Effect of hidden geometry and higher-order interactions on the synchronization and hysteresis behaviour of phase oscillators on 5-cliques simplicial assemblies

Imagine a giant dance floor where thousands of dancers (oscillators) are trying to move in perfect unison. Some dancers are naturally fast, some are slow, and they are all connected to each other by invisible strings. This is the classic "Kuramoto model" of synchronization, often used to explain how fireflies flash together or how heart cells beat as one.

But this paper asks a much more complex question: What happens if the dance floor isn't just a flat grid of connections, but a twisted, multi-dimensional shape made of triangles, tetrahedrons, and even higher-dimensional shapes? And what happens if the dancers are connected not just by holding hands (pairwise), but by forming tight little groups of three or more (higher-order interactions)?

Here is the story of the paper, broken down into simple concepts and analogies.

1. The Dance Floor: Building with "5-Clubs"

The researchers built three different types of dance floors using a specific building block: a 5-clique. Think of a 5-clique as a tight-knit group of 5 friends who all know each other perfectly.

They grew these dance floors by sticking new groups of 5 friends onto the existing structure. But how they stuck them together depended on a "chemical affinity" knob (parameter $\nu$ ):

The Compact Club (High Affinity): Imagine the new groups are glued together by their entire faces. They share 4 people with the existing group. The result is a dense, compact ball where everyone is connected to almost everyone else. It's like a crowded elevator where you can't move without touching someone.
The Sparse Web (Low Affinity): Imagine the new groups only share one single person with the existing group. The result is a sparse, sprawling web of isolated clusters connected by single bridges. It's like a chain of houses connected only by a single porch.
The Mixed Crowd: A random mix of both styles.

The Hidden Geometry: Even though these structures look different, they all share a "hidden geometry" (they are all 1-hyperbolic), but they have different "spectral dimensions." Think of spectral dimension as the effective size of the dance floor. The compact one feels huge and connected; the sparse one feels small and isolated.

2. The Rules of the Dance: Two Types of Connections

The dancers follow two types of rules:

Pairwise (The Handshake): $K_1$ . This is the standard rule: "If my neighbor is dancing fast, I speed up." This can be positive (encouraging) or negative (discouraging).
Triangle/Group (The Huddle): $K_2$ . This is a higher-order rule: "If my two neighbors are dancing together, I join their specific rhythm." This represents the "group pressure" of a clique.

3. The Experiment: The Hysteresis Loop

The researchers played a game of "push and pull." They slowly turned up the "encouragement" (positive $K_1$ ) to see if the dancers would sync up, and then slowly turned it down into "discouragement" (negative $K_1$ ) to see if they would fall apart.

The Surprise: The "Hysteresis" Effect
In a normal system, if you push a swing forward and then pull it back, it follows the same path. But here, the path didn't match.

Going Forward: It took a lot of effort to get everyone to dance together.
Going Backward: Once they were dancing together, they stayed together even when the encouragement was turned off or made negative. They got "stuck" in a synchronized state.

This is called hysteresis (like a magnet that stays magnetic even after you remove the magnet).

4. The Big Discovery: The "Ghost" Groups

The most fascinating part happened when the researchers turned the encouragement negative ( $K_1 < 0$ ). Usually, negative pressure causes chaos. But here, they found partial synchronization.

In the Compact Club: Even with negative pressure, the dancers didn't stop. They formed small, tight-knit groups that danced in sync with each other, but these groups were out of sync with the rest of the floor. It was like a crowded room where small circles of friends are chatting happily, but the whole room isn't listening to the same song.
In the Sparse Web: The dancers barely synchronized at all. They were too far apart to form these stable groups.

Why does this matter?
The paper shows that the shape of the network (how the 5-cliques are glued together) determines whether these "ghost groups" can form.

If the groups share a large face (many people), they can maintain a local rhythm even when the whole system is trying to break apart.
If they only share one person, the local rhythm collapses easily.

5. The "Chaos" of the Rhythm

When the system is in this "partially synchronized" state (where small groups are dancing but the whole room isn't), the global rhythm isn't steady. It wobbles.

The researchers used a mathematical tool called Multifractal Analysis to study these wobbles.

