Changing 3D heading direction in terrestrial locomotion: Effectiveness of visual landmarks and the onset of turning in stick insects

Changing 3D heading direction in terrestrial locomotion: Effectiveness of visual landmarks and the onset of turning in stick insects | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Changing 3D heading direction in terrestrial locomotion: Effectiveness of visual landmarks and the onset of turning in stick insects Ronja Bigge, Nicoletta Schwermann, Volker Dürr This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9243999/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 9 You are reading this latest preprint version Abstract Climbing animals need to be able to adjust their heading in both vertical and horizontal directions, requiring pitch and yaw rotation of the body axis, respectively. Stick insects are climbing herbivors that dwell among leaves and branches of the vegetation they feed on. By nature, they need to adjust their 3D heading in a complex environment. Since insects in general turn towards visual landmarks, we test whether or not stick insects initiate pitch rotation in response to landmarks that reliably elicit yaw rotation. We show that this is not the case. Instead, tactile cues alone are sufficient to mask the effect of visual deprivation, whereas lack of tactile cues strongly delays body inclination and the onset of climbing, even when same visual landmark cues are present that are sufficient to induce turning reliably. In a second set of experiments, we use a setup that constrains the onset of turning to tell whether visually induced yaw rotation is initiated by swing movements or rather by stance movements of the front legs. We show that a turn-related left-right asymmetry of front leg movement begins at least one stance phase after passing the constraint and may begin with the first swing movement thereafter. We conclude that stick insects do not exploit visual cues for initiating pitch rotation, despite they reliably initiate yaw rotation towards the exact same landmark. Moreover, we demonstrate that yaw rotation towards visual landmarks may begin with a swing movement of a front leg, much like the pitch rotation at the onset of climbing. yaw rotation pitch rotation visually-induced turning swing movement locomotion Carausius Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Climbing about in any three-dimensional spatial structure, such as the branches of a tree or the leaves of a shrub, requires the ability of an animal to adjust its heading in both the perceived vertical and perceived horizontal direction. Other than in aquatic or aerial maneuverability, where active pitch and yaw rotation of the body axis are not linked to discontinuities of the surrounding medium (swimming: Webb, 2004 ; flying: Dudley, 2002 ; Taylor, 2001 ), any change of heading during walking or climbing is bound to surface structure and, therefore, must be affected by surface discontinuities such as edges, walls or margins. Experimentally, surface discontinuities are commonly used to enforce pitch rotation by disturbance of horizontal walking(Pelletier and McLeod, 1994 ; Harley et al., 2009 ; Krause and Dürr, 2012 ; Clifton et al., 2023 ) or running (Zurek and Gilbert, 2014 ), whereas they are commonly avoided when enforcing yaw rotation on a horizontal surface through visual cues. This raises the question whether pitch and yaw rotations in walking animals rely on the same visuomotor mechanisms. Here we use the Indian stick insect Carausius morosus (Sinéty, 1901), a climbing herbivorous insect that dwells in the foliage of the plants which it feeds on, to investigate two general aspects of omnidirectional heading in surface-bound, terrestrial locomotion. The first aspect concerns the equal effectiveness of visual landmark cues for inducing either vertical (pitch) or horizontal (yaw) rotations. The second aspect concerns the type of leg movement that initiates turning on a horizontal surface, as opposed to climbing onto a vertical surface. Given the evidence that stick insects make use of tactile cues to initiate pitch rotation for climbing(Schütz and Dürr, 2011 ) and yaw rotation in turning (Berendes and Dürr, 2022 ), we will relate evidence on tactually versus visually induced changes in 3D heading. During locomotion, stick insects actively explore the space ahead with persistent, quasi-rhythmic movement of both antennae (Dürr et al., 2001 ; Harischandra et al., 2015 ; Dürr and Mesanovic, 2023 ). As in many other insect species (Dürr et al., 2022 ), active antennal movement in stick insects improves detection and negotiation of obstacles (Dürr et al., 2001 ; Krause and Dürr, 2012 ).. Tactile localisation of a vertical surface typically induces an upward climbing response that is initiated by a directed reach-to-grasp movement (Schütz and Dürr, 2011 ) within the spatial overlap region of antennal and front leg movement ranges (Dürr and Schilling, 2018 ). Such directed reach-to-grasp movements of a front leg may be considered as variations of targeted swing movements during walking on horizontal surfaces (Dürr et al., 2018). In horizontal walking and running cockroaches, tactile cues may induce strong and persistent yaw rotation towards the contacted location (Okada and Toh, 2000 ; 2006 ) or rapid steering in thigmotaxis paradigms (Camhi and Johnson, 1999 ; Cowan et al., 2006 ). In comparison, tactually induced turning in stick insects is more moderate and often accompanied by attempts to grasp the touched object with a front leg (Berendes and Dürr, 2022 ). Visually induced yaw turns appear to be more pronounced in stick insects than tactually induced yaw turns, despite their nocturnal lifestyle. Large-field visual motion (Dürr and Ebeling, 2005 ; Gruhn et al., 2009 ) as well as static landmarks (Kalmus, 1937 ; Jander and Volk-Heinrichs, 1970 ; Zeng et al., 2020 ; Meschenmoser and Dürr, 2025 ) have been shown to elicit strong and reliable yaw turns (Meschenmoser and Dürr, 2026 ). Whether visual cues about obstacle height may also initiate climbing by pitch rotation, is unknown. Though Dürr et al. ( 2003 ) showed that antennal tactile cues are sufficient to mask any potential effect of blindfolding in a stair-climbing paradigm, the lack of salient visual features in their experimental setup did not rule out the potential use of visual cues for directed reaching and/or the initiation of climbing. Meanwhile, flies (Pick and Strauss, 2005 ) and locusts (Niven et al., 2010 ) have been shown to use visual cues for targeted, upward-directed swing movements during climbing. Accordingly, the first objective of this study was to test whether stick insects make use of visual cues about obstacle height in the initiation of climbing, too. If so, this would indicate that vision, like touch, can provide directional cues for 3D heading. Perhaps it could even provide spatial coordinates required for spatially directed movement of the foot, as it is known for prey capture in mantids (Cleal and Prete, 1996 ). The second objective was to test whether visually induced yaw rotation begins a directional change of a swing movement, stance movement or either of those. Pitch rotation during the initiation of climbing is initiated by an inclination of the body axis, i.e., rearing (cockroach: Watson et al., 2002 ; Harley et al., 2009 ; stick insect: Schütz and Dürr, 2011 ). Rearing can account for small changes in body pitch angle through a front-to-rear difference in extension of the legs in stance (Watson et al., 2002 ), but climbing onto an obstacle or wall requires a change in foothold height that can be achieved by an elevated swing movement, only. In contrast, a change in body yaw angle requires asymmetric action of left and right legs that can be achieved during either swing or stance (Gruhn et al., 2009 ), though the transient change in stance direction of front legs is fastest and most consistent (Dürr and Ebeling, 2005 ), complemented by a left-right asymmetry of inter-leg coupling (Dürr, 2005 ). Recent optogenetic manipulation experiments on Drosophila induced a left-right asymmetry of stance trajectory length, highlighting a potential mechanism underlying a left-right asymmetry of stance movement (Chockley et al., 2022 ). Similarly, left-right asymmetries of stance movements are also commonly implemented in insect locomotion models, both in software and hardware (Cruse et al., 1998 ; Dürr et al., 2019 ). Here we use static visual landmarks that reliably induce yaw rotation in order to test whether the onset of turning is related to a preceding change in swing direction or rather to an ongoing change in stance direction. As a prerequisite for telling apart the two alternatives, we used a setup that hindered animals to change their heading until they reached a widening of the walkway. This allowed us to define a virtual “decision line” and to relate swing and stance directions on the narrow walkway, i.e. before the onset of rotation, to those during rotation on the widened walkway. Moreover, it allowed us to tell the number of steps necessary to achieve a significant turn angle. If turning began during swing, directional information about the landmark would be transferred into a change in touch-down position which would lead to a change in pull direction of the subsequent stance movement. Essentially, this would be the same kind of directional information transfer as used for the initiation of climbing by targeted reaching (Dürr and Schilling, 2018 ), except that the new foothold would be on the same plane as the preceding one. Alternatively, if turning began during stance, directional information about the visual landmark would induce a left-right asymmetry of stance length, pull direction, or both. Methods Animals All experiments were performed on adult female stick insects of the species Carausius morosus (de Sinéty, 1901) which were bred in a laboratory culture at Bielefeld University. Each of the tested animals had intact legs with equally long limbs and feelers. During the experiments, light conditions were kept constant by fully covering the windows and using an artificial light source instead. Experiments were conducted at room temperature. Different groups of animals were used for each one of the four experiments described below. Except for experiment 3 , animals were labelled with a set of retro-reflective, spherical markers with a diameter of 1.5 mm. Markers were glued to the cuticle of the animals by use of transparent nail polish. During the gluing process care was taken not to affect the mobility of the joints in any way. Stair-climbing paradigm Freely walking stick insects were tested in a climbing paradigm, in which they walked along a straight, flat, black walkway which carried a cuboid at its end that served as a stair. Once the animals reached the stair, they readily climbed it. To maximise the visual contrast of the upper edge of the stair, the front side of the stair was painted white with a black, horizontal bar beneath its upper edge. Both, the walkway and the stair were 40 mm wide, which is slightly broader than the average distance between left and right foot contacts during planar walking. A white cardboard shield was placed behind the stair to exclude any other visual landmarks from the paradigm and to increase visual contrast of the black edge against the background. Experiment 1: The cuboid stair was either 24 mm or 48 mm high, though only data from the 24 mm stair will be presented here. The size of the black bar beneath the edge was 40×8 mm. The setup was filmed from the side, using an infrared-sensitive digital video camera (Basler A602f, Ahrensburg, Germany; equipped with a 70 mm zoom lens) operating at 50 frames per second, fps. Camera exposure was synchronised with a custom-built infrared flash light. Each animal was labelled with four retro-reflective markers placed onto the posterior mesothorax, mid prothorax, between the eyes on the head, and at the distal tibia of the right front leg. Movement analysis was done using a custom-written Matlab program (The Mathworks, Natick/MA, USA) to track the head marker position over time. Head trajectories relative to the setup were recorded at a spatial resolution of 0.18 mm/pixel. Each trial was summarised by four head positions: (i) the position at the time of the 1 st tactile contact with the stair (either with a front leg or an antenna), (ii) the closest point to the corner at the bottom of the edge, (iii) the point directly above the edge, and (iv) the head position 10 mm behind the upper edge of the stair. The mean speed of the animal was calculated for the head trajectory between the 1 st and 4 th of these positions. Ten animals were divided into two cohorts of five. After at least ten trials with the intact animal, cohort 1 was first tested ten times after bilateral antennectomy at the flagellum base and another ten times after additional blindfolding, i.e. after covering both eyes with black, water-soluble paint. Cohort 2 was first tested intact, then blindfolded and finally antennectomized. The four head positions listed above were measured for each trial. Horizontal distances of positions (i) and (ii), clearances above the stair for positions (iii) and (iv), and mean speed were averaged per animal and tested with paired t-tests on per-animal means. To test whether the observed effects of antennectomy were specific to the visual pattern used, we conducted further trials with an additional cohort 3 (N = 5), with 4×8 trials per animal. As visual patterns on the front surface of the stair, we either replaced the horizontal black bar by a vertical black bar of the same width, or added a checkerboard pattern below the horizontal bar (see Supplementary Material). Experiment 2: For more accurate, 3-dimensional measurement of head trajectories, head pitch, and the apparent visual image size of the black bar beneath the edge, a similar setup as in experiment 1 was surrounded with a multi-camera motion-capture system (Vicon MX10, equipped with eight T10 cameras; Vicon, Oxford, UK), operating at 200 fps with a spatial resolution of approximately 0.1 mm. The stair was 36 mm high and carried a 40×9 mm black bar beneath its upper edge. Each animal was labelled with 10 retro-reflective markers: A triangle of three markers on the dorsal prothorax was used to define the body-fixed coordinate system; one marker on the head was used to estimate head pitch; 2×2 markers were placed on the distal femur and tibia of either front leg, and two further markers were fixed to the proximal third of each antenna. A total of five intact animals was tested for 20 trials each. An additional digital video camera (Basler A602f2, equipped with a Pentax H6Z810 zoom lens) recorded a side view of the experiment at 50 fps. Its videos were used for visual inspection and validation of the motion capture series and the corresponding kinematic analysis. Two-choice paradigm and analysis of turning A two-choice paradigm was used to induce yaw turns as part of behavioural orientation responses towards visual landmarks. Free-walking, intact animals were tested on a flat, black walkway (width: 40 mm) which broadened over a stretch of 6 cm to become either 12 or 14 cm wide (for cuboids and pillars, respectively, see below). The transition occurred at 20 cm and will be referred to as the ‘decision line’. Two equal-sized objects were placed at each side of the wide area, at equal distance from the decision line (90 mm). These objects were either white cuboids (40 mm wide; 36 mm high; with or without a black bar at the front) or pillars (20 mm diameter; 100 mm high; either black or white). Similar to the stair-climbing paradigm, a white cardboard shield was positioned behind the setup to increase visual contrast. Experiment 3: The front sides of the cuboids had one out of three possible appearances: (i) a blank, white surface, (ii) a white surface with a horizontal black bar beneath the upper edge, or (iii) a white surface with a vertical black bar in its middle. The dimensions of the horizontal and vertical bars were the same (15 mm x 36 mm) and were designed such that its size reached an apparent angular width of 10 degrees as seen from the decision line. Two combinations of appearances were tested: either blank against the horizontal bar, or horizontal against vertical bar. For both of these combinations the left-right arrangement was varied pseudo-randomly. In total, each one of five animals performed 80 trials, with 2×20 trials per combination. For each animal, the proportion of decisions to climb either one of the stairs was calculated. Effect sizes were quantified as one-sample Cohen’s delta, defined as the difference between the mean proportion across animals and the chance level (0.5), divided by the standard deviation across animals. Another five animals were tested in a paradigm with two vertical pillars, one black and one white. The preference for either the black or the white pillar was tested in 20 trials per animal, again with pseudo-randomised left-right arrangement of the two pillars. Experiment 4: The two-choice paradigm with pillars was also used in combination with the Vicon motion-capture system, using the same 10-marker arrangement as in experiment 2 . Eight animals were tested. Acquired marker trajectories were used to analyse front leg kinematics and head orientation, so as to reconstruct the shape and size of the visual image of the landmarks, but also to analyse the shape of the trajectories of both front feet as the animals turned towards one of the two objects. In addition, the yaw angle of the body axis was calculated and used to determine the onset of yaw rotation beyond the decision line. For the analysis of the step sequences, four subsequent step cycles were first subdivided into stance and swing episodes and then assigned to one of four categories, depending on where they occurred relative to the (virtual) decision line. The swing movement that first crossed the decision line divided the step sequence into two step cycles on the narrow part of the walkway (prior to crossing the decision line) and the subsequent two step cycles on the broad part of the walkway, as the animal turned towards one of the objects. Marker-based body model and kinematic analysis After attaching the markers to the animals, detailed pictures of the animals were taken by a stereo lens (Olympus SZ61), equipped with a digital camera. The pictures were used to measure the exact body and limb dimensions in relation the positions of the markers. These measurements were used to create a unique body model of each of the animals. This was done with a custom-written MATLAB script. The