Collective drift and pinning in active rotator networks with Kuramoto coupling and mixed-sign feedback disorder

This paper investigates fully connected active rotator networks with Kuramoto coupling and mixed-sign Gaussian feedback disorder, revealing how the competition between local pinning and collective alignment shapes drift regimes and can be mapped through numerical simulations and analytical limits.

The Gist

Imagine a large group of people standing in a circle, each holding a spinning top. In the world of physics, these spinning tops are called "active rotators," and they represent things that naturally want to move, like heart cells beating or fireflies flashing.

This paper explores what happens when you try to get all these spinning tops to move together in a synchronized dance, but you introduce two competing forces: a common push and random local tugs.

Here is the story of the paper, broken down into simple concepts:

1. The Setup: A Common Push vs. Random Tugs

Imagine every person in the circle is trying to spin their top forward at the same speed. This is the "common intrinsic drive." It's like a gentle, uniform wind blowing on everyone, trying to make them all spin forward together.

However, each person also has a "local feedback" mechanism. Think of this as a personal brake or accelerator attached to their specific top.

• Some people have a brake that slows them down.
• Some have an accelerator that speeds them up.
• Crucially, these brakes and accelerators are random. Some are strong, some are weak, and they are equally likely to be "brakes" (negative) or "accelerators" (positive). This is the "mixed-sign feedback disorder."

The paper asks: Can the group still spin together, or will the random brakes and accelerators stop them?

2. The Two Main Forces at Play

The researchers studied how two factors compete:

• The Local Tugs (Pinning): If a person's local brake is strong enough, it can stop their top completely, "pinning" them in place. If too many people get pinned, the whole group stops.
• The Group Hug (Coupling): The people are holding hands (mathematically, this is "Kuramoto coupling"). If they hold hands tightly, they try to pull each other into sync. If one person is stuck, the group might pull them free. If one person is spinning fast, the group might slow them down to match.

3. What Happens When You Change the Rules?

The authors created a "map" of what happens when you change the strength of the random brakes (disorder) and the strength of the hand-holding (coupling).

• Scenario A: Weak Hand-Holding, Strong Random Brakes
  If the people aren't holding hands very tightly, but everyone has strong, random brakes, the group falls apart. Many people get stuck (pinned) because their local brakes are too strong for their own momentum to overcome. The group drifts very slowly or stops.

• Scenario B: Strong Hand-Holding, Weak Random Brakes
  If the people hold hands very tightly, but their local brakes are weak, the "common push" wins. The group ignores the small, random tugs and spins forward together in a synchronized rhythm. The hand-holding pulls everyone into a collective drift.

• Scenario C: Strong Hand-Holding, Strong Random Brakes
  This is the most interesting part. If the brakes are very strong and the hand-holding is very strong, the group doesn't necessarily spin faster. Instead, the strong brakes pin so many people that even the strong hand-holding can't get them moving. The whole system becomes "stuck" in a stationary state. The collective effort isn't enough to overcome the sheer number of local anchors.

4. The "Backward" Drift

The paper also noticed something surprising. Even though the "wind" (the common drive) is blowing everyone forward, some people in the group actually spin backward.
This happens when a person has a very strong local brake (or accelerator in the wrong direction) and the group dynamics pull them the other way. It's like a swimmer trying to swim upstream; if the current is strong enough, they might get pushed backward despite their own effort. This "backward drift" only happens in specific zones where the random brakes are just strong enough to fight the wind, but not strong enough to stop the person completely.

5. What If Everyone Had Different Natural Speeds?

The authors also tested a variation where, instead of everyone having the same "wind" pushing them forward, everyone had a different natural speed (some fast, some slow, some backward).
In this case, even if they hold hands tightly, they don't start spinning forward together. Instead, they tend to cancel each other out and stop moving entirely. This highlights that the "common push" in the main experiment was essential for getting the group to drift in one direction.

The Big Takeaway

The main discovery of this paper is that randomness in how individuals react to their environment can completely change how a group moves.

Even if everyone is being pushed in the same direction, if their local "brakes" are random and mixed (some strong, some weak, some forward, some backward), the group can end up:

1. Spinning freely together.
2. Getting stuck in place.
3. Or even having some members spin backward.

The paper shows that you don't need everyone to be different to get complex behavior; you just need the local forces acting on them to be random and mixed. It's a study of how a crowd of individuals balances between being stuck in their own spot and moving together as a single unit.

Technical Summary

Problem Statement

The paper investigates the competition between local pinning and collective phase alignment in a network of fully connected active rotators. While the classical Kuramoto model addresses synchronization in the presence of intrinsic frequency heterogeneity, and active rotator models typically study the transition between excitable and oscillatory states, this work introduces a distinct source of disorder: mixed-sign local feedback amplitudes.

The system consists of $N$ oscillators sharing a common, positive intrinsic drive ($\omega > 0$). However, each oscillator is subject to a local feedback term with an amplitude $A_i$ drawn from a zero-mean Gaussian distribution. This creates a scenario where the intrinsic drive pushes all oscillators in the same direction, but the local feedback can either support or oppose this drive depending on the sign of $A_i$. The central problem is to understand how this specific type of disorder—randomness in the pinning mechanism rather than the drive itself—interacts with Kuramoto-type coupling to determine whether the system exhibits collective drift or becomes collectively pinned.

Methodology

The authors employ a combination of analytical limits and extensive numerical simulations to map the system's behavior.

