Collective motion as a distinct behavioral state of the individual

•Locusts were monitored before, during, and after experiencing collective motion•In each condition the locusts showed distinct walking kinematics•This indicates that adopt collective-motion-dependent behavioral states•Simulations show these states may be advantageous for swarm integrity The motion of swarms depends on adaptations at individual level. We explored their effects formation maintenance in locusts. kinematics insects under laboratory settings, as well during a group, again separation from group. It was found taking part induced unique kinematics, suggesting existence mode we term “collective-motion-state.” This state, characterized by adaptation to social context, is long lasting, not crowding per se, but only motion. Utilizing computational models, this adaptability increases robustness swarm. Overall, our findings suggest an emergent property group also mode, rooted endogenous mechanisms individual. ability form groups move collectively key feature many species (Sumpter, 2006Sumpter D.J.T. principles animal behaviour.Philos. Trans. R. Soc. B Biol. Sci. 2006; 361: 5-22Crossref PubMed Scopus (610) Google Scholar; Ward Webster, 2016Ward A. Webster M. Sociality: Behaviour Group-Living Animals. Springer International Publishing, 2016Crossref (190) Scholar), assumed increase survival both individuals (Be’er Ariel, 2019Be’er Ariel G. A statistical physics view swarming bacteria.Mov. Ecol. 2019; 7: 1-17PubMed Yang Schmickl, 2019Yang W.C. Schmickl T. Collective ultimate effect crowded selfish herds.Sci. Rep. 9: 1-11PubMed Scholar). Collectively moving organisms, however, differ levels peer-to-peer interactions, ranging minimal cooperation complex behaviors (Attanasi et al., 2014Attanasi Cavagna Del Castello L. Giardina I. Melillo S. Parisi Pohl O. Rossaro B. Shen E. Silvestri Viale behaviour without order wild midges.PLoS Comput. 2014; 10: 1-10Crossref (112) 2010Cavagna Cimarelli Santagati Stefanini F. Scale-free correlations starling flocks.Proc. Natl. Acad. U S 2010; 107: 11865-11870Crossref (527) Furthermore, differences among individuals, heterogenic environments, variability interactions between its direct environment are all sources variance affect coordinated behavior collective. Accordingly, it clear how synchronized constitutes such robust phenomenon, maintaining across various sizes densities, heterogeneous unpredictable environmental conditions. One most interesting, albeit disastrous, examples marching These millions, migrating mass large distances, devastating vegetation, agriculture (Ayali, 2019Ayali puzzle locust density-dependent phase polyphenism.Curr. Opin. Insect 35: 41-47Crossref (22) Cullen 2017Cullen D.A. Cease A.J. Latchininsky A.V. Ayali Berry K. Buhl J. De Keyser Foquet Hadrich J.C. Matheson al.From molecules management: consequences polyphenism.Adv. Physiol. 2017; 53: 167-285Crossref (68) Zhang 2019Zhang Lecoq Hunter D. Locust grasshopper management.Annu. Rev. Entomol. 64: 15-34Crossref (72) In context level individuals: collectivity, which based local mostly manifested alignment neighboring overall movement same general direction (e.g., 2014aAriel Ophir Y. Levi Ben-Jacob Individual pause-and-go instrumental nymphs.PLoS One. e101636Crossref (21) Bazazi 2008Bazazi Hale J.J. Anstey M.L. Sword G.A. Simpson S.J. Couzin I.D. cannibalism migratory bands.Curr. 2008; 18: 735-739Abstract Full Text PDF (182) Nonetheless, phenomenon extremely robust, with huge demonstrating moderate high collectivity scales (up 6–7 orders magnitudes), terms number animals spatiotemporal distribution (Ellis Ashall, 1957Ellis P.E. Ashall C. Field studies diurnal behaviour, aggregation desert (Schistocerca gregaria forskǻl).Anti-locust Bull. 1957; 25: 1-103Google Uvarov, 1977Uvarov Grasshoppers Locusts: Handbook General Acridology. Centre Overseas Pest Research, 1977Google see further references Ayali, 2015Ariel modeling.PLoS 2015; 11: 1-25Crossref (54) Thus, exhibit considerable disparity little large-scale collectivity. What maintain integrity? Here, series carefully controlled experiments induces internal switch gregarious locust, activating refer kinematic notably differs non-collective-motion-state. important emphasize “collective-motion-state” “non-collective-motion-state” swarming-gregarious referring well-known solitarious-gregarious transition How, then, does collective-motion-state swarm? Interestingly, into state seems occur rapidly, response walking. particular, indicate alone sufficient. Switching out occurs over longer timescale—significantly than typical timescale normal fluctuations around dynamics. Hence, stochastic fluctuations, (Algar 2019Algar S.D. Stemler Small From flocs flocks.in: Macau E.E.N. Mathematical Modeling Approach Nonlinear Dynamics Complex Systems. Springer, 2019: 157-175Crossref Escaff 2018Escaff Toral Van Den Broeck Lindenberg continuous-time persistent random walk model flocking.Chaos. 2018; 28: 075507Crossref (15) “smoothed-out,” leading highly dynamics behavior, turn beneficial integrity. Using simplified computer model, simulated properties locust-like agents different parameters, representing states. results support functional advantages collective-motion-state, allowing us conclude provides individual-based mechanism stability presence preventing collapsing. main objective report examine upon joining leaving conspecifics. Here studied locusts, reared dense populations, one developmental stage before becoming adults developing wings (i.e., fifth-instar larva). comprised three consecutive stages, conditions (Figure 1 Video S1): (1) Isolation stage: single taken rearing cage, tagged barcode, introduced ring-shaped arena (outer inner diameters: 60 30 cm, respectively). (2) Grouping nine other individually added arena. (3) Re-isolation removed arena, original alone. duration h, enough yet did cause changes due exhaustion, hunger, etc. trajectories fully reconstructed using