software Vicon Nexus 1.8.5 was used to control the motion capture system and to track and reconstruct all marker positions in space. For trajectory analysis, each marker was identified and labelled once per recording by hand, and then tracked automatically by the software. Marker tracking was generally robust, provided that a given marker was recorded by least two cameras at any time. In case of small gaps (<10 frames, i.e. 50 ms), marker trajectories were interpolated with an algorithm of the Nexus software. In case of bigger tracking issues, trials were excluded from the analysis. Marker trajectories were then exported into Matlab where they were analysed further, in conjunction with the corresponding individual body model. Whole-body kinematics were calculated using kinematic calculations as described by Theunissen and Dürr (2013). For the results described here, we used positions and yaw/pitch angles of the body axis and of the head, as well as body-centred position coordinates and speed of the tibia-tarsus joints of both front legs. The transition from swing to stance, i.e., touch-down of the foot, was detected as a drop in foot velocity below a threshold of 60 mm/s. In contrast, the transition from stance to swing, i.e. the lift-off of the foot, was detected as an increase beyond a threshold of 110 mm/s. To simplify the analysis of the foot trajectories, each swing or stance episode was described by four points: the posterior extreme position (PEP) or point of lift-off, the anterior extreme position (AEP) or point of touch-down, and the points at one third or two thirds of its duration. To capture the turning tendency of the animal, we quantified the left-right asymmetry of the left and right front leg movement. To do so we calculated the position difference for the four points per swing or stance trajectories of the left and right foot in the body-fixed coordinate system. The left-right asymmetry was then calculated as the slope of the regression line through the four “difference points”. To account for the step sequence, this was done separately for trials in which either the left or right front legs crossed the decision line first. Step cycles were pooled according to whether they belonged to the inner or outer leg with respect to turning direction and curvature of the walked path. Statistics were calculated for per-animal means. In case of left-right asymmetry, this was done for mean regression line slopes per animal, separately for swing and stance phases, using a Friedman test for repeated (matched) samples consisting of four subsequent steps, followed by Wilcoxon’s test for matched pairs as post-hoc tests. Results Stick insects do not use visual cues for the initiation of climbing In order to test the hypothesis that stick insects use visual cues to initiate climbing, we used the same stair-climbing paradigm as Dürr et al., (2003) and Krause and Dürr (2012), but with a high-contrast visual landmark, i.e. a black horizontal bar, marking the upper edge of the stair (Fig. 1A). The width of black bar was chosen such that it subtended a visual angle of approximately 10 degrees at a distance of 45 mm in front of the stair. After a first set of trials with a stair of the height 48 mm, it became clear that antennectomized animals hardly ever climbed the front wall of the stair, but climbed onto one of the side walls instead. As a consequence, we reduced the height of the stair to 24 mm, which is equivalent 2 to 3 times the typical clearing beneath the prothorax. A total of ten animals were tested sequentially in three conditions, with ten trials per condition. A first cohort of five animals was first tested intact, then without antennae, and finally after additional blindfolding. A second cohort was first tested intact, then after blindfolding, and finally after additional antennectomy (cut antennae). The initiation of climbing was assessed based on the shape of the head trajectory, which Dürr et al. (2003; their Fig. 5B) had shown to get much closer to the stair whenever the antennae were cut. Our hypothesis was that animals without antennae but with intact vision would exploit the presence of the visual cue such that their head would not approach the stair as much as in antennectomized and blind animals. Instead, we expected that the head trajectory would rise as early and smoothly as in intact animals. Fig. 1B shows that this was not the case. As expected, blindfolded animals with intact antennae climbed the stair virtually the same as intact animals did (left panels in Fig. 1B). However, head trajectories of animals with shortened antennae, irrespective of whether they could see or not, got much closer to the stair and followed much steeper head trajectories than animals with antennae. On average, intact animals first touched the stair at a distance of 29.1 ± 4.9 mm. This distance remained nearly the same after blindfolding (29.6 ± 3.9 mm) but approximately halved whenever the antennae were cut (sighted: 14.5 ± 8.0 mm; blind: 14.6 ± 7.6 mm). The same effect was found for the shortest distance to the bottom of the edge, which was nearly the same for intact and blindfolded animals (intact: 14.4 ± 5.8 mm; blind: 15.0 ± 5.3 mm) and approximately halved after antennectomy (sighted: 7.1 ± 3.7 mm; blind: 8.4 ± 4.2 mm). Accordingly, paired t-tests on per-animal means confirmed that contact distances of cohort 1 were significantly shorter after antennectomy (1 st contact: p < 0.001; bottom of edge: p = 0.005) but remained unchanged after additional blindfolding (1 st contact: p = 0.699; bottom of edge: p = 0.734). In cohort 2, contact distances remained the same after blindfolding (1 st contact: p = 0.845; bottom of edge: p = 0.801) but significantly shortened after additional antennectomy (1 st contact: p < 0.001; bottom of edge: p = 0.009). To test whether animals reared their body axis, i.e. increased pitch after the first contact with the obstacle, we assessed the height difference of the head between the instant of first contact (red symbols in Fig. 1B) and the closest point to the obstacle (green symbols in Fig. 1B). Pairwise t-tests revealed significantly less rearing in sighted but antennectomized animals of cohort 1 (top row in Fig. 1B, 3.94 ± 1.59 mm less vertical shift, p = 0.0077) but no difference in blindfolded animals of cohort 2 (left column in Fig. 1B, 0.61 ± 2.21 mm less vertical shift, p = 0.6074). In both cohorts, rearing was significantly reduced after both treatments (lower right corner in Fig. 1B; with 4.11 ± 1.83, p = 0.0109 and 4.22 ± 1.86 mm, p = 0.0105 less vertical shift in cohorts 1 and 2, respectively). We conclude that animals that could see the landmark but did not detect the obstacle with their antennae (cohort 1 after antennectomy) did not respond to the landmark with increased body pitch. Compared to these effects in front of the stair, treatments had no significant effect on clearance above the upper edge (intact: 10.2 ± 3.1 mm; 8.4 ± 3.1 mm after antennectomy, p = 0.271; 11.0 ± 2.9 mm after blindfolding, p = 0.515), suggesting that neither vision or antennal touch were important for the regulation of clearance near the edge. However, vision but not touch affected the overall speed of climbing, as blindfolded animals and animals with both impairments were significantly slower than intact animals (intact: 40.9 ± 4.7 mm/s; blindfolded: 38.1 ± 2.0 mm/s, p = 0.006; combined treatment: 33.9 ± 5.0 mm/s, p = 0.0013), whereas antennectomy alone had no significant effect (36.4 ± 8.8 mm/s, p = 0.336). We conclude that the effect of vision on overall speed is more pronounced than that of antennal touch. From the stick insects’ perspective, the horizontal black bar might have appeared to hover above the ground, as the bar had no connection to the black walkway. To make sure that sighted but antennectomized animals did not disregard visual cues because of a particular feature of the visual pattern, we conducted additional tests with two visual patterns that had vertical black-and-white contrast edges between the walkway and the top of the stair. The effect of antennectomy remained the same, with significantly less rearing and significantly shorter distances to the stair despite intact vision (Fig. S1 and Table S1 in the Suppl. Mat.). This corroborates our interpretation that sighted animals make no use of visual cues about obstacle height during the onset of climbing. At initiation of climbing, landmarks were visually discernible Since our results suggest that sighted stick insects without antennae do not use the visual cues to initiate climbing as early as intact animals do, it was possible that this happened because the animals were unable to see the visual landmark sufficiently early to respond to it. To assess whether this was the case, we conducted high-precision motion capture measurements to estimate the image size of the black bar and its increase during the approach of the insect. We used intact animals for this because we wanted to know the image size at the beginning of the smooth and early rise of the head of intact animals. Fig. 2A shows the 3D reconstruction of a sample trial with four postures per second (every 50 frames). Based on this reconstruction, the location and apparent angular size of the bar within the visual field was estimated for three instants during the approach (Fig. 2B). The first of these images was calculated for the beginning of the trajectory. Here, the image of the black bar subtended approximately 20 x 4 degrees. The second image was calculated for the instant of the first contact with the stair, when the bar subtended approximately 45 x 15 degrees. The third image was calculated for the middle of the climbing process, when the head was just beneath the lower edge of the landmark. Here, the bar subtended apparent size of 100 x 20 degrees. Assuming a spatial resolution of approximately 5 degrees (Jander and Volk-Heinrichs, 1970), this image sequence suggests that the stick insect could have easily resolved the black bar at the time of initial contact. The bar width had reached approximately three times the visual resolution by then. Figures 2 C and 2 D further substantiate this conclusion. They show three single trials (Fig. 2 C) and per-animal means of head trajectories (top), head pitch (middle) and the corresponding apparent bar width in the visual field of the animal (bottom). The head trajectories within one animal (Fig. 2 C), but also in comparison among animals (Fig. 2D) had an overall uniform sigmoid shape, indicating that both the distance at which the animals started to climb and the course of the climb were consistent both among trials and among animals. The same was found to be true for the head pitch of the animals (mid panels of Fig. 2 C, D). Along the straight walkway, head pitch fluctuated mildly about a constant, neutral value (approximately 0 degrees) until the insect started to climb. The head then began to face upwards towards the top edge of the stair. As the animal pulled itself up, its head pitch decreased such that it faced downwards in relation to the body axis. The apparent visual size of the bar changed accordingly (lower panels in Fig. 2 C, D). About 50 mm in front of the stair, the angular width of the bar exceeded 10 degrees. Given the high image contrast and an inter-ommatidial angle of approximately 5 degrees (Jander and Volk-Heinrichs, 1970), we conclude that stick insects would have been able to discern the visual landmark prior to the initiation of climbing. Landmarks not used for climbing reliably induce yaw turns Using a two-choice paradigm, we were able to prove that stick insects do indeed show visually induced responses to a bar-shaped landmark of width 10 degrees (Fig. 3). When given the choice between a horizontal bar of the same size as the one previously ignored and a blank white surface, stick insects walked towards the stair with the black bar in 87.5 % (effect size: Cohen’s d = 6.4) of all trials (Fig. 3, left). The positions of the two alternative objects were swapped pseudo-randomly, and no side preference was observed in any of the animals. When given the choice between horizontal and vertical black landmarks of equal size, animals reliably chose the vertical landmark in 93.5% (effect size: Cohen’s d = 11.5) of all trials (Fig. 3, right). We conclude that stick insects reliably change their movement direction towards the same visual landmark that was not used during climbing. The disregard of the landmark during climbing was not due to the image size or the type of landmark presented, but rather due to the kind of directional movement to be controlled: the same landmark that was ignored during climbing reliably induced yaw-turning. Visually induced yaw turns become overt during the first swing phase past the decision line To investigate when and how visually induced yaw turns were initiated and executed, we analysed the foot trajectories in a paradigm in which they would reliably turn towards a static visual landmark. Given the results shown in Fig. 3, we opted for vertical landmarks. Preliminary behavioural tests with five animals revealed that in a choice between a white and a black pillar, animals turned towards the white pillar in only 0.5 % of cases (Fig. 4 A; only a single trial of a single animal in a total of 200 trials). Using the same motion capture technique as for Fig. 2, body postures were reconstructed for entire walking trials at a temporal resolution of 0.05 s. Fig. 4 B shows a subset of postures for a representative trial of a visually induced turn towards the black pillar. The analysis began as the animal entered a zone 100 mm prior to the (virtual) decision line and ended as soon as either a front foot or an antenna entered a circular environment at 5 mm distance around the pillar (see red ring for exclusion criterion in Fig. 4 C). As the objective of this analysis was to tell when and how stick insects initiated a yaw turn, we focused on two target variables. The first was the yaw angle of the prothorax, which we used to tell the onset of yaw rotation of the body axis (Fig. 5). The second was the left-right asymmetry between corresponding points on the foot trajectories of both front legs, as front legs are known to initiate yaw turns in tethered walking stick insects (Dürr and Ebeling, 2005). To assess asymmetries of front leg movements, the sequence of four step cycles per front leg was taken into account, two before and two after crossing the decision line (Fig. 4 C). Step cycles were separated into swing and stance movements (Fig. 6 A), the trajectories of which were described by per-animal means (Fig. 6 B). Concerning the body axis, the onset of turning was judged from prothorax trajectories (Fig. 5, top panels; left panel shows the mean ± s.d. of the y position of two animals; right panel shows mean y positions of all eight animals). Overall, the narrow part of the walkway constrained the body trajectories to a consistently narrow range near the middle of the straight walkway. In particular, we used the yaw angle of the prothorax, as a function of prothorax position on the walkway (Fig. 5, mid panels). In order to remove potential directional biases, we subtracted the mean yaw angle measured within a reference range of -80 mm to -20 mm before reaching the decision line. Given the fluctuation of the body axis yaw angle during locomotion, we used three times its standard deviation within that same reference range as a threshold for detecting the onset of yaw rotation. The eight colored dots in the mid panel of Fig. 5B indicate the mean distances where this threshold was exceeded. Among the eight animals, this happened between 14 and 37 mm (25 ± 9 mm, N = 8) behind the decision line, indicating that some animals turned earlier than others. At the time of crossing the decision line, the angular landmark width had already exceeded 10 degrees (Fig. 5, bottom), such that animals must have been able to discern the pillar no later than when the decision line was reached. Since the torque that is required for initiating a turn of the body axis must be generated by some sort of a left-right asymmetry in the leg movement, we wanted to know when such an asymmetry occurred for the first time. To this end, we compared the foot trajectories of the two front legs by matching corresponding pairs of swing and stance phases of the left and right front legs (Fig. 4 C). This was done for a set of four subsequent step cycles of interest, resulting in 14 to 30 body-centred foot trajectories per animal, step category and step cycle phase (Fig. 6A). Note that the movement of stance and swing trajectories was counter-directed, with stance phases beginning at the anterior extreme position (AEP) and ending at the posterior extreme position (PEP), and swing phases moving from PEP to AEP. Furthermore, the y-coordinates of leftward turns were flipped (owing to the pseudo-randomised position of the black pillar and corresponding turning direction), such that bluish colors label the outer leg and reddish colors label the inner leg with respect to the turning direction. Finally, average trajectories were first calculated per animal (black lines in Fig. 6 A) and then summarised as mean positions among animals (mean ± s.d. of per-animal means), using four points per movement phase (Fig. 6 B). Accordingly, mean trajectories of swing phases turned out to be strongly curved, whereas curvature was weak during stance. Mean trajectories for subsequent step cycles had the same overall shape and appeared bilaterally symmetrical for step categories 1 and 2 (before crossing the decision line) but with increasing left-right asymmetry for step categories 3 and 4 (behind the decision line). In order to quantify this left-right asymmetry, we calculated the position differences between inner and outer legs (black lines and symbols in Fig. 6 B) and determined the slope Δy/Δx for each set of four position differences. We reasoned that the first step cycle phase to initiate the yaw turn would be the first for which the slope deviated from zero in a statistically significant manner. To account for repeated, matched measurements of slopes we first tested for overall changes in slope, separately for swing and stance trajectories. Mean slopes differed significantly across the four subsequent steps, both for swing and stance phases (Friedman test; N = 8; stance: p = 0.0045; p = 0.001). Significant differences between step categories were then tested post-hoc with Wilcoxon’s test for matched pairs (again on per-animal means, N = 8). Within the entire data set shown in Fig. 6 B, the stance phase of step 4 was the first to differed significantly from all preceding