1. Model Definition:
The dynamics of the $i$-th oscillator are governed by:
$$\frac{d\theta_i}{dt} = \omega - A_i \sin \theta_i + \frac{K}{N-1} \sum_{j \neq i} \sin(\theta_j - \theta_i)$$
where $\omega$ is the common drive, $K$ is the global coupling strength, and $A_i \sim \mathcal{N}(0, \sigma_A^2)$. The system is analyzed using dimensionless parameters: $\kappa = K/\omega$ (coupling), $r = \sigma_A/\omega$ (feedback disorder strength), and $a_i = A_i/\omega$.

2. Key Observables:

• Mean Absolute Late-Time Drift ($D$): Defined as the average of the absolute late-time angular velocities of all oscillators. The absolute value is crucial because oscillators may drift in opposite directions; a signed average could mask persistent motion through cancellation.
• Drifting Fractions ($f_+$ and $f_-$): The fractions of oscillators exhibiting positive or negative late-time drift, respectively, used to distinguish the direction of motion.

3. Analytical Approaches:

• Uncoupled Limit ($\kappa=0$): The authors derive an exact analytical expression for the drift by integrating the single-oscillator drift velocity over the Gaussian distribution of $a_i$, considering only the regime where $|a_i| < 1$ (where no fixed points exist).
• Strong-Coupling Limit: A coherent-phase approximation is used where all phases cluster around a common center $\Theta$. This leads to a scaling estimate relating the coupling strength $\kappa$ and disorder strength $r$ required for collective pinning.
• Necessary Condition: A condition for collective pinning is derived by averaging the equations of motion, showing that a non-trivial correlation between local feedback amplitudes and phases is required.

4. Numerical Simulations:
Regime maps are constructed in the $(r, \kappa)$ plane for system sizes up to $N=200$. Simulations utilize Euler integration to track late-time velocities. The study also examines finite-size effects and a variant model where the intrinsic drive is replaced by zero-mean distributed intrinsic frequencies.

Key Results

1. Competition Between Pinning and Drift:
The numerical regime maps reveal a clear competition:

• Weak Coupling: Increasing the feedback disorder strength ($r$) suppresses the mean absolute drift. As $r$ increases, a larger fraction of oscillators satisfy $|a_i| > 1$, causing them to become locally pinned (stuck at fixed points).
• Strong Coupling: When coupling is strong, it can restore collective drift even in the presence of moderate disorder. The coupling aligns the phases, allowing the common intrinsic drive to overcome the local pinning forces, provided the disorder is not too strong.
• Suppression at High Disorder: In the limit of very large $r$ and large $\kappa$, the system enters a regime of near-complete suppression of drift, where the collective state becomes effectively stationary.

2. Direction of Motion:
While the intrinsic drive is positive, negative drift is observed in specific regimes. Negative drift occurs primarily at intermediate coupling and moderate-to-large disorder, where local feedback and coupling terms can collectively overcome the positive drive for a subset of oscillators without the entire system becoming pinned.

3. Analytical Validation:

• Uncoupled Limit: The analytical prediction for $D_0(r)$ matches numerical simulations perfectly. It shows that for weak disorder ($r \ll 1$), drift is near maximal ($D \approx 1$), while for strong disorder ($r \gg 1$), drift scales as $1/r$ due to the increasing fraction of pinned oscillators.
• Strong Coupling: Under the coherent-phase approximation, the boundary between low-drift and restored drift scales quadratically: $\kappa \sim r^2$. Numerical results with coherent initial conditions confirm this parabolic scaling, identifying the region below the curve as the pinned regime.

4. Finite-Size Effects:
The qualitative behavior is robust across system sizes ($N=20$ to $200$). Weak-coupling drift suppression is largely independent of $N$. However, the recovery of drift at strong coupling shows a slight dependence on system size, with larger systems exhibiting slightly higher drift levels in the intermediate disorder regime.

5. Zero-Mean Intrinsic Frequencies:
In a variant model where the common drive is replaced by zero-mean distributed intrinsic frequencies, strong coupling does not restore a directed collective drift. Instead, because the intrinsic frequency distribution is centered at zero, strong coupling suppresses motion entirely, leading to a near-stationary collective state. This highlights the distinct role of the fixed positive drive in the primary model.

Significance and Claims

The paper claims that mixed-sign local feedback disorder alone is sufficient to reshape the balance between pinning and drifting in coupled active rotator networks, even when the intrinsic drive is homogeneous.

• Distinction from Frequency Disorder: The work emphasizes that disorder in the local pinning mechanism (feedback amplitude) is fundamentally different from disorder in intrinsic frequencies. The former creates a competition where local terms can oppose the global drive, whereas the latter typically leads to incoherence or synchronization depending on the frequency spread.
• Mechanism of Restoration: The study demonstrates that coupling can act as a mechanism to "restore" drift that would otherwise be suppressed by local pinning, provided the disorder strength is not excessive.
• Minimal Framework: The authors present this model as a minimal framework for studying how local pinning, feedback disorder, and collective alignment compete. They suggest that these results are relevant for understanding excitable and oscillatory systems in physical and biological contexts where local heterogeneity in feedback mechanisms exists.

The paper concludes modestly, noting that the results are specific to fully connected networks and zero-mean Gaussian disorder, and suggests future work on random network topologies, non-Gaussian distributions, and more complex correlations between parameters.