barcode tracking system. middle 40 min analyzed, detailed transparent methods section. range statistics collected classify compare locusts' stages. https://www.cell.com/cms/asset/5b354008-f03e-46d2-a09b-3e3dad207ac9/mmc2.mp4Loading ... Download .mp4 (1.51 MB) Help files S1. Experimental procedure, related figure To verify grouping indeed inducing (swarming), calculated synchronization grouped parameter (see definition), fundamental estimator Knebel 2019Knebel Guershon Intra- versus intergroup behavior.Sci. Adv. 5: eaav0695Crossref (10) median significantly higher obtained computationally randomized (as presented previous report, medians: 0.632 0.239, respectively; Wilcoxon rank-sum test: p < 0.001). Consequently, experimental setup demonstrated Locusts intermittent pattern (Ariel 2012Bazazi Bartumeus Intermittent locusts: behavioural complexity simple environments.PLoS 2012; 8: e1002498Crossref (58) i.e., sequences alternating bouts pauses. characterize measured four parameters: fraction time spends walking, average speed while bout duration, (4) pause duration. Comparing values several statistically significant 2). following, correspond Friedman test followed multiple comparison Bonferroni method. when comparing isolation grouping, spent durations differed (p 0.01 0.001, respectively), showing larger shorter pauses grouping. accordance reports (Knebel consistent known propensity more rest times Scholar, However, revealed new swarming. re-isolation parameters significantly. Specifically, speed, stage, whereas decreased 0.05, Interpreting together, account low (or make U-turn starting walk; area much second state. increased following 0.01). data combinations exact points trends can Figure activity surprising suggests dictated instantaneous or immediate se. Rather, induce outlasts results, schematically 3. observed represent transient rather permanent modulation, control experiment performed. six experiments, returned cage (in motion) tested next day (isolation condition). no difference latter first day. Hence collective-motion transient. (n = 6) performed exclude potential experiments. end, hours. above-described analysis procedure then separately 40-min segments 3-h tests, found. Despite major interested know whether there any changing conditions: isolation, re-isolation. would although change throughout stages 2), they relative position compared others, thus some tendencies. tendencies generally persisted. within-individual correlation durations, correlated Table numerical details). dependent itself, cannot predicted therefore influenced animal's properties.Table 1Correlation conditionsp valuerho valueFraction walkingIsolation-grouping<0.0010.706Isolation-re-isolation0.0010.679Grouping-re-isolation0.0020.638SpeedIsolation-grouping0.0010.679Isolation-re-isolation0.0100.561Grouping-re-isolation0.0020.636Walking durationIsolation-grouping0.0020.628Isolation-re-isolation0.1480.391Grouping-re-isolation0.3540.314Pause durationIsolation-grouping0.0070.580Isolation-re-isolation0.0030.612Grouping-re-isolation<0.0010.708The averaged bouts, underlined mark correlations. Open table tab undergo sociobehavioral confers benefit maintenance, if so, what kind, remained unanswered. explore aspect developed simulates swarms, could manipulated. simulate used agent-based square domain periodic boundaries. Agents designed rectangles circular receptive field center. agents' programmed resemble (pause-and-go) pattern: every step simulation, agent made decision stop, current (walking stopping) predefined probabilities (pW pS, S2). While moving, constant. allowed stopping An agent's weighted sum own (inertia), visual field, short memory Rimer 2017Rimer Kinetic order-disorder transitions memory.J. Theor. 419: 90-99Crossref Scholar importance simulations), noise. set generate value approximately similar See details S1 values. spatial temporal follows: cm considered distance 0.1 s corresponded simulation step. fixed dimensions real size parameters. (0.1 × 0.4) maintains proportions larva Additionally, radius, 10 rectangle width, represents proportional (range just cm; 0.25, corresponding shown 2B. Finally, 7 7, 12 within. generated slightly lower density better mimicked limited experienced evaluated pW swarm:1The (the vector).2Spread pairs).3The neighbors (within view, denoted NN).4Regrouping steps takes has lost re-obtain least neighbor). 4 shows independent repetitions set. decreases 4B). spread, hand, increasing pausing 4C). NN 3D) accord spread (small implies NN). Its (∼1–1.5), denser, explained earlier. regrouping 4E) dependency durations: shortest longest high. Yet, mid reduction regroup. Worth noting fact because (similar locusts; length turns durations. relate states—isolation, re-isolation—were indicated tiles interquartile empirical Figures 2C 2D. marked point. As seen, calculated, reflects crowd improves coherency rigidity observation possible finds itself scenarios, 75 75, agents, thereby reducing considerably, regroup 3E). density, where practically present S3), reduce Therefore, sparse (even cost decreasing collectivity), shorten until reunites Our reported here that, sensorimotor act accompanied locust—a specific kinematics. experience synchronous, marching. turn, role consistency Next, discuss aspects implications finding. should stressed once study focused gregarious, crowded-reared only. described confused, therefore, researched polyphenism swarming, phase. (crowded-reared) breeding colony, maintained generations. cages, physical constraints abundance food, despite very rarely, all, demonstrate adopted experiencing, in, within recent have (density) contexts. reconfirmed elucidated initial novel idea posited spatially temporally concept arises time: re-isolation, critically (initial isolation). expected, If time, isolated more, loses walking-related previously motion, collective-motion-state. decay explored, likely affected external factors, availability food day-night cycle. retained repo