steps (difference to steps 3, 2 and 1 with p = 0.015, 0.007 and 0.007, respectively), whereas step 3 did not differ from step 2. In contrast, the swing phases of step 3 was the first to differ significantly from all preceding steps (i.e., steps 2 and 1, with p = 0.015 and 0.023, respectively). The swing phase of step 4 also differed from that of steps 2 and 1 (p = 0.007 and 0.007, respectively), but not from step 3. We conclude that the left-right asymmetry of front leg movement was detectable first in a directional change of the first swing movement after crossing the decision line (step 3), with the first significant change in stance direction occurring during the immediately subsequent step 4. Since this kind of analysis was designed to detect spatial differences in the sequence of steps, it neglected the temporal delay between steps of the inner and outer leg. However, animals crossed the decision line first with their inner leg in some trials, but with the outer leg first in other trials. Therefore, the pooled data set of Fig. 6 B was split according to which front leg crossed the decision line first. Fig. 7 shows the corresponding mean trajectories for “inner-leg-first trials” (Fig. 7 A) and “outer-leg-first trials” (Fig 7 B). Overall, the statistical results substantiated the analysis of the pooled data set, with both step phases becoming significantly asymmetric in inner-leg-first trials (Friedman test, N = 8, stance: p = 0.0359; swing: p = 0.001), however with only the stance phase doing so in outer-leg-first trials (Friedman test, N = 8, stance: p = 0.0239; swing: p = 0.0703). In inner-leg-first trials (Fig. 7 A) post-hoc tests revealed statistically significant differences for stance phases between steps 4 and 2 only (p= 0.007; n.s. for steps 3 or 1), whereas for swing phases, the first statistical difference occurred between steps 3 and 2 (p = 0.007), followed by a further change between steps 4 and 3 (p= 0.04). Similarly, in outer-leg-first trials (Fig. 7 B) the first significant difference between stance phases occurred between steps 4 and 2 (p = 0.007). Outer-leg-first trials showed no significant differences between any pair of subsequent steps. We conclude that, like in the pooled sample of Fig. 6, first significant spatial asymmetry occurred in the first swing phase after crossing the decision line (step 3), and/or in the second stance phase (step 4). If the outer leg was first to cross the decision line, it took one complete step cycle of that leg (step 3) before its second stance phase changed direction. If the inner leg was first to cross the decision line, it took only one stance phase of that leg before its first swing phase changed direction. We conclude that the delay for initiating a yaw turn is one stance phase and an entire step cycle. Discussion As a climbing insect that dwells in complex-structured environments like thickets of bramble or ivy, stick insects need to be able to change their heading with both vertical (pitch) and horizontal (yaw) turns. Our results show that stick insects reliably turn towards a visual landmark in the yaw direction (Fig. 3 ), whereas they do not exploit the presence of the very same landmark for a timely change of body pitch (Fig. 1 ). Using a paradigm that constrains the onset of visually induced yaw turns (Fig. 4 ) we found that a change in heading occurred approximately at the distance covered during one swing or stance movement (Fig. 5 ). On average, the first left-right asymmetry in stepping occurred during a swing, with a delay of an entire stance phase (Fig. 6 ). Depending on which front leg first stepped beyond the turning constraint, the delay could be an entire step cycle, and the first asymmetry could occur during stance (Fig. 7 ). Visual landmarks do not affect the initiation of climbing When climbing an obstacle, insects must change their body pitch angle. While this change in body pitch has been studied in relation to tactile and/or visual cues in several insect species (e.g., (Pelletier and McLeod, 1994 ; Zurek and Gilbert, 2014 ; Clifton et al., 2020 ) only few studies have related the change in pitch to the leg movements that cause it. Those that did, identified two distinct leg movement strategies: One strategy is to rear their body axis by differential extension of front, middle and hind legs in stance, thus raising the prothorax more than their metathorax. This has been documented for cockroaches (Watson et al., 2002 ) and, indirectly, for stick insects (Schütz and Dürr, 2011 ). In stick insects, the change in pitch was small but sufficiently fast to be detectable between two subsequent antennal contacts with the obstacle (in the range of 100 ms, i.e. faster than a swing movement of a leg). The other strategy is to grasp hold of the obstacle with a front leg and pull the body upwards. This has been found to occur in combination with rearing in cockroaches and stick insects (same studies as before), with stick insects showing tactually induced changes in leg movement within 40 ms, for example during re-targeting of an ongoing swing movement. So far, only few insect species have been shown to adjust foothold or grasping position in response to visual cues: Locusts and grasshoppers adjust their front leg swing movements so as to reach for footholds in climbing paradigms (Niven et al., 2010 ; Niven et al., 2012 ). Praying mantises adjust their front leg strike direction to visual targets (Prete and Cleal, 1996 ). In contrast to this, we found no evidence for stick insects to make use of a high-contrast landmark to adjust body pitch so as to improve climbing performance. The animals tested in experiment 1 behaved in the exact same way as reported for climbing in darkness (Dürr et al., 2003 ): The presence of a high-contrast visual landmark did not rescue the effect of antennectomy, such that sighted animals without antennae showed the same head trajectory as blind animals without antennae (Fig. 1 , Suppl. Fig. S1 ). This could not be explained by the animals not being able to see the landmark, for two reasons: First, the apparent landmark size in the visual field of the animals (Fig. 2 ) was sufficiently large for the animals to discern the black bar against the white surrounding. Second, a same landmark with the same shape and visual image size reliably elicited a yaw turning response (Fig. 3 ). We conclude that seeing the landmark did not improve their climbing efficiency by a timely inclination of the body axis. There are two possible explanations for this: (i) Stick insects could have ignored the elevated visual landmark unless it was accompanied by near-range information from tactile cues, or (ii) they may not have had enough time to respond with a change in body pitch or targeted reaching. Both of these explanations are related to the estimate of distance. In insects, visual distance information can be inferred from motion parallax, e.g. through peering movements in locusts (Collett, 1978 ), from stereopsis, e.g. in striking mantids (Rossel, 1983 ) or from image expansion, e.g. during landing or escape (Wagner, 1982 ; Fotowat and Gabbiani, 2007 ). To date, none of these mechanisms have been described for stick insects. Moreover, even in insects that can estimate distance visually, a relation between visual size and distance needs not exist (Rossel, 1991 ). In other words, the perceived angular size of an elevated landmark alone may be too unreliable to adjust body pitch in preparation of climbing. Regarding reaction time, visually triggered movements in stick insects must be expected to be considerably slower than tactually elicited movements. Touch-related retargeting of a front leg may be initiated only 40 ms after the first contact (Schütz and Dürr, 2011 ), and descending interneurons convey antennal mechanosensory information to the thorax with an 11 ms delay only (Ache and Dürr, 2013 ). In comparison, visually induced responses in stick insects are slow. For example, changes in front leg movement during turns elicited by large-field visual motion have time constants of 1.5 or 1.7 s in inner and outer front legs, respectively (Dürr and Ebeling, 2005 ). Even when considering the long time constant of muscle activation dynamics in insect muscle (Harischandra et al., 2019 ) the onset of the related muscle activity would occur with a delay of at least 1 s. Similarly, we could reliably detect yaw turns towards static landmarks only 17 to 37 mm behind the decision line (Fig. 5 ). This is similar to the 24 to 40 mm distance expected to be covered during one stance phase, assuming an average speed of 40 to 50 mm/s and stance durations between 600 to 800 ms. Further considering that the significant change in body pitch found in Fig. 1 occurred only ca. 15 mm beyond the first contact in blindfolded animals, we conclude that even if the animals could have estimated distance from the expanding image of the visual landmark, they couldn’t have responded as early as the blindfolded animals and not early enough to prevent bumping into the obstacle (as many of them did). Visually induced yaw turns tend to begin with a change in swing direction Theoretically, the torque driving yaw rotation must be generated by some left-right asymmetry of leg movement action. This asymmetry can be generated by means of three, potentially independent movement parameters: step frequency, step length and step direction. In stick insects, visual-motion-induced yaw turns involve asymmetric changes in all three of these parameters, both with and without mechanical coupling among the legs in stance (Dürr and Ebeling, 2005 ; Gruhn et al., 2009 ). Associated gait changes are governed by changes in pair-wise coupling of neighboring legs, particularly in front legs, including left-right asymmetry of coupling efficacy (Dürr, 2005 ). Recently, stick insects were found to turn by means of a combination of changed step direction and the likelihood of a distinct turn-related step class (Meschenmoser and Dürr, 2026 ). To complement this earlier work, the present study focused on step direction asymmetries of the front leg movements, with separate treatment of swing and stance phases. In steady state locomotion, for example when walking a curve of fixed radius with constant speed, directions of stance or swing movements will be inverted by 180°, while step lengths will be equal. At the onset of turning, however, directions of stance and swing will transiently deviate from this 180° inversion. Although it is common to study the mentioned left-right asymmetry during stance (e.g., (Yang et al., 2024 ; Chockley et al., 2022 ), and tethered walking stick insects were found to change stance parameters slightly faster than swing parameters (Dürr and Ebeling, 2005 ), it is unclear what initiates turning in free walking animals. Transient asymmetries could arise through an obliquely directed swing movement that was to be followed by a stance movement to pull the body towards the new foothold. Our results show that the earliest step phase with a statistically significant left-right asymmetry is the first swing phase past the “decision line”, i.e. past the physical widening of the walkway (Fig. 6 ). After splitting the trials according to the body side that stepped across the decision line first, this pattern became more pronounced in inner-leg-first trials (Fig. 7 A) and weakened in outer-leg-first trials (Fig. 7 B). Assuming that animals decided to turn towards the dark object soon after stepping across the widening, this means that a significant change in leg movement asymmetry required at least the duration of an entire stance phase to take effect (step 3 in Figs. 6 and 7 ). While this timing is very similar to that observed by Dürr and Ebeling ( 2005 ), their earliest change was a lateral shift in lift-off position and the preceding stance direction, whereas in our pooled data it was a change in swing direction. This difference could be related to (i) a difference in time course analysis, (ii) the biomechanical differences associated to the experimental setup, or (iii) the visual cue used to initiate turning. Regarding methodological differences, Dürr and Ebeling ( 2005 ) inferred the time constant of an exponential response variable from pooled data, comprising all step cycles of a trial and varying numbers of steps per animal. In contrast, here, we use matched samples of an ordered sequence of step cycle phases, with means per animal. Furthermore the free walking animals of the present study needed to generate natural yaw torques only (as they rotated their own body, only), whereas the animals of Dürr and Ebeling ( 2005 ) needed to rotate a hollow sphere whose moment of inertia was more than 100-fold that of their own body. Finally, Dürr and Ebeling ( 2005 ) triggered curve walking by means of a large-field visual-motion cue (provided by a rotating stripe drum) as opposed to the stationary landmark used here. At least for flying insects, visually induced turning responses have been dissociated into motion-related turns as opposed to position-related cues (Buchner, 1984 ). Accordingly, the position of our stationary landmark may have triggered turning through a change in swing direction, whereas a large-field motion stimulus could trigger turning by a change in stance. Whether this differentiation of position and motion-induced turning holds for walking insects may be tested by future studies. We conclude that visually induced yaw turns towards static visual landmarks may be initiated by a change in front leg swing direction. This is reminiscent of the fact that tactually induced turning is accompanied with an increased likelihood of reaching movements (Berendes and Dürr, 2022 ). Other than horizontal turning, climbing must involve a change in foothold height, which can only occur during swing during swing. The initiation of both pitch and yaw turns by means of a directed swing movement could be a general strategy for initiating omnidirectional changes in heading in terrestrial locomotion. Declarations Author Contribution R.B. conducted experiments, analysed data, prepared figures 2-7, wrote and reviewed the manuscript; N.S. conducted experiments, analysed data, prepared figure 1 and Suppl. Fig. 1, and reviewed the manuscript; V.D. conceived the experiments, supervised R.B. and N.S., analysed data, wrote and reviewed the manuscript. Acknowledgement We thank Florian P. Schmidt for technical assistance and Merit Meschenmoser for critically reading a previous version of the manuscript. References Ache, J. M. and Dürr, V . (2013). Encoding of near-range spatial information by descending interneurons in the stick insect antennal mechanosensory pathway. J. Neurophysiol. 110 , 2099–2112. Berendes, V. and Dürr, V . (2022). Active tactile exploration and tactually induced turning in tethered walking stick insects. J. Exp. Biol. 225 , jeb243190. Buchner, E . (1984). Behavioral analysis of spatial vision in insects. In Photoreception and Vision in Invertebrates (ed. M. Ali), pp. 561–621. New York, London: Plenum Press. Camhi, J. M. and Johnson, E. N . (1999). High-frequency steering maneuvers mediated by tactile cues: Antennal wall-following in the cockroach. J. Exp. Biol. 202 , 631–643. Chockley, A. S., Dinges, G. F., Di Cristina, G., Ratican, S., Bockemühl, T. and Büschges, A . (2022). Subsets of leg proprioceptors influence leg kinematics but not interleg coordination in Drosophila melanogaster walking. J. Exp. Biol. https://doi.org/10.1242/jeb.244245. Cleal, K. S. and Prete, F. R . (1996). The predatory strike of free ranging praying mantises, Sphodromantis lineola (Burmeister). 2. Strikes in the horizontal plane. Brain Behavior and Evolution 48 , 191–204. Clifton, G., Stark, A. Y., Li, C. and Gravish, N . (2023). The bumpy road ahead: the role of substrate roughness on animal walking and a proposed comparative metric. J. Exp. Biol. https://doi.org/10.1242/jeb.245261. Clifton, G. T., Holway, D. and Gravish, N . (2020). Vision does not impact walking performance in Argentine ants. J. Exp. Biol. https://doi.org/10.1242/jeb.228460. Collett, T. S . (1978). Peering - Locust behavior pattern for obtaining motion parallax information. J. Exp. Biol. 76 , 237–241. Cowan, N. J., Lee, J. and Full, R. J . (2006). Task-level control of rapid wall following in the American cockroach. J. Exp. Biol. 209 , 1617–1629. Cruse, H., Dean, J., Kindermann, T., Schmitz, J. and Schumm, M . (1998). Simulation of complex movements using artificial neural networks. Zeitschrift fur Naturforschung C-A Journal of Biosciences 53 , 628–638. Dudley, R . (2002). Mechanisms and implications of animal flight maneuverability. Integr Comp Biol 42 , 135–140. Dürr, V., König, Y. and Kittmann, R . (2001). The antennal motor system of the stick insect Carausius morosus: anatomy and antennal movement pattern during walking. J. Comp. Physiol. A 187 , 131–144. Dürr, V., Krause, A. F., Schmitz, J. and Cruse, H . (2003). Neuroethological concepts and their transfer to walking machines. Int. J. Robot. Res. 22 , 151–167. Dürr, V . (2005). Context-dependent changes in strength and efficacy of leg coordination mechanisms. J. Exp. Biol. 208 , 2253–2267. Dürr, V. and Ebeling, W . (2005). The behavioural transition from straight to curve walking: kinetics of leg movement parameters and the initiation of turning. J. Exp. Biol. 208 , 2237–2252. Dürr, V. and Schilling, M . (2018). Transfer of spatial contact information among limbs and the notion of peripersonal space in insects. Front. Comput. Neurosci. 12 , 101. Dürr, V., Arena, P., Cruse, H., Dallmann, C. J., Drimus, A., Hoinville, T., Krause, T., Matefi-Tempfli, S., Paskarbeit, J., Patane, L. et al . (2019). Integrative biomimetics of autonomous hexapedal locomotion. Frontiers in Neurorobotics . https://doi.org/10.3389/fnbot.2019.00088. Dürr, V., Berendes, V. and Strube-Bloss, M. F . (2022). Sensorimotor ecology of the insect antenna: Active sampling by a multimodal sensory organ. Adv. Insect Physiol. 63 , 1–105. Dürr, V. and Mesanovic, A . (2023). Behavioural function and development of body-to-limb proportions and active movement ranges in three stick insect species. J. Comp. Physiol. A , 265–284. Fotowat, H. and Gabbiani, F . (2007). Relationship between the Phases of Sensory and Motor Activity during a Looming-Evoked Multistage Escape Behavior. J. Neurosci. 27 , 10047–10059. Gruhn, M., Zehl, L. and Büschges, A . (2009). Straight walking and turning on a slippery surface. J. Exp. Biol. 212 , 194–209. Harischandra, N., Krause, A. F. and Dürr, V . (2015). Stable phase-shift despite quasi-rhythmic movements: a CPG-driven dynamic model of active tactile exploration in an insect. Front. Comput. Neurosci. 