Analogy: Imagine listening to a drumbeat that isn't just a simple "boom-bap." It's a complex, rhythmic pattern that has small beats inside big beats, and tiny ripples inside those.
The paper found that the "wobble" of the dance floor has a fractal nature. It's not random noise; it's a structured, complex pattern. This suggests that the system is in a delicate, critical state where order and chaos are fighting for control.

6. The Real-World Takeaway

Why should you care about dancing 5-cliques?

The Brain: Our brains are full of these complex, higher-order connections (neurons forming cliques). This research suggests that the brain might use these "partial synchronization" states to process information. We don't need everything to fire at once (which would be a seizure); we need small, synchronized groups to handle different tasks simultaneously.
Resilience: The study shows that complex structures are surprisingly resilient. Even when you try to break the system (negative coupling), the internal geometry holds it together in pockets of order.
Design: If you want to build a network (like a power grid or a social media platform) that stays stable even under stress, you need to design the "glue" between your groups carefully. Sharing too little (sparse) makes it fragile; sharing too much (compact) might make it hard to break out of a bad state.

Summary

This paper is a study of how the shape of a network changes the way it behaves. It reveals that when you add complex, multi-dimensional shapes to a network, you get a "memory" effect (hysteresis) and the ability to form stable, local islands of order even when the whole system is being pushed apart. It's a beautiful demonstration of how geometry dictates destiny in complex systems.

Here is a detailed technical summary of the paper "Effect of hidden geometry and higher-order interactions on the synchronization and hysteresis behaviour of phase oscillators on 5-cliques simplicial assemblies."

1. Problem Statement

The paper investigates how the hidden geometry of simplicial complexes and higher-order interactions (beyond pairwise) influence the collective dynamics of phase oscillators. Specifically, it addresses:

How different architectural arrangements of high-dimensional simplicial complexes (built from 5-cliques) affect synchronization and desynchronization.
The role of hysteresis in synchronization transitions when varying pairwise coupling strengths ( $K_1$ ) in the presence of triangle-based (3-body) couplings ( $K_2$ ).
The nature of partial synchronization states, particularly those arising from negative pairwise couplings, and how they are modulated by internal frequency distributions and network topology.
The characterization of fluctuations in partially synchronized states using multifractal analysis.

2. Methodology

A. Network Construction (Simplicial Assemblies)

The authors generate 4-dimensional simplicial complexes using a cooperative self-assembly model where 5-cliques (complete graphs of 5 nodes) serve as building blocks. The growth is controlled by a chemical affinity parameter ( $\nu$ ):

$\nu = +5$ (Compact): Cliques share maximal faces (tetrahedra), resulting in a dense, compact structure.
$\nu = 0$ (Mixed): Faces of any size are shared based on geometric compatibility.
$\nu = -5$ (Sparse): Cliques share minimal faces (single nodes), resulting in a sparse, tree-like structure.
Key Structural Metrics: All three structures are 1-hyperbolic but possess distinct spectral dimensions ( $d_s$ ): $d_s \approx 1.57$ (sparse), $2.11 $(mixed), and$ 4.01$ (compact).

B. Dynamical Model

The system consists of $N \approx 1000$ Kuramoto oscillators attached to the nodes of the simplicial complexes. The phase evolution $\dot{\theta}_i$ is governed by:
$\dot{\theta}_i = \omega_i + \frac{K_1}{k_i^{(1)}} \sum_{j} A_{ij} \sin(\theta_j - \theta_i) + \frac{K_2}{2k_i^{(2)}} \sum_{j,l} B_{ijl} \sin(\theta_j + \theta_l - 2\theta_i)$

$K_1$ : Pairwise coupling strength (1-simplex), varied from negative to positive values.
$K_2$ : Triangle-based coupling strength (2-simplex), fixed at various values.
$\omega_i$ : Intrinsic frequencies, drawn from either a uniform distribution ( $\delta$ -function) or a Gaussian distribution with width $\sigma$ .

