Collective Turns in Spinless Flocks
Joao Lizárraga, Marcus de Aguiar · 2025
Read Collective Turns in Spinless Flocks by Joao Lizárraga, Marcus de Aguiar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-17804.txt.
Introduction
Joao Lizarraga and Marcus de Aguiar model how a turning signal spreads through a flock when one member starts to turn. In natural flocks the signal moves through the group at a fairly steady speed, letting the whole flock turn together without a leader directing every member.
Using a minimal model without spin like variables, they show that wave like spread comes from interactions that are not reciprocal, meaning members affect each other unevenly. They explain related spectral features with an approximate analysis. The ideas can be tested in natural flocks and extended to other systems with aligned motion.
Catalog record
| CREATOR | Joao Lizárraga, Marcus de Aguiar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-17804
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Collective Turns in Spinless Flocks Joao Lizárraga, Marcus de Aguiar Abstract Using a minimal aggregation-based model, we address the efficient information transfer observed in natural flocks during collective turns. Specifically, we demonstrate that this feature can arise solely from the non-reciprocal nature of local interactions. Through a perturbative analysis, moreover, we find that velocity fluctuations (in the continuum) can be described by a Born approximation. We then show that a wave propagating across the flock undergoes scattering. Our model provides testable predictions and can be extended to study other physical contexts exhibiting polar order. Birds gather and fly together, forming a flock. The early observation of this phenomenon led to the study of such collectives using frameworks based on a velocity alignment mechanism. Indeed, for a long time, the Vicsek model 23 ; 9 ; 18 was considered a pillar in the study of related problems 3 ; 15 ; 10 ; 13 . This status, however, faced challenges as subsequent field observations were carried out. Particularly, studies on starling flocks revealed hallmarks of their collective behavior that were not contemplated by Vicsek’s formulation: strong alignment and cohesion are maintained during both straight flight and turns 2 ; 5 . Moreover, despite the absence of leaders to reach collective coherence, the flocks’ turning events were found to be triggered by single individuals 1 . After recognizing the limitations of the Vicsek model in reproducing the phenomena exhibited by starlings, Cavagna et al. introduced the Inertial Spin Model (ISM) 8 ; 7 . At its core, the ISM stems from the incompleteness of the Vicsek model. It does rely on the velocity-alignment framework; yet, rather than directly promoting the alignment of individual velocities, it does so through internal rotation generators: the spins . As argued by Cavagna et al. , the Vicsek model characterizes an overdamped system: if waves emerge, the system’s nature causes them to dissipate rather than propagate. In contrast, the ISM allows the system to operate in one of two regimes: overdamped or underdamped. Therefore, since real starling flocks allow for the propagation of waves during collective turns, they were hypothesized to match the ISM underdamped scenario. Although the ISM effectively reproduces the turning phenomenon of starling flocks, recent experiments have given rise to an apparent paradox 6 . Under the ISM formulation, if the flock were underdamped, two peaks, representing the spin-waves, should emerge in the frequency domain. However, in real starling flocks, these peaks are not observed. Thus, to reconcile the experimental evidence with the theoretical framework, Cavagna et al. proposed that starlings may operate adaptively depending on the amplitude of fluctuations. That is, for small fluctuations, the system is overdamped, and for larger ones ( turning signals ), the system becomes underdamped. In this letter, we address the propagation phenomenon experienced by turning flocks without explicitly relying on the standard alignment-based framework. Instead, we extend a model of aggregation introduced in 14 . In its simplest form, this model describes a collective of individuals who align their velocities as a by-product of achieving cohesion. Each individual’s self-propelling velocity is constant. Here, in contrast to the form presented in 14 , we consider that individuals are affected by their intention to match neighboring velocities. That is, self-propulsion is adaptive . Thus, a system of N individuals is governed by \displaystyle\begin{split}\dot{\vec{r}}_{i}&=\vec{f}_{i}(t)+\alpha\left(\langle\vec{r}\rangle-\vec{r}_{i}\right),\\ \dot{\vec{f}}_{i}&=\gamma\sum_{j=1}^{N}A_{ij}(\dot{\vec{r}}_{j}-\dot{\vec{r}}_{i}),\end{split} (1) where \vec{r}_{i} and \vec{f}_{i}(t) respectively represent the i -th individual’s position and self-propelling velocity, and \alpha and \gamma are constant coupling strengths. Topological velocity interactions are defined by the adjacency matrix A : the self-propulsion of the i -th individual is affected by those of its \eta closest neighbors in space. Note that this restriction does not imply reciprocity; that is, A may not be symmetric. Indeed, hereafter we consider the general case where A_{ij}\neq A_{ji} . Fig. 1 illustrates the behavior of the system described by Eqs. ( 1 ) during both straight and turning motion. Regardless of its trajectory, the non-reciprocity of interactions is present throughout the entire event [Fig. 1 (b)]. During the straight flight, when the flock’s velocity is roughly constant (fluctuations vanish), the interaction asymmetry index, \psi_{A} , reaches a plateau. Nevertheless, during the collective turn, small variations in \psi_{A} reveal that the network of interactions is also time-varying: A=A(t) . In this state, the acceleration differences among individuals are small [Fig. 1 (c)], thus evidencing that all individuals have effectively received the turning signal (see the Supplemental Material (SM) for details on the turning hierarchy). Figure 1: Collective behavior of a flock over time. (a) Individuals start positioned in the red dots; the black curve describes the trajectory followed by the individual initiating the turn. (b) The asymmetry index, ranging from 0 (perfect asymmetry) to 1 (perfect symmetry), is calculated for the event in (a) via \psi_{A}=\|0.5(A+A^{T})\|_{F}/\|A\|_{F} . The flock follows a sequence of aggregation (blue arrow), straight flight (orange arrow), turning (green arrow), and straight flight again. (c) During the turn, the radial acceleration, a_{i} , is computed for the initiator (dashed line) and random neighbors. The simulation was performed for N=100 individuals, each influenced by \eta=6 neighbors, and coupling strengths (\alpha,\gamma)=(1,10) . Time ( t ) is reported in simulation units ( \Delta\tau=0.01 ; see SM for computation details and additional simulations). Our main interest lies in how information is transferred among individuals. Specifically, we focus on the propagation of velocity fluctuations, \vec{\varphi}_{i}=\dot{\vec{r}}_{i}-\langle\dot{\vec{r}}\rangle , across the flock (which acts as a medium). To this end, we start by rearranging Eqs. ( 1 ) into the fluctuation dynamics: \dot{\vec{\varphi}}_{i}+\alpha\vec{\varphi}_{i}+\gamma\sum_{j=1}^{N}L_{ij}\vec{\varphi}_{j}-\vec{\xi}_{i}=0, (2) where L is the in-degree Laplacian associated with A , and \vec{\xi}_{i} represents a stimulus that remains constant for all the nodes. Since this global term arises from the non-reciprocity of interactions, it can be written in terms of the nodes’ in- and out-degrees as \vec{\xi}_{i}=\gamma\sum_{j}\vec{\varphi}_{j}(d_{j}^{out}-d_{j}^{in}) . To explore the system’s spatiotemporal behavior, we extend Eq. ( 2 ) to the continuum limit: \dot{\vec{\mu}}+\alpha\vec{\mu}+\gamma\mathcal{L}\vec{\mu}-\vec{\xi}=0, (3) where \vec{\mu}(\vec{r},t) is the field of velocity fluctuations and \mathcal{L} is the Laplace operator. In the frequency domain, where our analysis lies, handling a time-dependent operator such as \mathcal{L}=\mathcal{L}(t) may be challenging, complicating the drawing of conclusions. To bypass this issue without compromising much of the system’s nature, we decompose the Laplacian as \mathcal{L}\approx\mathcal{L}_{0}+\sigma\mathcal{L}_{1}(t). Here, \mathcal{L}_{0} and \sigma are constant, and a small \sigma allows us to treat the time-varying component of the interaction operator as a perturbation. The variations exhibited by \psi_{A} [Fig. 1 (b)] provide a qualitative indication on the small time-varying effects of the interactions. However, the validity of the approximation is better justified by the relative magnitudes between the constant and time-varying components (see SM). In the zeroth order ( \vec{\mu}_{0} ), the constant Laplacian is defined as \gamma\mathcal{L}_{0}=\vec{C}_{1}\nabla+\nabla\cdot(\mathbf{C}_{2}\nabla) , where the coefficients \vec{C}_{1} and \mathbf{C}_{2} are constant, and \vec{C}_{1}\nabla and \nabla\cdot(\mathbf{C}_{2}\nabla) respectively represent advection and diffusion terms (see SM for details). Then, by plugging \mathcal{L}_{0} into Eq. ( 3 ), and performing the Fourier transform, we obtain \hat{G}_{0}^{-1}\hat{\mu}_{0}=0 , with \hat{.} denoting that the variable is in the frequency domain, and \hat{G}_{0}^{-1}=\alpha-\vec{k}\cdot\mathbf{C}_{2}\vec{k}+i(\vec{C}_{1}\cdot\vec{k}-\omega). (4) Notice that transforming \vec{\xi} leads to a term proportional to \delta(\vec{k}) , which vanishes as we restrict our study to cases where \vec{k}\neq\vec{0} .