9 , 107. Harischandra, N., Clare, A. J., Zakotnik, J., Blackburn, L. M. L., Matheson, T. and Dürr, V . (2019). Evaluation of linear and non-linear activation dynamics models for insect muscle. Plos Computational Biology 15 , e1007437. Harley, C. M., English, B. A. and Ritzmann, R. E . (2009). Characterization of obstacle negotiation behaviors in the cockroach, Blaberus discoidalis . J. Exp. Biol. 212 , 1463–1476. Jander, R. and Volk-Heinrichs, I . (1970). Das Strauch-spezifische Perceptor-System der Stabheuschrecke ( Carausius morosus ). Z. vergl. Physiol. 70 , 425–447. Kalmus, H . (1937). Photohorotaxis, eine neue Reaktionsart, gefunden an den Eilarven von Dixippus morosus . Z. vergl. Physiol. 24 , 644–655. Krause, A. F. and Dürr, V . (2012). Active tactile sampling by an insect in a step-climbing paradigm. Front. Behav. Neurosci. 6 , 1–17. Meschenmoser, M. and Dürr, V . (2025). Contrast and luminance dependence of target choice and visual orientation in walking stick insects. Scientific Reports 15 , 12226. Meschenmoser, M. and Dürr, V . (2026). Walking in circles: freely walking stick insects turn by a combination of step-parameter scaling and altered likelihood of particular step types. bioRxiv . Niven, J. E., Buckingham, C. J., Lumley, S., Cuttle, M. F. and Laughlin, S. B . (2010). Visual targeting of forelimbs in ladder-walking locusts. Curr. Biol. 20 , 86–91. Niven, J. E., Ott, S. R. and Rogers, S. M . (2012). Visually targeted reaching in horse-head grasshoppers. Proc. R. Soc. Lond. B 279 , 3697–3705. Okada, J. and Toh, Y . (2000). The role of antennal hair plates in object-guided tactile orientation of the cockroach ( Periplaneta americana ). J. Comp. Physiol. A 186 , 849–857. Okada, J. and Toh, Y . (2006). Active tactile sensing for localization of objects by the cockroach antenna. J. Comp. Physiol. A 192 , 715–726. Pelletier, Y. and McLeod, C. D . (1994). Obstacle perception by insect antennae during terrestrial locomotion. Physiol. Entomol. 19 , 360–362. Pick, S. and Strauss, R . (2005). Goal-driven behavioral adaptations in gap-climbing Drosophila . Curr. Biol. 15 , 1473–1478. Prete, F. R. and Cleal, D. S . (1996). The predatory strike of free ranging praying mantises, Sphodromantis lineola (Burmeister).1. Strikes in the mid-sagittal plane. Brain Behavior and Evolution 48 , 173–190. Rossel, S . (1983). Binocular stereopsis in an insect. Nature 302 , 821–822. Rossel, S . (1991). Spatial vision in the praying mantis: is distance implicated in size detection ? J. Comp. Physiol. 169 , 101–108. Schütz, C. and Dürr, V . (2011). Active tactile exploration for adaptive locomotion in the stick insect. Phil. Trans. R. Soc. Lond. B 366 , 2996–3005. Taylor, G. K . (2001). Mechanics and aerodynamics of insect flight control. Biol. Rev. 76 , 449–471. Theunissen, L. M. and Dürr, V . (2013). Insects use two distinct classes of steps during unrestrained locomotion. PLOS one 8 , e85321. Wagner, H . (1982). Flow-field variables trigger landing in flies. Nature 297 , 147–148. Watson, J. T., Ritzmann, R. E., Zill, S. N. and Pollack, A. J . (2002). Control of obstacle climbing in the cockroach, Blaberus discoidalis . I. Kinematics. J. Comp. Physiol. A 188 , 39–53. Webb, P. W . (2004). Maneuverability - general issues. IEEE Journal of Oceanic Engineering 29 , 547–555. Yang, H. H., Brezovec, B. E., Serratosa Capdevila, L., Vanderbeck, Q. X., Adachi, A., Mann, R. S. and Wilson, R. I . (2024). Fine-grained descending control of steering in walking Drosophila . Cell 187 , 6290-6308.e27. Zeng, Y., Chang, S. W., Williams, J. Y., Nguyen, L. Y.-N., Tang, J., Naing, G. and Dudley, R . (2020). Canopy parkour: movement ecology of post-hatch dispersal in a gliding nymphal stick insect ( Extatosoma tiaratum ). J. Exp. Biol. 223 , jeb226266. Zurek, D. B. and Gilbert, C . (2014). Static antennae act as locomotory guides that compensate for visual motion blur in a diurnal, keen-eyed predator. Proc. R. Soc. Lond. B . https://doi.org/10.1098/rspb.2013.3072. Additional Declarations No competing interests reported. Supplementary Files BiggeEtAl2026SupplMaterial.pdf Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 08 May, 2026 Reviews received at journal 08 May, 2026 Reviews received at journal 11 Apr, 2026 Reviewers agreed at journal 07 Apr, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers invited by journal 31 Mar, 2026 Editor assigned by journal 30 Mar, 2026 Submission checks completed at journal 30 Mar, 2026 First submitted to journal 27 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9243999","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":615939147,"identity":"04809dbf-7646-4e46-966e-61f6cd046bce","order_by":0,"name":"Ronja Bigge","email":"","orcid":"","institution":"Bielefeld University","correspondingAuthor":false,"prefix":"","firstName":"Ronja","middleName":"","lastName":"Bigge","suffix":""},{"id":615939151,"identity":"c06523e3-dfa0-454d-a5bc-d29186a1bf0e","order_by":1,"name":"Nicoletta Schwermann","email":"","orcid":"","institution":"Bielefeld University","correspondingAuthor":false,"prefix":"","firstName":"Nicoletta","middleName":"","lastName":"Schwermann","suffix":""},{"id":615939153,"identity":"30c0612e-d780-44bb-82a0-679e7be5d4ba","order_by":2,"name":"Volker Dürr","email":"data:image/png;base64,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","orcid":"","institution":"Bielefeld University","correspondingAuthor":true,"prefix":"","firstName":"Volker","middleName":"","lastName":"Dürr","suffix":""}],"badges":[],"createdAt":"2026-03-27 11:10:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9243999/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9243999/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106403912,"identity":"037bd4c4-a46b-4411-898e-187b7e6858b4","added_by":"auto","created_at":"2026-04-08 09:15:12","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":426795,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStick insects use touch but not vision to initiate climbing.\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eA\u003c/strong\u003e Stick insects walked towards a stair that was marked by a horizontal bar just below the edge (width 8 mm). Top: View of the stair as seen during approach; bottom: schematic side view. \u003cstrong\u003eB\u003c/strong\u003e Mean head trajectories per animal were reduced to four head positions, as depicted by symbols of different color (means of 9 to 10 trials per animal). Red: first contact with stair. Green: Closest point to bottom edge. Dark blue: Height above the upper edge. Turquoise: Height 10 mm behind the edge of the stair. The four conditions were intact (top left, N = 10, both cohorts), sighted but with cut antennae (top right, N = 5, cohort 2), blind but with intact antennae (lower left, N = 5, cohort 1) and blind with cut antennae (lower right, N = 10). Black crosses show means and standard deviations of the pooled data. Black dashed lines connect the grand mean positions as an idealized head trajectory. Grey dotted lines copy mean positions of intact animals, for comparison.\u003c/p\u003e","description":"","filename":"Fig1NoVisionForClimbing.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/654f04b17ac68d841be823ec.jpg"},{"id":106300576,"identity":"ca6432c9-1833-44c7-a971-630cdcf63f4d","added_by":"auto","created_at":"2026-04-07 09:14:13","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2766783,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe visual image of the bar was visible before the onset of climbing.\u003c/strong\u003e \u003cstrong\u003eA \u003c/strong\u003e3D reconstruction of a sample trial, captured every 50 frames (250 ms). \u003cstrong\u003eB \u003c/strong\u003eReconstruction of the bar size at three different distances to the stair. Distances are given for the head position in relation to the center of the bar. \u003cstrong\u003eC/D \u003c/strong\u003eHead trajectories, head pitch and angular landmark size of three single trials of one animal (\u003cstrong\u003eC\u003c/strong\u003e) and average trajectories of three animals (\u003cstrong\u003eD\u003c/strong\u003e). Dashed line indicates a threshold of 10° at which stick insects should be able to discern visually, given the high contrast of the bar and literature values of their visual resolution. Arrows in \u003cstrong\u003eD\u003c/strong\u003e refer to Fig. 1B and indicate the positions of the 1\u003csup\u003est\u003c/sup\u003e contact of intact animals (left) and of animals with cut antennae (right).\u003c/p\u003e","description":"","filename":"Fig2HeadPitchEtccolorFix.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/ced298fcc222a440ff412e90.jpg"},{"id":106300568,"identity":"d43e5b4b-7b91-4701-b1bc-120c60c672f4","added_by":"auto","created_at":"2026-04-07 09:14:13","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2341458,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLandmarks that were not used during climbing reliably induced yaw turns.\u003c/strong\u003e When confronted with a choice of two stairs, stick insects reliable turned towards one of two alternative visual patterns. Two pattern combinations were tested. Left: When given the choice between a horizontal bar and no bar, animals reliably went for the horizontal bar. Right: when given the choice between equal-sized horizontal and vertical bars, they reliably went for the vertical bar. N = 5 animals with n = 2×20 trials per animal and pattern combination. Error bars show 95% confidence intervals for ratios for pooled data (n = 200).\u003c/p\u003e","description":"","filename":"Fig3PatternPreference.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/565dfcea934f0e956eb4a174.jpg"},{"id":106300563,"identity":"210bf40e-c8b5-48f5-8dfe-becd2c313700","added_by":"auto","created_at":"2026-04-07 09:14:11","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1604354,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eReliable visual induction of yaw turns for high-precision analysis. A \u003c/strong\u003eIn a two-choice paradigm with black and white pillars (top), stick insects reliably turn towards the black pillar (bottom; N = 5 animals with n = 40 trials each). Error bars indicate 95% confidence intervals for the entire sample.\u003cstrong\u003e B \u003c/strong\u003e3D reconstruction of a single decision trial, showing instances every 50 frames. \u003cstrong\u003eC \u003c/strong\u003eStep cycle categories for kinematic analysis of foot trajectories. Black line shows prothorax trajectory, blue line the outer (left) foot, and red line the inner (right) foot trajectory during turning. The last two step cycles before crossing the ‘decision line’ (dashed line) and the subsequent two step cycles were taken into account for the analysis. Step cycles were defined to begin with a stance episode (coloured dots). Red circle indicates the exclusion criterion, i.e., a 5 mm environment around the pillar. The latter was used to exclude the effect of tactile contacts. Once either a front foot or an antennal tip entered this environment, the last swing episode was excluded from the analysis.\u003c/p\u003e","description":"","filename":"Fig4ParadigmBraodenedWalkway.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/422c35aa606c4b1302cef450.jpg"},{"id":106300506,"identity":"4f3345cc-b161-418e-9d33-dac6e1c2221a","added_by":"auto","created_at":"2026-04-07 09:14:01","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2993037,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eProthorax trajectory, body orientation and angular landmark size during turning\u003c/strong\u003e. \u003cstrong\u003eA\u003c/strong\u003e Mean and standard deviation for two individuals; n\u003csub\u003e1\u003c/sub\u003e = 26, n\u003csub\u003e2\u003c/sub\u003e = 30. \u003cstrong\u003eB\u003c/strong\u003e Per-animal means; N = 8, n = {14, 17, 20, 29, 26, 19, 30, 30} for ascending animal number. Top panels: Prothorax trajectory in top view, with (0, 0) marking the middle of the decision line. Mid panels: Prothorax orientation as a function of its x-position on the walkway. As an estimate of turning onset, circles mark where yaw angles exceeded 3 times the standard deviation of the angles before reaching the decision line. This reference range was calculated in the shaded area. Bottom panels: Angular width of the landmark (pillar) in the visual fields of the animals at corresponding prothorax position.\u003c/p\u003e","description":"","filename":"Fig5YawTurns.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/ec6dc47f21a5e90ff9a2d446.jpg"},{"id":106300510,"identity":"1374397d-7a63-40e0-9abd-319800561595","added_by":"auto","created_at":"2026-04-07 09:14:03","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3552657,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStance and swing trajectories for four subsequent step categories.\u003c/strong\u003e Color code and steps as indicated in Fig. 4C. Foot trajectories are shown in body-fixed coordinate system, with its origin at the prothorax-mesothorax joint. Mean trajectories (black) were calculated for four points of each trajectory: for both extreme positions (AEP and PEP) and after one third and two thirds of the trajectory. \u003cstrong\u003eA\u003c/strong\u003e Stance and swing trajectories of several trials (n = 29) from one example animal. \u003cstrong\u003eB \u003c/strong\u003eAverage stance and swing trajectories across all animals (N = 8). Whereas black solid lines show the mean difference between inner and outer leg for x- and y-coordinates (error bars show s.d.). Asterisks in subpanels 6 and 7 label first significantly slanted line of mean differences for swing and stance phase, respectively. Corresponding black lines and asterisks above subpanels indicate significant pairwise tests.\u003c/p\u003e","description":"","filename":"Fig6SwingStancePooledcolorFix.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/a59fd9af90f7ee223050cdd2.jpg"},{"id":106300507,"identity":"50ed4b72-61f5-413e-8b30-94ffb86f74f8","added_by":"auto","created_at":"2026-04-07 09:14:01","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2869427,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAverage stance and swing trajectories for split samples, depending on which foot crossed the decision line first.\u003c/strong\u003e \u003cstrong\u003eA \u003c/strong\u003eInner-leg-first sample. \u003cstrong\u003eB \u003c/strong\u003eOuter-leg-first sample. Same graph details as in Fig. 6 B.\u003c/p\u003e","description":"","filename":"Fig7SwingStanceLeftRightcolorFix.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/541ecf6b1b4d981bf8bc1881.jpg"},{"id":106406020,"identity":"e42a82f1-7d30-4ae2-b136-5fcecc37afb3","added_by":"auto","created_at":"2026-04-08 09:29:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5378049,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/d3af7e46-0f3d-4780-bd5e-eda800900165.pdf"},{"id":106300566,"identity":"c5806d23-d8ed-44cb-bf69-3e19290c9d80","added_by":"auto","created_at":"2026-04-07 09:14:11","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":758648,"visible":true,"origin":"","legend":"","description":"","filename":"BiggeEtAl2026SupplMaterial.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9243999/v1/df675fe533f8399c9d079cca.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Changing 3D heading direction in terrestrial locomotion: Effectiveness of visual landmarks and the onset of turning in stick insects","fulltext":[{"header":"Introduction","content":"\u003cp\u003eClimbing about in any three-dimensional spatial structure, such as the branches of a tree or the leaves of a shrub, requires the ability of an animal to adjust its heading in both the perceived vertical and perceived horizontal direction. Other than in aquatic or aerial maneuverability, where active pitch and yaw rotation of the body axis are not linked to discontinuities of the surrounding medium (swimming: Webb, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; flying: Dudley, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Taylor, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2001\u003c/span\u003e), any change of heading during walking or climbing is bound to surface structure and, therefore, must be affected by surface discontinuities such as edges, walls or margins. Experimentally, surface discontinuities are commonly used to enforce pitch rotation by disturbance of horizontal walking(Pelletier and McLeod, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e1994\u003c/span\u003e; Harley et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Krause and D\u0026uuml;rr, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Clifton et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) or running (Zurek and Gilbert, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), whereas they are commonly avoided when enforcing yaw rotation on a horizontal surface through visual cues. This raises the question whether pitch and yaw rotations in walking animals rely on the same visuomotor mechanisms. Here we use the Indian stick insect \u003cem\u003eCarausius morosus\u003c/em\u003e (Sin\u0026eacute;ty, 1901), a climbing herbivorous insect that dwells in the foliage of the plants which it feeds on, to investigate two general aspects of omnidirectional heading in surface-bound, terrestrial locomotion. The first aspect concerns the equal effectiveness of visual landmark cues for inducing either vertical (pitch) or horizontal (yaw) rotations. The second aspect concerns the type of leg movement that initiates turning on a horizontal surface, as opposed to climbing onto a vertical surface. Given the evidence that stick insects make use of tactile cues to initiate pitch rotation for climbing(Sch\u0026uuml;tz and D\u0026uuml;rr, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) and yaw rotation in turning (Berendes and D\u0026uuml;rr, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), we will relate evidence on tactually versus visually induced changes in 3D heading.