Chimera States

1. Kuramoto and Battogtokh, 2002

The Kuramoto model describes populations of coupled phase oscillators and how they transition to full synchrony as coupling strength increases. Kuramoto's original model used global (all-to-all) coupling: every oscillator feels every other oscillator equally.

In 2002, Kuramoto and Battogtokh studied a ring of identical phase oscillators with non-local coupling — each oscillator is influenced mainly by its neighbours, with influence decaying smoothly with distance around the ring. They expected either full synchrony or full incoherence. Numerically, they instead found stable coexistence: roughly half the ring locked into a coherent, synchronized cluster, and the other half remained incoherent and effectively chaotic — indefinitely, with a sharp, persistent boundary between the two regions.

2. Non-local coupling: the model

The Kuramoto-Battogtokh (KB) model places N identical phase oscillators θᵢ on a ring, each coupled to all others through a kernel G that decays with distance:

Where:

• ω — natural frequency, identical for every oscillator
• G(x) — coupling kernel, e.g. G(x) ∝ exp(−κ|x|), decaying with distance around the ring
• α — phase-lag parameter; a nonzero α is essential for chimeras to appear at all

The critical ingredient is the phase-lag α (sometimes written as π/2 − α). Without it, the model reduces to standard attractive coupling, which always synchronizes fully. With α close to but below π/2, the coupling becomes almost (but not quite) purely dissipative — this is exactly the regime where chimeras appear.

3. Why is this strange?

Chimera states violate the intuition that identical elements with identical coupling should behave identically. There is no external field, no heterogeneity in frequencies, no special initial condition planted at a particular oscillator — the equations are perfectly symmetric under relabelling and rotation of the ring. Yet the stable solution spontaneously breaks that symmetry, splitting the ring into a coherent domain and an incoherent domain.

This is an example of spontaneous symmetry breaking in a dynamical system: the governing equations have more symmetry than their stable solutions do. The location of the coherent cluster is itself arbitrary — it can sit anywhere on the ring depending on initial conditions, but once formed, the split persists indefinitely.

4. Mechanism: the mean-field self-consistency argument

In the continuum limit (N → ∞), Kuramoto and Battogtokh reduced the problem to a self-consistency equation for a complex local order parameter Z(x,t) = R(x,t)e^(iΨ(x,t)), where R measures local phase coherence at position x on the ring.

In the coherent region, R ≈ 1 — the oscillators there rotate at a common effective frequency, all locked to each other. In the incoherent region, R < 1 and the local oscillators drift at frequencies spread around the mean, never locking. The self-consistency equation admits a stable solution in which R(x) has exactly this two-region profile, with a smooth but rapid transition zone between them — matching what was seen numerically.

Crucially, the incoherent oscillators are not simply "unsynchronized" — they experience an effective mean field generated by the coherent cluster, and this mean field is strong enough to prevent them from drifting into full incoherence but too weak to lock them into the coherent group. They sit in a genuinely intermediate, partially-locked regime.

5. Laboratory confirmations

Chimera states were purely theoretical until 2012, when two independent groups produced experimental evidence:

• Hagerstrom et al. (2012) — a ring of coupled optoelectronic (light-modulated) oscillators demonstrating stable chimeras.
• Tinsley, Nkomo & Showalter (2012) — a population of chemical Belousov-Zhabotinsky oscillators (see also our article on Belousov-Zhabotinsky reaction-diffusion), coupled through a camera-and-light feedback loop, showing coexisting synchronized and desynchronized subgroups.

Since then, chimeras have been observed in mechanical metronomes, chemical oscillators, laser arrays and simulated neural networks — confirming the phenomenon is not an artefact of one particular model but a robust feature of coupled oscillator systems more generally.

6. Relevance: unihemispheric sleep and beyond

Chimera states offer a compelling mathematical metaphor for unihemispheric slow-wave sleep, observed in dolphins, some birds, and seals: one brain hemisphere shows synchronized slow-wave (sleep) activity while the other remains active and desynchronized (awake), allowing the animal to keep one eye open for predators while resting the other half of its brain. While the biological mechanism is more complex than the idealized KB ring, the qualitative picture — identical neural populations splitting into a coherent and an incoherent domain — matches the chimera phenomenon closely.

Chimera-like dynamics have also been proposed as models for epileptic seizure onset (localized hypersynchrony against a background of irregular activity) and for power-grid desynchronization events, where part of a network of coupled generators falls out of step with the rest.

7. Pseudocode

// N identical phase oscillators on a ring, non-local coupling
const N = 200
const omega = 1.0       // identical natural frequency
const K = 4            // coupling strength
const alpha = 1.45     // phase lag, close to PI/2
const kappa = 4        // kernel decay rate

// Exponential coupling kernel, periodic on the ring
function kernel(dx):
  return (kappa/2) * Math.exp(-kappa * Math.abs(dx))

theta = randomArray(N)  // near-identical random initial phases

loop (each step):
  dtheta = new Array(N)
  for i = 0 to N-1:
    sum = 0
    for j = 0 to N-1:
      dx = ringDistance(i, j, N)
      sum += kernel(dx) * Math.sin(theta[i] - theta[j] + alpha)
    dtheta[i] = omega - (K/N) * sum
  for i = 0 to N-1:
    theta[i] += dtheta[i] * dt

  renderPhases(theta)  // coherent arc vs incoherent arc emerges