understanding of descending control and the contributions of individual, command-like descending neurons and population codes. To highlight how insights in insects help discover fundamental motor control principles, we cross-reference findings from other animals, particularly vertebrates. In addition, we discuss methodological advances that enabled breakthroughs in motor control research, including neurogenetics and connectomics, and discuss key open questions.

Campaniform sensilla (CS) are mechanosensors embedded in the cuticle of insects. They are often found at locations near the joints of leg segments. On legs, CS are generally considered to respond directionally to cuticle bending during legged locomotion. It is currently unclear how CS locations affect strain levels at the CS, but this information is crucial for understanding how CS respond to stimuli. Here we present a parametric finite element model of the femoral CS field for Drosophila hind legs with 12 general and seven CS-specific parameters each. This model allows testing how changes in CS location, orientation and material property affect strain levels at each CS. We used experimentally acquired kinematic data and computed ground reaction forces to simulate in vivo -like forward stepping. The displacements found in this study at the physiological CS field location near the trochanter–femur joint are smaller than those necessary for conformation changes of ion channels involved in signal elicitation. Also, variation of material properties of the CS had little influence on displacement magnitudes at the CS cap where the sensory neuron attaches. Thus, our results indicate that ground reaction forces alone are unlikely to serve CS field activation during forward walking.

Load proprioceptive feedback measured by campaniform sensilla (CS) substantially contributes to the generation of leg stepping during walking. In stick insects, the effect of load feedback on the antagonistic protractor and retractor motor neurons (MNs) reverses depending on the walking direction to support the functional stance MN pool in both stepping directions. However, this was shown in highly reduced preparations in which the stimulated leg did not perform actual stepping movements, but the effect was inferred from MN activity profiles. It is still largely unknown whether the load feedback similarly shows a direction dependent reversal in its effects on protractor and retractor activity during actual stepping of the stimulated leg. Here, we implemented load stimulation in a single-middle leg preparation and recorded protractor and retractor MN activity extracellularly during fictive forward and backward stepping. We show that CS stimulation during the stance phase prolonged stance (retractor) activity for the duration of the stimulation and induced stance (retractor) activity when applied during the swing phase in forward stepping. During backward stepping, when protractor and retractor switch phases, the same stimulation again prolonged stance (protractor) activity for the duration of the stimulation. However, it rarely induced stance (protractor) activity when applied during the swing phase. We provide evidence that this is due to the absence of signals about the stepping direction of the neighboring legs, thus highlighting the importance of intersegmental signals for task-specific processing of load proprioceptive feedback. New and Noteworthy Proprioceptive feedback needs to be processed in a task-dependent manner to serve its function, even when the leg kinematics change to generate different behaviors. We show for forward and backward stepping that load signals initiate and maintain the activity of the functional stance phase motor neuron pool and that intersegmental signals about stepping of the neighboring legs are relevant for the task-specific processing of load feedback.