\u003c/p\u003e \u003cp\u003eDuring locomotion, stick insects actively explore the space ahead with persistent, quasi-rhythmic movement of both antennae (D\u0026uuml;rr et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Harischandra et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; D\u0026uuml;rr and Mesanovic, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). As in many other insect species (D\u0026uuml;rr et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), active antennal movement in stick insects improves detection and negotiation of obstacles (D\u0026uuml;rr et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Krause and D\u0026uuml;rr, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).. Tactile localisation of a vertical surface typically induces an upward climbing response that is initiated by a directed reach-to-grasp movement (Sch\u0026uuml;tz and D\u0026uuml;rr, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) within the spatial overlap region of antennal and front leg movement ranges (D\u0026uuml;rr and Schilling, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Such directed reach-to-grasp movements of a front leg may be considered as variations of targeted swing movements during walking on horizontal surfaces (D\u0026uuml;rr et al., 2018). In horizontal walking and running cockroaches, tactile cues may induce strong and persistent yaw rotation towards the contacted location (Okada and Toh, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2006\u003c/span\u003e) or rapid steering in thigmotaxis paradigms (Camhi and Johnson, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Cowan et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). In comparison, tactually induced turning in stick insects is more moderate and often accompanied by attempts to grasp the touched object with a front leg (Berendes and D\u0026uuml;rr, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eVisually induced yaw turns appear to be more pronounced in stick insects than tactually induced yaw turns, despite their nocturnal lifestyle. Large-field visual motion (D\u0026uuml;rr and Ebeling, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Gruhn et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) as well as static landmarks (Kalmus, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1937\u003c/span\u003e; Jander and Volk-Heinrichs, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1970\u003c/span\u003e; Zeng et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Meschenmoser and D\u0026uuml;rr, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) have been shown to elicit strong and reliable yaw turns (Meschenmoser and D\u0026uuml;rr, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2026\u003c/span\u003e). Whether visual cues about obstacle height may also initiate climbing by pitch rotation, is unknown. Though D\u0026uuml;rr et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) showed that antennal tactile cues are sufficient to mask any potential effect of blindfolding in a stair-climbing paradigm, the lack of salient visual features in their experimental setup did not rule out the potential use of visual cues for directed reaching and/or the initiation of climbing. Meanwhile, flies (Pick and Strauss, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) and locusts (Niven et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) have been shown to use visual cues for targeted, upward-directed swing movements during climbing. Accordingly, the first objective of this study was to test whether stick insects make use of visual cues about obstacle height in the initiation of climbing, too. If so, this would indicate that vision, like touch, can provide directional cues for 3D heading. Perhaps it could even provide spatial coordinates required for spatially directed movement of the foot, as it is known for prey capture in mantids (Cleal and Prete, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e1996\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe second objective was to test whether visually induced yaw rotation begins a directional change of a swing movement, stance movement or either of those. Pitch rotation during the initiation of climbing is initiated by an inclination of the body axis, i.e., rearing (cockroach: Watson et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Harley et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; stick insect: Sch\u0026uuml;tz and D\u0026uuml;rr, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Rearing can account for small changes in body pitch angle through a front-to-rear difference in extension of the legs in stance (Watson et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), but climbing onto an obstacle or wall requires a change in foothold height that can be achieved by an elevated swing movement, only. In contrast, a change in body yaw angle requires asymmetric action of left and right legs that can be achieved during either swing or stance (Gruhn et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), though the transient change in stance direction of front legs is fastest and most consistent (D\u0026uuml;rr and Ebeling, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), complemented by a left-right asymmetry of inter-leg coupling (D\u0026uuml;rr, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Recent optogenetic manipulation experiments on \u003cem\u003eDrosophila\u003c/em\u003e induced a left-right asymmetry of stance trajectory length, highlighting a potential mechanism underlying a left-right asymmetry of stance movement (Chockley et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Similarly, left-right asymmetries of stance movements are also commonly implemented in insect locomotion models, both in software and hardware (Cruse et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e1998\u003c/span\u003e; D\u0026uuml;rr et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHere we use static visual landmarks that reliably induce yaw rotation in order to test whether the onset of turning is related to a preceding change in swing direction or rather to an ongoing change in stance direction. As a prerequisite for telling apart the two alternatives, we used a setup that hindered animals to change their heading until they reached a widening of the walkway. This allowed us to define a virtual \u0026ldquo;decision line\u0026rdquo; and to relate swing and stance directions on the narrow walkway, i.e. \u003cem\u003ebefore\u003c/em\u003e the onset of rotation, to those \u003cem\u003eduring\u003c/em\u003e rotation on the widened walkway. Moreover, it allowed us to tell the number of steps necessary to achieve a significant turn angle. If turning began during swing, directional information about the landmark would be transferred into a change in touch-down position which would lead to a change in pull direction of the subsequent stance movement. Essentially, this would be the same kind of directional information transfer as used for the initiation of climbing by targeted reaching (D\u0026uuml;rr and Schilling, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), except that the new foothold would be on the same plane as the preceding one. Alternatively, if turning began during stance, directional information about the visual landmark would induce a left-right asymmetry of stance length, pull direction, or both.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch3\u003eAnimals\u003c/h3\u003e\n\u003cp\u003eAll experiments were performed on adult female stick insects of the species \u003cem\u003eCarausius morosus\u003c/em\u003e (de Sin\u0026eacute;ty, 1901) which were bred in a laboratory culture at Bielefeld University. Each of the tested animals had intact legs with equally long limbs and feelers. During the experiments, light conditions were kept constant by fully covering the windows and using an artificial light source instead. Experiments were conducted at room temperature.\u003c/p\u003e\n\u003cp\u003eDifferent groups of animals were used for each one of the four experiments described below. Except for \u003cem\u003eexperiment 3\u003c/em\u003e, animals were labelled with a set of retro-reflective, spherical markers with a diameter of 1.5 mm. Markers were glued to the cuticle of the animals by use of transparent nail polish. During the gluing process care was taken not to affect the mobility of the joints in any way.\u003c/p\u003e\n\u003ch3\u003eStair-climbing paradigm\u003c/h3\u003e\n\u003cp\u003eFreely walking stick insects were tested in a climbing paradigm, in which they walked along a straight, flat, black walkway which carried a cuboid at its end that served as a stair. Once the animals reached the stair, they readily climbed it. To maximise the visual contrast of the upper edge of the stair, the front side of the stair was painted white with a black, horizontal bar beneath its upper edge. Both, the walkway and the stair were 40 mm wide, which is slightly broader than the average distance between left and right foot contacts during planar walking. A white cardboard shield was placed behind the stair to exclude any other visual landmarks from the paradigm and to increase visual contrast of the black edge against the background.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 1:\u003c/em\u003e The cuboid stair was either 24 mm or 48 mm high, though only data from the 24 mm stair will be presented here. The size of the black bar beneath the edge was 40\u0026times;8\u0026nbsp;mm. The setup was filmed from the side, using an infrared-sensitive digital video camera (Basler A602f, Ahrensburg, Germany; equipped with a 70 mm zoom lens) operating at 50 frames per second, fps. Camera exposure was synchronised with a custom-built infrared flash light.\u003c/p\u003e\n\u003cp\u003eEach animal was labelled with four retro-reflective markers placed onto the posterior mesothorax, mid prothorax, between the eyes on the head, and at the distal tibia of the right front leg.\u0026nbsp;Movement analysis was done using a custom-written\u0026nbsp;Matlab program (The Mathworks, Natick/MA, USA) to track the head marker position over time. Head trajectories relative to the setup were recorded at a spatial resolution of 0.18 mm/pixel. Each trial was summarised by four head positions: (i) the position at the time of the 1\u003csup\u003est\u003c/sup\u003e tactile contact with the stair (either with a front leg or an antenna), (ii) the closest point to the corner at the bottom of the edge, (iii) the point directly above the edge, and (iv) the head position 10 mm behind the upper edge of the stair. The mean speed of the animal was calculated for the head trajectory between the 1\u003csup\u003est\u003c/sup\u003e and 4\u003csup\u003eth\u003c/sup\u003e of these positions.\u003c/p\u003e\n\u003cp\u003eTen animals were divided into two cohorts of five. After at least ten trials with the intact animal, cohort 1 was first tested ten times after bilateral antennectomy at the flagellum base and another ten times after additional blindfolding, i.e. after covering both eyes with black, water-soluble paint. Cohort 2 was first tested intact, then blindfolded and finally antennectomized. The four head positions listed above were measured for each trial. Horizontal distances of positions (i) and (ii), clearances above the stair for positions (iii) and (iv), and mean speed were averaged per animal and tested with paired t-tests on per-animal means. To test whether the observed effects of antennectomy were specific to the visual pattern used, we conducted further trials with an additional cohort 3 (N = 5), with 4\u0026times;8 trials per animal. As visual patterns on the front surface of the stair, we either replaced the horizontal black bar by a vertical black bar of the same width, or added a checkerboard pattern below the horizontal bar (see Supplementary Material).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 2:\u003c/em\u003e For more accurate, 3-dimensional measurement of head trajectories, head pitch, and the apparent visual image size of the black bar beneath the edge, a similar setup as in \u003cem\u003eexperiment 1\u0026nbsp;\u003c/em\u003ewas surrounded with a multi-camera motion-capture system (Vicon MX10, equipped with eight T10 cameras; Vicon, Oxford, UK), operating at 200 fps with a spatial resolution of approximately 0.1 mm. The stair was 36 mm high and carried a 40\u0026times;9\u0026nbsp;mm black bar beneath its upper edge. Each animal was labelled with 10 retro-reflective markers: A triangle of three markers on the dorsal prothorax was used to define the body-fixed coordinate system; one marker on the head was used to estimate head pitch; 2\u0026times;2\u0026nbsp;markers were placed on the distal femur and tibia of either front leg, and two further markers were fixed to the proximal third of each antenna.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA total of five intact animals was tested for 20 trials each. An additional digital video camera (Basler A602f2, equipped with a Pentax H6Z810 zoom lens) recorded a side view of the experiment at 50 fps. Its videos were used for visual inspection and validation of the motion capture series and the corresponding kinematic analysis.\u003c/p\u003e\n\u003ch3\u003eTwo-choice paradigm and analysis of turning\u003c/h3\u003e\n\u003cp\u003eA two-choice paradigm was used to induce yaw turns as part of behavioural orientation responses towards visual landmarks. Free-walking, intact animals were tested on a flat, black walkway (width: 40 mm) which broadened over a stretch of 6 cm to become either 12 or 14 cm wide (for cuboids and pillars, respectively, see below). The transition occurred at 20 cm and will be referred to as the \u0026lsquo;decision line\u0026rsquo;. Two equal-sized objects were placed at each side of the wide area, at equal distance from the decision line (90 mm). These objects were either white cuboids (40 mm wide; 36 mm high; with or without a black bar at the front) or pillars (20 mm diameter; 100 mm high; either black or white). Similar to the stair-climbing paradigm, a white cardboard shield was positioned behind the setup to increase visual contrast.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 3:\u003c/em\u003e The front sides of the cuboids had one out of three possible appearances: (i) a blank, white surface, (ii) a white surface with a horizontal black bar beneath the upper edge, or (iii) a white surface with a vertical black bar in its middle. The dimensions of the horizontal and vertical bars were the same (15 mm x 36 mm) and were designed such that its size reached an apparent angular width of 10 degrees as seen from the decision line. Two combinations of appearances were tested: either blank against the horizontal bar, or horizontal against vertical bar. For both of these combinations the left-right arrangement was varied pseudo-randomly. In total, each one of five animals performed 80 trials, with 2\u0026times;20 trials per combination. For each animal, the proportion of decisions to climb either one of the stairs was calculated. Effect sizes were quantified as one-sample Cohen\u0026rsquo;s delta, defined as the difference between the mean proportion across animals and the chance level (0.5), divided by the standard deviation across animals.\u003c/p\u003e\n\u003cp\u003eAnother five animals were tested in a paradigm with two vertical pillars, one black and one white. The preference for either the black or the white pillar was tested in 20 trials per animal, again with pseudo-randomised left-right arrangement of the two pillars.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 4:\u003c/em\u003e The two-choice paradigm with pillars was also used in combination with the Vicon motion-capture system, using the same 10-marker arrangement as in \u003cem\u003eexperiment 2\u003c/em\u003e. Eight animals were tested. Acquired marker trajectories were used to analyse front leg kinematics and head orientation, so as to reconstruct the shape and size of the visual image of the landmarks, but also to analyse the shape of the trajectories of both front feet as the animals turned towards one of the two objects. In addition, the yaw angle of the body axis was calculated and used to determine the onset of yaw rotation beyond the decision line.\u003c/p\u003e\n\u003cp\u003eFor the analysis of the step sequences, four subsequent step cycles were first subdivided into stance and swing episodes and then assigned to one of four categories, depending on where they occurred relative to the (virtual) decision line. The swing movement that first crossed the decision line divided the step sequence into two step cycles on the narrow part of the walkway (prior to crossing the decision line) and the subsequent two step cycles on the broad part of the walkway, as the animal turned towards one of the objects.\u003c/p\u003e\n\u003ch3\u003eMarker-based body model and kinematic analysis\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eAfter attaching the markers to the animals, detailed pictures of the animals were taken by a stereo lens (Olympus SZ61), equipped with a digital camera. The pictures were used to measure the exact body and limb dimensions in relation the positions of the markers. These measurements were used to create a unique body model of each of the animals. This was done with a custom-written MATLAB script.\u003c/p\u003e\n\u003cp\u003eThe software Vicon Nexus 1.8.5 was used to control the motion capture system and to track and reconstruct all marker positions in space. For trajectory analysis, each marker was identified and labelled once per recording by hand, and then tracked automatically by the software. Marker tracking was generally robust, provided that a given marker was recorded by least two cameras at any time. In case of small gaps (\u0026lt;10 frames, i.e. 50 ms), marker trajectories were interpolated with an algorithm of the Nexus software. In case of bigger tracking issues, trials were excluded from the analysis.\u003c/p\u003e\n\u003cp\u003eMarker trajectories were then exported into Matlab where they were analysed further, in conjunction with the corresponding individual body model. Whole-body kinematics were calculated using kinematic calculations as described by Theunissen and D\u0026uuml;rr (2013). For the results described here, we used positions and yaw/pitch angles of the body axis and of the head, as well as body-centred position coordinates and speed of the tibia-tarsus joints of both front legs. The transition from swing to stance, i.e., touch-down of the foot, was detected as a drop in foot velocity below a threshold of 60 mm/s. In contrast, the transition from stance to swing, i.e. the lift-off of the foot, was detected as an increase beyond a threshold of 110 mm/s. To simplify the analysis of the foot trajectories, each swing or stance episode was described by four points: the posterior extreme position (PEP) or point of lift-off, the anterior extreme position (AEP) or point of touch-down, and the points at one third or two thirds of its duration. To capture the turning tendency of the animal, we quantified the left-right asymmetry of the left and right front leg movement. To do so we calculated the position difference for the four points per swing or stance trajectories of the left and right foot in the body-fixed coordinate system. The left-right asymmetry was then calculated as the slope of the regression line through the four \u0026ldquo;difference points\u0026rdquo;. To account for the step sequence, this was done separately for trials in which either the left or right front legs crossed the decision line first. Step cycles were pooled according to whether they belonged to the inner or outer leg with respect to turning direction and curvature of the walked path. Statistics were calculated for per-animal means. In case of left-right asymmetry, this was done for mean regression line slopes per animal, separately for swing and stance phases, using a Friedman test for repeated (matched) samples consisting of four subsequent steps, followed by Wilcoxon\u0026rsquo;s test for matched pairs as post-hoc tests.