C. Simulation Protocol

Hysteresis Sweep: The system is simulated by adiabatically sweeping $K_1$ from negative to positive (forward sweep) and back (backward sweep) in steps of $0.1$.
Order Parameter: Global synchronization is measured by the Kuramoto order parameter $r = |\frac{1}{N}\sum e^{i\theta_j}|$ .
Analysis: The study analyzes the hysteresis loops, individual phase evolution patterns, and the temporal fluctuations of $r$ using Detrended Multifractal Analysis (MF-DFA) to calculate singularity spectra.

3. Key Contributions and Results

A. Impact of Architecture on Hysteresis and Partial Synchronization

Negative Coupling ( $K_1 < 0$ ):
- Compact Network ( $\nu=5$ ): Exhibits significant partial synchronization ( $r \approx 0.5$ ) even with negative pairwise coupling and no higher-order interactions ( $K_2=0$ ). This is attributed to the large shared faces (tetrahedra) between cliques.
- Sparse/Mixed Networks ( $\nu=-5, 0$ ): Show negligible partial synchronization ( $r \approx 0.03$ ) under the same conditions.
- Conclusion: The level of partial synchronization at negative $K_1$ is determined by the minimal size of shared faces between neighboring cliques, largely independent of $K_2$ or frequency distribution.
Desynchronization and Hysteresis Loops:
- In compact networks, the hysteresis loop closes (desynchronization occurs) only when a sufficiently strong $K_2$ is applied.
- In sparse networks, desynchronization occurs abruptly even at $K_2=0$ .
- Higher-order interactions ( $K_2 > 0$ ) generally induce sudden desynchronization (abrupt drops in $r$ ) and broaden the hysteresis loop, delaying the transition to full synchronization.

B. Effect of Internal Frequency Distribution

Homogeneous Frequencies ( $\sigma \to 0$ ): The hysteresis loops are well-defined.
Distributed Frequencies ( $\sigma > 0$ ):
- Broadening the frequency distribution ( $\sigma$ ) reduces the overall synchronization level.
- Instability: In sparse and mixed networks, increasing $\sigma$ leads to instabilities where the order parameter fluctuates significantly with small changes in $K_1$ , preventing a smooth transition to full synchronization even at high positive $K_1$ .
- This instability is absent in the compact network, highlighting the stabilizing role of high spectral dimension and dense connectivity.

C. Multifractal Analysis of Fluctuations

In partially synchronized states (especially in sparse networks), the order parameter $r(t)$ exhibits quasi-cyclical fluctuations.
Multifractal Spectrum: Analysis reveals that these fluctuations are multifractal, characterized by broad singularity spectra ( $\Psi(\alpha)$ ) and non-trivial Hurst exponents ( $H_q$ ).
This indicates that the dynamics are not simple periodic oscillations but involve complex, scale-invariant temporal correlations arising from the co-evolution of small, roughly synchronized clusters.

4. Significance and Implications

Geometric Frustration: The paper demonstrates that simplicial geometry induces "geometric frustration" in synchronization, leading to stable partial synchronization states that are robust against parameter changes.
Beyond Spectral Dimension: While spectral dimension ( $d_s$ ) is a known predictor for global synchronization (requiring $d_s \ge 4$ for stability), this work shows that local architectural features (specifically the size of shared faces) are critical for understanding partial synchronization and hysteresis.
Higher-Order Interactions: The study clarifies that triangle-based interactions do not merely enhance synchronization; they can induce abrupt switching and complex desynchronization pathways, acting as a control mechanism for network dynamics.
Real-World Applications: The findings are relevant for understanding complex systems where higher-order interactions are prevalent, such as:
- Neuroscience: Brain networks where partial synchronization is functional (e.g., memory formation) and full synchronization is pathological (e.g., epilepsy).
- Social Systems: Emergence of consensus or polarization in groups with complex group interactions.
- Materials Science: Cooperative self-assembly of nanoparticles.

Conclusion

The authors establish that the architecture of simplicial complexes (specifically the size of shared faces between cliques) and the strength of higher-order interactions are decisive factors in determining the synchronization landscape. They reveal a rich phase space where negative couplings can sustain partial order, and where multifractal fluctuations characterize the transition between synchronized and desynchronized states, offering a deeper understanding of collective dynamics in high-dimensional networks.