\u003c/p\u003e"},{"header":"Results","content":"\u003ch3\u003eStick insects do not use visual cues for the initiation of climbing\u003c/h3\u003e\n\u003cp\u003eIn order to test the hypothesis that stick insects use visual cues to initiate climbing, we used the same stair-climbing paradigm as D\u0026uuml;rr et al., (2003) and Krause and D\u0026uuml;rr (2012), but with a high-contrast visual landmark, i.e. a black horizontal bar, marking the upper edge of the stair (Fig. 1A). The width of black bar was chosen such that it subtended a visual angle of approximately 10 degrees at a distance of 45 mm in front of the stair. After a first set of trials with a stair of the height 48 mm, it became clear that antennectomized animals hardly ever climbed the front wall of the stair, but climbed onto one of the side walls instead. As a consequence, we reduced the height of the stair to 24 mm, which is equivalent 2 to 3 times the typical clearing beneath the prothorax. A total of ten animals were tested sequentially in three conditions, with ten trials per condition. A first cohort of five animals was first tested intact, then without antennae, and finally after additional blindfolding. A second cohort was first tested intact, then after blindfolding, and finally after additional antennectomy (cut antennae). The initiation of climbing was assessed based on the shape of the head trajectory, which\u0026nbsp;D\u0026uuml;rr et al. (2003; their Fig. 5B) had shown to get much closer to the stair whenever the antennae were cut. Our hypothesis was that animals without antennae but with intact vision would exploit the presence of the visual cue such that their head would not approach the stair as much as in antennectomized and blind animals. Instead, we expected that the head trajectory would rise as early and smoothly as in intact animals.\u003c/p\u003e\n\u003cp\u003eFig. 1B shows that this was not the case. As expected, blindfolded animals with intact antennae climbed the stair virtually the same as intact animals did (left panels in Fig. 1B). However, head trajectories of animals with shortened antennae, irrespective of whether they could see or not, got much closer to the stair and followed much steeper head trajectories than animals with antennae. On average, intact animals first touched the stair at a distance of 29.1\u0026nbsp;\u0026plusmn;\u0026nbsp;4.9 mm. This distance remained nearly the same after blindfolding (29.6\u0026nbsp;\u0026plusmn;\u0026nbsp;3.9 mm) but approximately halved whenever the antennae were cut (sighted: 14.5\u0026nbsp;\u0026plusmn; 8.0 mm; blind: 14.6 \u0026plusmn; 7.6 mm). The same effect was found for the shortest distance to the bottom of the edge, which was nearly the same for intact and blindfolded animals (intact: 14.4\u0026nbsp;\u0026plusmn; 5.8 mm; blind: 15.0 \u0026plusmn; 5.3 mm) and approximately halved after antennectomy (sighted: 7.1\u0026nbsp;\u0026plusmn; 3.7 mm; blind: 8.4\u0026nbsp;\u0026plusmn; 4.2 mm). Accordingly, paired t-tests on per-animal means confirmed that contact distances of cohort 1 were significantly shorter after antennectomy (1\u003csup\u003est\u003c/sup\u003e contact: p \u0026lt; 0.001; bottom of edge: p = 0.005) but remained unchanged after additional blindfolding (1\u003csup\u003est\u003c/sup\u003e contact: p = 0.699; bottom of edge: p = 0.734). In cohort 2, contact distances remained the same after blindfolding (1\u003csup\u003est\u003c/sup\u003e contact: p = 0.845; bottom of edge: p = 0.801) but significantly shortened after additional antennectomy (1\u003csup\u003est\u003c/sup\u003e contact: p \u0026lt; 0.001; bottom of edge: p = 0.009).\u003c/p\u003e\n\u003cp\u003eTo test whether animals reared their body axis, i.e. increased pitch after the first contact with the obstacle, we assessed the height difference of the head between the instant of first contact (red symbols in Fig. 1B) and the closest point to the obstacle (green symbols in Fig. 1B). Pairwise t-tests revealed significantly less rearing in sighted but antennectomized animals of cohort 1 (top row in Fig. 1B, 3.94 \u0026plusmn; 1.59 mm less vertical shift, p = 0.0077) but no difference in blindfolded animals of cohort 2 (left column in Fig. 1B, 0.61 \u0026plusmn; 2.21 mm less vertical shift, p = 0.6074). In both cohorts, rearing was significantly reduced after both treatments (lower right corner in Fig. 1B; with 4.11 \u0026plusmn; 1.83, p = 0.0109 and 4.22 \u0026plusmn; 1.86 mm, p = 0.0105 less vertical shift in cohorts 1 and 2, respectively). We conclude that animals that could see the landmark but did not detect the obstacle with their antennae (cohort 1 after antennectomy) did not respond to the landmark with increased body pitch.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCompared to these effects in front of the stair, treatments had no significant effect on clearance above the upper edge (intact: 10.2\u0026nbsp;\u0026plusmn; 3.1 mm; 8.4 \u0026plusmn; 3.1 mm after\u0026nbsp;antennectomy, p = 0.271;\u0026nbsp;11.0 \u0026plusmn; 2.9 mm after\u0026nbsp;blindfolding, p = 0.515), suggesting that neither vision or antennal touch were important for the regulation of clearance near the edge. However, vision but not touch affected the overall speed of climbing, as blindfolded animals and animals with both impairments were significantly slower than intact animals (intact: 40.9\u0026nbsp;\u0026plusmn;\u0026nbsp;4.7 mm/s; blindfolded: 38.1\u0026nbsp;\u0026plusmn;\u0026nbsp;2.0 mm/s, p = 0.006; combined treatment: 33.9\u0026nbsp;\u0026plusmn;\u0026nbsp;5.0 mm/s, p = 0.0013), whereas antennectomy alone had no significant effect (36.4\u0026nbsp;\u0026plusmn;\u0026nbsp;8.8 mm/s, p = 0.336). We conclude that the effect of vision on overall speed is more pronounced than that of antennal touch.\u003c/p\u003e\n\u003cp\u003eFrom the stick insects\u0026rsquo; perspective, the horizontal black bar might have appeared to hover above the ground, as the bar had no connection to the black walkway. To make sure that sighted but antennectomized animals did not disregard visual cues because of a particular feature of the visual pattern, we conducted additional tests with two visual patterns that had vertical black-and-white contrast edges between the walkway and the top of the stair. The effect of antennectomy remained the same, with significantly less rearing and significantly shorter distances to the stair despite intact vision (Fig. S1 and Table S1 in the Suppl. Mat.). This corroborates our interpretation that sighted animals make no use of visual cues about obstacle height during the onset of climbing.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eAt initiation of climbing, landmarks were visually discernible\u003c/h3\u003e\n\u003cp\u003eSince our results suggest that sighted stick insects without antennae do not use the visual cues to initiate climbing as early as intact animals do, it was possible that this happened because the animals were unable to see the visual landmark sufficiently early to respond to it. To assess whether this was the case, we conducted high-precision motion capture measurements to estimate the image size of the black bar and its increase during the approach of the insect. We used intact animals for this because we wanted to know the image size at the beginning of the smooth and early rise of the head of intact animals. Fig. 2A shows the 3D reconstruction of a sample trial with four postures per second (every 50 frames). Based on this reconstruction, the location and apparent angular size of the bar within the visual field was estimated for three instants during the approach (Fig. 2B). The first of these images was calculated for the beginning of the trajectory. Here, the image of the black bar subtended approximately 20 x 4 degrees. The second image was calculated for the instant of the first contact with the stair, when the bar subtended approximately 45 x 15 degrees. The third image was calculated for the middle of the climbing process, when the head was just beneath the lower edge of the landmark. Here, the bar subtended apparent size of 100 x 20 degrees. Assuming a spatial resolution of approximately 5 degrees (Jander and Volk-Heinrichs, 1970), this image sequence suggests that the stick insect could have easily resolved the black bar at the time of initial contact. The bar width had reached approximately three times the visual resolution by then.\u003c/p\u003e\n\u003cp\u003eFigures 2 C and 2 D further substantiate this conclusion. They show three single trials (Fig. 2 C) and per-animal means of head trajectories (top), head pitch (middle) and the corresponding apparent bar width in the visual field of the animal (bottom). The head trajectories within one animal (Fig. 2 C), but also in comparison among animals (Fig. 2D) had an overall uniform sigmoid shape, indicating that both the distance at which the animals started to climb and the course of the climb were consistent both among trials and among animals. The same was found to be true for the head pitch of the animals (mid panels of Fig. 2 C, D). Along the straight walkway, head pitch fluctuated mildly about a constant, neutral value (approximately 0 degrees) until the insect started to climb. The head then began to face upwards towards the top edge of the stair. As the animal pulled itself up, its head pitch decreased such that it faced downwards in relation to the body axis. The apparent visual size of the bar changed accordingly (lower panels in Fig. 2 C, D). About 50 mm in front of the stair, the angular width of the bar exceeded 10 degrees. Given the high image contrast and an inter-ommatidial angle of approximately 5 degrees (Jander and Volk-Heinrichs, 1970), we conclude that stick insects would have been able to discern the visual landmark prior to the initiation of climbing.\u003c/p\u003e\n\u003ch3\u003eLandmarks not used for climbing reliably induce yaw turns\u003c/h3\u003e\n\u003cp\u003eUsing a two-choice paradigm, we were able to prove that stick insects do indeed show visually induced responses to a bar-shaped landmark of width 10 degrees (Fig. 3). When given the choice between a horizontal bar of the same size as the one previously ignored and a blank white surface, stick insects walked towards the stair with the black bar in 87.5 % (effect size: Cohen\u0026rsquo;s\u0026nbsp;d\u0026nbsp;= 6.4) of all trials (Fig. 3, left). The positions of the two alternative objects were swapped pseudo-randomly, and no side preference was observed in any of the animals.\u003c/p\u003e\n\u003cp\u003eWhen given the choice between horizontal and vertical black landmarks of equal size, animals reliably chose the vertical landmark in 93.5% (effect size: Cohen\u0026rsquo;s d = 11.5) of all trials (Fig. 3, right). We conclude that stick insects reliably change their movement direction towards the same visual landmark that was not used during climbing. The disregard of the landmark during climbing was not due to the image size or the type of landmark presented, but rather due to the kind of directional movement to be controlled: the same landmark that was ignored during climbing reliably induced yaw-turning.\u003c/p\u003e\n\u003ch3\u003eVisually induced yaw turns become overt during the first swing phase past the decision line\u003c/h3\u003e\n\u003cp\u003eTo investigate when and how visually induced yaw turns were initiated and executed, we analysed the foot trajectories in a paradigm in which they would reliably turn towards a static visual landmark. Given the results shown in Fig. 3, we opted for vertical landmarks. Preliminary behavioural tests with five animals revealed that in a choice between a white and a black pillar, animals turned towards the white pillar in only 0.5 % of cases (Fig. 4 A; only a single trial of a single animal in a total of 200 trials). Using the same motion capture technique as for Fig. 2, body postures were reconstructed for entire walking trials at a temporal resolution of 0.05 s. Fig. 4 B shows a subset of postures for a representative trial of a visually induced turn towards the black pillar. The analysis began as the animal entered a zone 100 mm prior to the (virtual) decision line and ended as soon as either a front foot or an antenna entered a circular environment at 5 mm distance around the pillar (see red ring for exclusion criterion in Fig. 4 C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs the objective of this analysis was to tell when and how stick insects initiated a yaw turn, we focused on two target variables. The first was the yaw angle of the prothorax, which we used to tell the onset of yaw rotation of the body axis (Fig. 5). The second was the left-right asymmetry between corresponding points on the foot trajectories of both front legs, as front legs are known to initiate yaw turns in tethered walking stick insects (D\u0026uuml;rr and Ebeling, 2005). To assess asymmetries of front leg movements, the sequence of four step cycles per front leg was taken into account, two before and two after crossing the decision line (Fig. 4 C). Step cycles were separated into swing and stance movements (Fig. 6 A), the trajectories of which were described by per-animal means (Fig. 6 B).\u003c/p\u003e\n\u003cp\u003eConcerning the body axis, the onset of turning was judged from prothorax trajectories (Fig. 5, top panels; left panel shows the mean \u0026plusmn; s.d. of the y position of two animals; right panel shows mean y positions of all eight animals). Overall, the narrow part of the walkway constrained the body trajectories to a consistently narrow range near the middle of the straight walkway. In particular, we used the yaw angle of the prothorax, as a function of prothorax position on the walkway (Fig. 5, mid panels). In order to remove potential directional biases, we subtracted the mean yaw angle measured within a reference range of -80 mm to -20 mm before reaching the decision line. Given the fluctuation of the body axis yaw angle during locomotion, we used three times its standard deviation within that same reference range as a threshold for detecting the onset of yaw rotation. The eight colored dots in the mid panel of Fig. 5B indicate the mean distances where this threshold was exceeded. Among the eight animals, this happened between 14 and 37 mm (25 \u0026plusmn; 9 mm, N = 8) behind the decision line, indicating that some animals turned earlier than others. At the time of crossing the decision line, the angular landmark width had already exceeded 10 degrees (Fig. 5, bottom), such that animals must have been able to discern the pillar no later than when the decision line was reached.\u003c/p\u003e\n\u003cp\u003eSince the torque that is required for initiating a turn of the body axis must be generated by some sort of a left-right asymmetry in the leg movement, we wanted to know when such an asymmetry occurred for the first time. To this end, we compared the foot trajectories of the two front legs by matching corresponding pairs of swing and stance phases of the left and right front legs (Fig. 4 C). This was done for a set of four subsequent step cycles of interest, resulting in 14 to 30 body-centred foot trajectories per animal, step category and step cycle phase (Fig. 6A). Note that the movement of stance and swing trajectories was counter-directed, with stance phases beginning at the anterior extreme position (AEP) and ending at the posterior extreme position (PEP), and swing phases moving from PEP to AEP. Furthermore, the y-coordinates of leftward turns were flipped (owing to the pseudo-randomised position of the black pillar and corresponding turning direction), such that bluish colors label the outer leg and reddish colors label the inner leg with respect to the turning direction.\u003c/p\u003e\n\u003cp\u003eFinally, average trajectories were first calculated per animal (black lines in Fig. 6 A) and then summarised as mean positions among animals (mean \u0026plusmn; s.d. of per-animal means), using four points per movement phase (Fig. 6 B). Accordingly, mean trajectories of swing phases turned out to be strongly curved, whereas curvature was weak during stance. Mean trajectories for subsequent step cycles had the same overall shape and appeared bilaterally symmetrical for step categories 1 and 2 (before crossing the decision line) but with increasing left-right asymmetry for step categories 3 and 4 (behind the decision line). In order to quantify this left-right asymmetry, we calculated the position differences between inner and outer legs (black lines and symbols in Fig. 6 B) and determined the slope \u0026Delta;y/\u0026Delta;x for each set of four position differences. We reasoned that the first step cycle phase to initiate the yaw turn would be the first for which the slope deviated from zero in a statistically significant manner. To account for repeated, matched measurements of slopes we first tested for overall changes in slope, separately for swing and stance trajectories. Mean slopes differed significantly across the four subsequent steps, both for swing and stance phases (Friedman test; N = 8; stance: p = 0.0045; p = 0.001). Significant differences between step categories were then tested post-hoc with Wilcoxon\u0026rsquo;s test for matched pairs (again on per-animal means, N = 8). Within the entire data set shown in Fig. 6 B, the stance phase of step 4 was the first to differed significantly from all preceding steps (difference to steps 3, 2 and 1 with p = 0.015, 0.007 and 0.007, respectively), whereas step 3 did not differ from step 2. In contrast, the swing phases of step 3 was the first to differ significantly from all preceding steps (i.e., steps 2 and 1, with p = 0.015 and 0.023, respectively). The swing phase of step 4 also differed from that of steps 2 and 1 (p = 0.007 and 0.007, respectively), but not from step 3. We conclude that the left-right asymmetry of front leg movement was detectable first in a directional change of the first swing movement after crossing the decision line (step 3), with the first significant change in stance direction occurring during the immediately subsequent step 4.\u003c/p\u003e\n\u003cp\u003eSince this kind of analysis was designed to detect spatial differences in the sequence of steps, it neglected the temporal delay between steps of the inner and outer leg. However, animals crossed the decision line first with their inner leg in some trials, but with the outer leg first in other trials. Therefore, the pooled data set of Fig. 6 B was split according to which front leg crossed the decision line first. Fig. 7 shows the corresponding mean trajectories for \u0026ldquo;inner-leg-first trials\u0026rdquo; (Fig. 7 A) and \u0026ldquo;outer-leg-first trials\u0026rdquo; (Fig 7 B). Overall, the statistical results substantiated the analysis of the pooled data set, with both step phases becoming significantly asymmetric in inner-leg-first trials (Friedman test, N = 8, stance: p = 0.0359; swing: p = 0.001), however with only the stance phase doing so in outer-leg-first trials (Friedman test, N = 8, stance: p = 0.0239; swing: p = 0.0703).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn inner-leg-first trials (Fig. 7 A) post-hoc tests revealed statistically significant differences for stance phases between steps 4 and 2 only (p= 0.007; n.s. for steps 3 or 1), whereas for swing phases, the first statistical difference occurred between steps 3 and 2 (p = 0.007), followed by a further change between steps 4 and 3 (p= 0.04). Similarly, in outer-leg-first trials (Fig. 7 B) the first significant difference between stance phases occurred between steps 4 and 2 (p = 0.007). Outer-leg-first trials showed no significant differences between any pair of subsequent steps. We conclude that, like in the pooled sample of Fig. 6, first significant spatial asymmetry occurred in the first swing phase after crossing the decision line (step 3), and/or in \u0026nbsp;the second stance phase (step 4). If the outer leg was first to cross the decision line, it took one complete step cycle of that leg (step 3) before its second stance phase changed direction. If the inner leg was first to cross the decision line, it took only one stance phase of that leg before its first swing phase changed direction. We conclude that the delay for initiating a yaw turn is one stance phase and an entire step cycle.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eAs a climbing insect that dwells in complex-structured environments like thickets of bramble or ivy, stick insects need to be able to change their heading with both vertical (pitch) and horizontal (yaw) turns. Our results show that stick insects reliably turn towards a visual landmark in the yaw direction (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), whereas they do not exploit the presence of the very same landmark for a timely change of body pitch (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Using a paradigm that constrains the onset of visually induced yaw turns (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) we found that a change in heading occurred approximately at the distance covered during one swing or stance movement (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). On average, the first left-right asymmetry in stepping occurred during a swing, with a delay of an entire stance phase (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Depending on which front leg first stepped beyond the turning constraint, the delay could be an entire step cycle, and the first asymmetry could occur during stance (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eVisual landmarks do not affect the initiation of climbing\u003c/h2\u003e \u003cp\u003eWhen climbing an obstacle, insects must change their body pitch angle. While this change in body pitch has been studied in relation to tactile and/or visual cues in several insect species (e.g., (Pelletier and McLeod, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e1994\u003c/span\u003e; Zurek and Gilbert, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Clifton et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) only few studies have related the change in pitch to the leg movements that cause it. Those that did, identified two distinct leg movement strategies: One strategy is to rear their body axis by differential extension of front, middle and hind legs in stance, thus raising the prothorax more than their metathorax. This has been documented for cockroaches (Watson et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) and, indirectly, for stick insects (Sch\u0026uuml;tz and D\u0026uuml;rr, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). In stick insects, the change in pitch was small but sufficiently fast to be detectable between two subsequent antennal contacts with the obstacle (in the range of 100 ms, i.e. faster than a swing movement of a leg). The other strategy is to grasp hold of the obstacle with a front leg and pull the body upwards. This has been found to occur in combination with rearing in cockroaches and stick insects (same studies as before), with stick insects showing tactually induced changes in leg movement within 40 ms, for example during re-targeting of an ongoing swing movement. So far, only few insect species have been shown to adjust foothold or grasping position in response to visual cues: Locusts and grasshoppers adjust their front leg swing movements so as to reach for footholds in climbing paradigms (Niven et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Niven et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Praying mantises adjust their front leg strike direction to visual targets (Prete and Cleal, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e1996\u003c/span\u003e). In contrast to this, we found no evidence for stick insects to make use of a high-contrast landmark to adjust body pitch so as to improve climbing performance. The animals tested in experiment 1 behaved in the exact same way as reported for climbing in darkness (D\u0026uuml;rr et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2003\u003c/span\u003e): The presence of a high-contrast visual landmark did not rescue the effect of antennectomy, such that sighted animals without antennae showed the same head trajectory as blind animals without antennae (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Suppl. Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). This could not be explained by the animals not being able to see the landmark, for two reasons: First, the apparent landmark size in the visual field of the animals (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) was sufficiently large for the animals to discern the black bar against the white surrounding. Second, a same landmark with the same shape and visual image size reliably elicited a yaw turning response (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe conclude that seeing the landmark did not improve their climbing efficiency by a timely inclination of the body axis. There are two possible explanations for this: (i) Stick insects could have ignored the elevated visual landmark unless it was accompanied by near-range information from tactile cues, or (ii) they may not have had enough time to respond with a change in body pitch or targeted reaching. Both of these explanations are related to the estimate of distance. In insects, visual distance information can be inferred from motion parallax, e.g. through peering movements in locusts (Collett, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1978\u003c/span\u003e), from stereopsis, e.g. in striking mantids (Rossel, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1983\u003c/span\u003e) or from image expansion, e.g. during landing or escape (Wagner, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1982\u003c/span\u003e; Fotowat and Gabbiani, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). To date, none of these mechanisms have been described for stick insects. Moreover, even in insects that can estimate distance visually, a relation between visual size and distance needs not exist (Rossel, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e1991\u003c/span\u003e). In other words, the perceived angular size of an elevated landmark alone may be too unreliable to adjust body pitch in preparation of climbing.\u003c/p\u003e \u003cp\u003eRegarding reaction time, visually triggered movements in stick insects must be expected to be considerably slower than tactually elicited movements. Touch-related retargeting of a front leg may be initiated only 40 ms after the first contact (Sch\u0026uuml;tz and D\u0026uuml;rr, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), and descending interneurons convey antennal mechanosensory information to the thorax with an 11 ms delay only (Ache and D\u0026uuml;rr, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). In comparison, visually induced responses in stick insects are slow. For example, changes in front leg movement during turns elicited by large-field visual motion have time constants of 1.5 or 1.7 s in inner and outer front legs, respectively (D\u0026uuml;rr and Ebeling, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Even when considering the long time constant of muscle activation dynamics in insect muscle (Harischandra et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) the onset of the related muscle activity would occur with a delay of at least 1 s. Similarly, we could reliably detect yaw turns towards static landmarks only 17 to 37 mm behind the decision line (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). This is similar to the 24 to 40 mm distance expected to be covered during one stance phase, assuming an average speed of 40 to 50 mm/s and stance durations between 600 to 800 ms. Further considering that the significant change in body pitch found in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e occurred only ca. 15 mm beyond the first contact in blindfolded animals, we conclude that even if the animals could have estimated distance from the expanding image of the visual landmark, they couldn\u0026rsquo;t have responded as early as the blindfolded animals and not early enough to prevent bumping into the obstacle (as many of them did).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eVisually induced yaw turns tend to begin with a change in swing direction\u003c/h2\u003e \u003cp\u003eTheoretically, the torque driving yaw rotation must be generated by some left-right asymmetry of leg movement action. This asymmetry can be generated by means of three, potentially independent movement parameters: step frequency, step length and step direction. In stick insects, visual-motion-induced yaw turns involve asymmetric changes in all three of these parameters, both with and without mechanical coupling among the legs in stance (D\u0026uuml;rr and Ebeling, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Gruhn et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Associated gait changes are governed by changes in pair-wise coupling of neighboring legs, particularly in front legs, including left-right asymmetry of coupling efficacy (D\u0026uuml;rr, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Recently, stick insects were found to turn by means of a combination of changed step direction and the likelihood of a distinct turn-related step class (Meschenmoser and D\u0026uuml;rr, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2026\u003c/span\u003e). To complement this earlier work, the present study focused on step direction asymmetries of the front leg movements, with separate treatment of swing and stance phases. In steady state locomotion, for example when walking a curve of fixed radius with constant speed, directions of stance or swing movements will be inverted by 180\u0026deg;, while step lengths will be equal. At the onset of turning, however, directions of stance and swing will transiently deviate from this 180\u0026deg; inversion. Although it is common to study the mentioned left-right asymmetry during stance (e.g., (Yang et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Chockley et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and tethered walking stick insects were found to change stance parameters slightly faster than swing parameters (D\u0026uuml;rr and Ebeling, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), it is unclear what initiates turning in free walking animals. Transient asymmetries could arise through an obliquely directed swing movement that was to be followed by a stance movement to pull the body towards the new foothold.\u003c/p\u003e \u003cp\u003eOur results show that the earliest step phase with a statistically significant left-right asymmetry is the first swing phase past the \u0026ldquo;decision line\u0026rdquo;, i.e. past the physical widening of the walkway (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). After splitting the trials according to the body side that stepped across the decision line first, this pattern became more pronounced in inner-leg-first trials (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA) and weakened in outer-leg-first trials (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). Assuming that animals decided to turn towards the dark object soon after stepping across the widening, this means that a significant change in leg movement asymmetry required at least the duration of an entire stance phase to take effect (step 3 in Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). While this timing is very similar to that observed by D\u0026uuml;rr and Ebeling (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), their earliest change was a lateral shift in lift-off position and the preceding stance direction, whereas in our pooled data it was a change in swing direction. This difference could be related to (i) a difference in time course analysis, (ii) the biomechanical differences associated to the experimental setup, or (iii) the visual cue used to initiate turning. Regarding methodological differences, D\u0026uuml;rr and Ebeling (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) inferred the time constant of an exponential response variable from pooled data, comprising all step cycles of a trial and varying numbers of steps per animal. In contrast, here, we use matched samples of an ordered sequence of step cycle phases, with means per animal. Furthermore the free walking animals of the present study needed to generate natural yaw torques only (as they rotated their own body, only), whereas the animals of D\u0026uuml;rr and Ebeling (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) needed to rotate a hollow sphere whose moment of inertia was more than 100-fold that of their own body. Finally, D\u0026uuml;rr and Ebeling (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) triggered curve walking by means of a large-field visual-motion cue (provided by a rotating stripe drum) as opposed to the stationary landmark used here. At least for flying insects, visually induced turning responses have been dissociated into motion-related turns as opposed to position-related cues (Buchner, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1984\u003c/span\u003e). Accordingly, the position of our stationary landmark may have triggered turning through a change in swing direction, whereas a large-field motion stimulus could trigger turning by a change in stance. Whether this differentiation of position and motion-induced turning holds for walking insects may be tested by future studies.\u003c/p\u003e \u003cp\u003eWe conclude that visually induced yaw turns towards static visual landmarks may be initiated by a change in front leg swing direction. This is reminiscent of the fact that tactually induced turning is accompanied with an increased likelihood of reaching movements (Berendes and D\u0026uuml;rr, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Other than horizontal turning, climbing must involve a change in foothold height, which can only occur during swing during swing. The initiation of both pitch and yaw turns by means of a directed swing movement could be a general strategy for initiating omnidirectional changes in heading in terrestrial locomotion.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eR.B. conducted experiments, analysed data, prepared figures 2-7, wrote and reviewed the manuscript; N.S. conducted experiments, analysed data, prepared figure 1 and Suppl. Fig. 1, and reviewed the manuscript; V.D. conceived the experiments, supervised R.B. and N.S., analysed data, wrote and reviewed the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe thank Florian P. Schmidt for technical assistance and Merit Meschenmoser for critically reading a previous version of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003e\u003cstrong\u003eAche, J. M. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2013). Encoding of near-range spatial information by descending interneurons in the stick insect antennal mechanosensory pathway. \u003cem\u003eJ. Neurophysiol.\u003c/em\u003e \u003cstrong\u003e110\u003c/strong\u003e, 2099\u0026ndash;2112.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eBerendes, V. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2022). Active tactile exploration and tactually induced turning in tethered walking stick insects. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e225\u003c/strong\u003e, jeb243190.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eBuchner, E\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1984). Behavioral analysis of spatial vision in insects. In \u003cem\u003ePhotoreception and Vision in Invertebrates \u003c/em\u003e(ed. M. Ali), pp. 561\u0026ndash;621. New York, London: Plenum Press.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCamhi, J. M. and Johnson, E. N\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1999). High-frequency steering maneuvers mediated by tactile cues: Antennal wall-following in the cockroach. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e202\u003c/strong\u003e, 631\u0026ndash;643.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eChockley, A. S., Dinges, G. F., Di Cristina, G., Ratican, S., Bockem\u0026uuml;hl, T. and B\u0026uuml;schges, A\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2022). Subsets of leg proprioceptors influence leg kinematics but not interleg coordination in \u003cem\u003eDrosophila melanogaster \u003c/em\u003ewalking. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e https://doi.org/10.1242/jeb.244245.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCleal, K. S. and Prete, F. R\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1996). The predatory strike of free ranging praying mantises, \u003cem\u003eSphodromantis lineola\u003c/em\u003e (Burmeister). 2. Strikes in the horizontal plane. \u003cem\u003eBrain Behavior and Evolution\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 191\u0026ndash;204.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eClifton, G., Stark, A. Y., Li, C. and Gravish, N\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2023). The bumpy road ahead: the role of substrate roughness on animal walking and a proposed comparative metric. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e https://doi.org/10.1242/jeb.245261.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eClifton, G. T., Holway, D. and Gravish, N\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2020). Vision does not impact walking performance in Argentine ants. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e https://doi.org/10.1242/jeb.228460.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCollett, T. S\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1978). Peering - Locust behavior pattern for obtaining motion parallax information. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 237\u0026ndash;241.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCowan, N. J., Lee, J. and Full, R. J\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2006). Task-level control of rapid wall following in the American cockroach. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e209\u003c/strong\u003e, 1617\u0026ndash;1629.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCruse, H., Dean, J., Kindermann, T., Schmitz, J. and Schumm, M\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1998). Simulation of complex movements using artificial neural networks. \u003cem\u003eZeitschrift fur Naturforschung C-A Journal of Biosciences\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 628\u0026ndash;638.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eDudley, R\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2002). Mechanisms and implications of animal flight maneuverability. \u003cem\u003eIntegr Comp Biol\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 135\u0026ndash;140.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V., K\u0026ouml;nig, Y. and Kittmann, R\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2001). The antennal motor system of the stick insect \u003cem\u003eCarausius morosus:\u003c/em\u003e anatomy and antennal movement pattern during walking. \u003cem\u003eJ. Comp. Physiol. A\u003c/em\u003e \u003cstrong\u003e187\u003c/strong\u003e, 131\u0026ndash;144.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V., Krause, A. F., Schmitz, J. and Cruse, H\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2003). Neuroethological concepts and their transfer to walking machines. \u003cem\u003eInt. J. Robot. Res.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 151\u0026ndash;167.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2005). Context-dependent changes in strength and efficacy of leg coordination mechanisms. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e208\u003c/strong\u003e, 2253\u0026ndash;2267.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V. and Ebeling, W\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2005). The behavioural transition from straight to curve walking: kinetics of leg movement parameters and the initiation of turning. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e208\u003c/strong\u003e, 2237\u0026ndash;2252.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V. and Schilling, M\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2018). Transfer of spatial contact information among limbs and the notion of peripersonal space in insects. \u003cem\u003eFront. Comput. Neurosci.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 101.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V., Arena, P., Cruse, H., Dallmann, C. J., Drimus, A., Hoinville, T., Krause, T., Matefi-Tempfli, S., Paskarbeit, J., Patane, L. et al\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2019). Integrative biomimetics of autonomous hexapedal locomotion. \u003cem\u003eFrontiers in Neurorobotics\u003c/em\u003e. https://doi.org/10.3389/fnbot.2019.00088.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V., Berendes, V. and Strube-Bloss, M. F\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2022). Sensorimotor ecology of the insect antenna: Active sampling by a multimodal sensory organ. \u003cem\u003eAdv. Insect Physiol.\u003c/em\u003e \u003cstrong\u003e63\u003c/strong\u003e, 1\u0026ndash;105.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eD\u0026uuml;rr, V. and Mesanovic, A\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2023). Behavioural function and development of body-to-limb proportions and active movement ranges in three stick insect species. \u003cem\u003eJ. Comp. Physiol. A\u003c/em\u003e, 265\u0026ndash;284.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eFotowat, H. and Gabbiani, F\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2007). Relationship between the Phases of Sensory and Motor Activity during a Looming-Evoked Multistage Escape Behavior. \u003cem\u003eJ. Neurosci.\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 10047\u0026ndash;10059.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eGruhn, M., Zehl, L. and B\u0026uuml;schges, A\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2009). Straight walking and turning on a slippery surface. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e212\u003c/strong\u003e, 194\u0026ndash;209.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eHarischandra, N., Krause, A. F. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2015). Stable phase-shift despite quasi-rhythmic movements: a CPG-driven dynamic model of active tactile exploration in an insect. \u003cem\u003eFront. Comput. Neurosci.\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 107.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eHarischandra, N., Clare, A. J., Zakotnik, J., Blackburn, L. M. L., Matheson, T. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2019). Evaluation of linear and non-linear activation dynamics models for insect muscle. \u003cem\u003ePlos Computational Biology\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, e1007437.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eHarley, C. M., English, B. A. and Ritzmann, R. E\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2009). Characterization of obstacle negotiation behaviors in the cockroach, \u003cem\u003eBlaberus discoidalis\u003c/em\u003e. \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e212\u003c/strong\u003e, 1463\u0026ndash;1476.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eJander, R. and Volk-Heinrichs, I\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1970). Das Strauch-spezifische Perceptor-System der Stabheuschrecke (\u003cem\u003eCarausius morosus\u003c/em\u003e). \u003cem\u003eZ. vergl. Physiol.\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 425\u0026ndash;447.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eKalmus, H\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1937). Photohorotaxis, eine neue Reaktionsart, gefunden an den Eilarven von \u003cem\u003eDixippus morosus\u003c/em\u003e. \u003cem\u003eZ. vergl. Physiol.\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 644\u0026ndash;655.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eKrause, A. F. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2012). Active tactile sampling by an insect in a step-climbing paradigm. \u003cem\u003eFront. Behav. Neurosci.\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 1\u0026ndash;17.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eMeschenmoser, M. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2025). Contrast and luminance dependence of target choice and visual orientation in walking stick insects. \u003cem\u003eScientific Reports\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 12226.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eMeschenmoser, M. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2026). Walking in circles: freely walking stick insects turn by a combination of step-parameter scaling and altered likelihood of particular step types. \u003cem\u003ebioRxiv\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eNiven, J. E., Buckingham, C. J., Lumley, S., Cuttle, M. F. and Laughlin, S. B\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2010). Visual targeting of forelimbs in ladder-walking locusts. \u003cem\u003eCurr. Biol.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 86\u0026ndash;91.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eNiven, J. E., Ott, S. R. and Rogers, S. M\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2012). Visually targeted reaching in horse-head grasshoppers. \u003cem\u003eProc. R. Soc. Lond. B\u003c/em\u003e \u003cstrong\u003e279\u003c/strong\u003e, 3697\u0026ndash;3705.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eOkada, J. and Toh, Y\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2000). The role of antennal hair plates in object-guided tactile orientation of the cockroach (\u003cem\u003ePeriplaneta americana\u003c/em\u003e). \u003cem\u003eJ. Comp. Physiol. A\u003c/em\u003e \u003cstrong\u003e186\u003c/strong\u003e, 849\u0026ndash;857.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eOkada, J. and Toh, Y\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2006). Active tactile sensing for localization of objects by the cockroach antenna. \u003cem\u003eJ. Comp. Physiol. A\u003c/em\u003e \u003cstrong\u003e192\u003c/strong\u003e, 715\u0026ndash;726.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ePelletier, Y. and McLeod, C. D\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1994). Obstacle perception by insect antennae during terrestrial locomotion. \u003cem\u003ePhysiol. Entomol.\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 360\u0026ndash;362.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ePick, S. and Strauss, R\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2005). Goal-driven behavioral adaptations in gap-climbing \u003cem\u003eDrosophila\u003c/em\u003e. \u003cem\u003eCurr. Biol.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1473\u0026ndash;1478.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ePrete, F. R. and Cleal, D. S\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1996). The predatory strike of free ranging praying mantises, \u003cem\u003eSphodromantis lineola\u003c/em\u003e (Burmeister).1. Strikes in the mid-sagittal plane. \u003cem\u003eBrain Behavior and Evolution\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 173\u0026ndash;190.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eRossel, S\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1983). Binocular stereopsis in an insect. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e302\u003c/strong\u003e, 821\u0026ndash;822.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eRossel, S\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1991). Spatial vision in the praying mantis: is distance implicated in size detection ? \u003cem\u003eJ. Comp. Physiol.\u003c/em\u003e \u003cstrong\u003e169\u003c/strong\u003e, 101\u0026ndash;108.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eSch\u0026uuml;tz, C. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2011). Active tactile exploration for adaptive locomotion in the stick insect. \u003cem\u003ePhil. Trans. R. Soc. Lond. B\u003c/em\u003e \u003cstrong\u003e366\u003c/strong\u003e, 2996\u0026ndash;3005.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eTaylor, G. K\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2001). Mechanics and aerodynamics of insect flight control. \u003cem\u003eBiol. Rev.\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 449\u0026ndash;471.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eTheunissen, L. M. and D\u0026uuml;rr, V\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2013). Insects use two distinct classes of steps during unrestrained locomotion. \u003cem\u003ePLOS one\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, e85321.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eWagner, H\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (1982). Flow-field variables trigger landing in flies. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e297\u003c/strong\u003e, 147\u0026ndash;148.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eWatson, J. T., Ritzmann, R. E., Zill, S. N. and Pollack, A. J\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2002). Control of obstacle climbing in the cockroach, \u003cem\u003eBlaberus discoidalis\u003c/em\u003e. I. Kinematics. \u003cem\u003eJ. Comp. Physiol. A\u003c/em\u003e \u003cstrong\u003e188\u003c/strong\u003e, 39\u0026ndash;53.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eWebb, P. W\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2004). Maneuverability - general issues. \u003cem\u003eIEEE Journal of Oceanic Engineering\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 547\u0026ndash;555.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eYang, H. H., Brezovec, B. E., Serratosa Capdevila, L., Vanderbeck, Q. X., Adachi, A., Mann, R. S. and Wilson, R. I\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2024). Fine-grained descending control of steering in walking \u003cem\u003eDrosophila\u003c/em\u003e. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e187\u003c/strong\u003e, 6290-6308.e27.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eZeng, Y., Chang, S. W., Williams, J. Y., Nguyen, L. Y.-N., Tang, J., Naing, G. and Dudley, R\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2020). Canopy parkour: movement ecology of post-hatch dispersal in a gliding nymphal stick insect (\u003cem\u003eExtatosoma tiaratum\u003c/em\u003e). \u003cem\u003eJ. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e223\u003c/strong\u003e, jeb226266.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eZurek, D. B. and Gilbert, C\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e (2014). Static antennae act as locomotory guides that compensate for visual motion blur in a diurnal, keen-eyed predator. \u003cem\u003eProc. R. Soc. Lond. B\u003c/em\u003e. https://doi.org/10.1098/rspb.2013.3072.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-comparative-physiology-a","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jcpa","sideBox":"Learn more about [Journal of Comparative Physiology A](http://link.springer.com/journal/359)","snPcode":"359","submissionUrl":"https://submission.nature.com/new-submission/359/3","title":"Journal of Comparative Physiology A","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"yaw rotation, pitch rotation, visually-induced turning, swing movement, locomotion, Carausius","lastPublishedDoi":"10.21203/rs.3.rs-9243999/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9243999/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eClimbing animals need to be able to adjust their heading in both vertical and horizontal directions, requiring pitch and yaw rotation of the body axis, respectively. Stick insects are climbing herbivors that dwell among leaves and branches of the vegetation they feed on. By nature, they need to adjust their 3D heading in a complex environment. Since insects in general turn towards visual landmarks, we test whether or not stick insects initiate pitch rotation in response to landmarks that reliably elicit yaw rotation. We show that this is not the case. Instead, tactile cues alone are sufficient to mask the effect of visual deprivation, whereas lack of tactile cues strongly delays body inclination and the onset of climbing, even when same visual landmark cues are present that are sufficient to induce turning reliably. In a second set of experiments, we use a setup that constrains the onset of turning to tell whether visually induced yaw rotation is initiated by swing movements or rather by stance movements of the front legs. We show that a turn-related left-right asymmetry of front leg movement begins at least one stance phase after passing the constraint and may begin with the first swing movement thereafter. We conclude that stick insects do not exploit visual cues for initiating pitch rotation, despite they reliably initiate yaw rotation towards the exact same landmark. Moreover, we demonstrate that yaw rotation towards visual landmarks may begin with a swing movement of a front leg, much like the pitch rotation at the onset of climbing.\u003c/p\u003e","manuscriptTitle":"Changing 3D heading direction in terrestrial locomotion: Effectiveness of visual landmarks and the onset of turning in stick insects","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-07 09:12:55","doi":"10.21203/rs.3.rs-9243999/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-08T06:50:41+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-08T06:07:39+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-11T19:31:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"40943042942607299283372971889030217312","date":"2026-04-07T23:32:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233542680038403963052921670457035772227","date":"2026-03-31T16:08:11+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-31T15:50:15+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-30T21:33:15+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-30T21:32:54+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Comparative Physiology A","date":"2026-03-27T10:53:11+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-comparative-physiology-a","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jcpa","sideBox":"Learn more about [Journal of Comparative Physiology A](http://link.springer.com/journal/359)","snPcode":"359","submissionUrl":"https://submission.nature.com/new-submission/359/3","title":"Journal of Comparative Physiology A","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"79b8c38a-6add-4e9b-bd98-d3bbad4929a4","owner":[],"postedDate":"April 7th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-08T06:50:41+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-08T06:07:39+00:00","index":16,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-08T06:56:27+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-07 09:12:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9243999","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9243999","identity":"rs-9243999","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}