Critical and quasicritical behavior in a three-species dynamical model of semi-directed percolation

This study investigates a one-dimensional three-species dynamical model of semi-directed percolation, demonstrating that it exhibits a directed percolation universality class phase transition and reveals a complex quasi-critical regime with spontaneous activity characterized by two distinct pseudo-thresholds: one maximizing dynamic susceptibility and another governing scale-free spatial and temporal correlations.

The Gist

The Big Picture: A Game of "Catch" with Three Rules

Imagine a long line of people standing shoulder-to-shoulder. Each person can be in one of three states:

1. The Ignorant (State 0): They don't know the secret and can't catch it. They are immune.
2. The Susceptible (State 1): They don't know the secret yet, but they can catch it if someone next to them tells them.
3. The Informed (State 2): They know the secret and can pass it on to their neighbors.

The researchers created a computer simulation of this line to see how a "secret" (or an infection, or a fire) spreads over time. They wanted to understand two things:

1. The Tipping Point: How likely does a person need to be to pass the secret for it to spread forever, rather than dying out?
2. The "Spark": What happens if people occasionally discover the secret on their own, even without a neighbor telling them?

---

Part 1: The Perfect Storm (No Spontaneous Activity)

First, the researchers turned off the "spontaneous discovery" feature. In this scenario, the secret can only spread if an "Informed" person tells a "Susceptible" neighbor.

The Rules of the Game:

• The Spread: If an "Informed" person is next to a "Susceptible" one, the Susceptible one instantly becomes Informed. It's like a wave of gossip moving down the line.
• The Fade: However, people get tired of talking. An "Informed" person might decide to stop talking and become "Ignorant" (immune) again. Also, "Susceptible" people might decide to stop listening and become "Ignorant" too.
• The Control Knob ($p$): The researchers had a dial called $p$. This dial controls how likely a person is to stay in their current state (either keep talking or keep listening) versus changing their mind.

The Discovery:
They found a magic number on the dial ($p \approx 0.632$).

• Below the number: The gossip dies out quickly. The "Informed" people stop talking before they can reach the whole line. The system falls into a "sleeping" state where no one knows the secret.
• Above the number: The gossip spreads forever, reaching everyone. The system stays "active."
• At the number: The system is in a delicate balance. The spread is neither too fast nor too slow; it follows a specific mathematical pattern (a "power law") that physicists call Directed Percolation.

The Metaphor: Think of this like a forest fire. If the trees are too wet (low $p$), the fire dies instantly. If they are dry enough (high $p$), the fire rages out of control. There is a specific humidity level where the fire burns perfectly, creating a fractal, self-similar shape. The researchers proved their model behaves exactly like this famous "forest fire" math.

---

Part 2: The "Spark" (Adding Spontaneous Activity)

Next, they introduced a twist: Spontaneous Activity ($\epsilon$).

Imagine that even if no one is talking to them, a "Susceptible" person might randomly decide to shout the secret just because they had a sudden idea (like a lightning strike starting a forest fire).

The Problem:
In the real world, if you have a constant spark of lightning, the fire never truly dies out. The system never reaches a "sleeping" state because the spark keeps reigniting the fire. This means the sharp "tipping point" disappears. You can't have a phase transition if the fire never stops.

The Surprise Discovery (Quasi-Criticality):
Even though the fire never stops, the researchers found something fascinating. As they turned up the "spark" dial, the system didn't just behave randomly. It showed Quasi-Critical Behavior.

1. The "Sweet Spot" (Susceptibility Peak): They measured how sensitive the system was to changes. They found a specific setting for the "stay-talking" dial ($p$) where the system was most responsive. It was like finding the perfect volume knob where the music sounds the clearest. This is called a Pseudo-Threshold.
2. The "Two-Threshold" Mystery: Here is the most interesting part.
   • They found one spot where the system was most responsive (the peak of the sensitivity curve).
   • They found a different spot where the connections between people (correlations) followed the perfect mathematical "fractal" patterns seen in critical systems.

The Metaphor: Imagine tuning a radio.

• Threshold A: The point where the signal is loudest (maximum volume).
• Threshold B: The point where the music sounds the most "crystal clear" and balanced (perfect frequency).
• In this model, when you add the "spontaneous spark," the point of maximum volume and the point of perfect clarity are no longer the same station! They drift apart.

---

Why Does This Matter?

1. Biology and Brains: This model helps scientists understand how neurons in the brain fire. Neurons sometimes fire spontaneously (like the "spark"). This research suggests that the brain might operate in a "quasi-critical" state, balancing between chaos and order, but with different "sweet spots" for reaction speed versus long-term coordination.
2. Forest Fires: It helps model how fires spread when lightning strikes randomly, not just from one burning tree to another.
3. Mathematical Insight: It proves that even when a system is "broken" (by having a constant spark that prevents a true phase transition), it still retains the beautiful, hidden mathematical structures of critical systems, just shifted slightly.

Summary in One Sentence

The researchers built a digital game of "gossip" to show that while a constant background noise (spontaneous activity) prevents a system from truly "sleeping," it creates a new, complex state where the system's reaction speed and its internal connections peak at two different settings, offering a new way to understand how complex systems like brains or ecosystems stay alive and active.

Technical Summary

1. Problem Statement

The paper investigates the Semi-Directed Percolation (SDP) problem, a variant of percolation theory that occupies an intermediate position between Isotropic Percolation (IP) and fully Directed Percolation (DP).

• Context: While IP models describe static geometric connectivity and DP models describe non-equilibrium dynamic processes with absorbing states, SDP models exhibit directionality in one dimension (typically time) while remaining isotropic in others.
• Gap: Previous studies on SDP relied on transfer matrix and phenomenological renormalization techniques. There was a need to formulate SDP as a one-dimensional dynamical model to study its active-absorbing phase transition and critical exponents directly through simulation.
• Secondary Problem: The authors also investigate the effect of spontaneous activity generation (an external noise term) on the SDP model. In standard absorbing state models, spontaneous activity destroys the true phase transition, but it is hypothesized to induce "quasicritical" behavior (pseudo-thresholds) similar to observations in neuronal and forest-fire models.

2. Methodology

The authors propose and simulate a one-dimensional three-species dynamical model on a lattice of size $L$ with periodic boundary conditions.

Model Definition

Each site $i$ can be in one of three states:

• 0 (Immune/Inactive): Cannot be activated.
• 1 (Susceptible): Can be activated.
• 2 (Active): Can activate neighbors.

The dynamics evolve in discrete time steps via two processes:

1. Process P1 (Spontaneous/Probabilistic Transitions):
   • $0 \to 1$ with probability $p$.
   • $1 \to 0$ with probability $(1-p)$.
   • $2 \to 0$ with probability $(1-p)$ (staying active with prob $p$).
   • Spontaneous Activity: A cluster of 1s can spontaneously become 2 with probability $\epsilon$.
2. Process P2 (Deterministic Spreading):
   • If a cluster of 1s is in contact with an active site (2), the entire contiguous cluster of 1s instantly becomes 2 with probability 1.

Key Distinction: The model requires three species because the "immune" state (0) is necessary to define the boundaries of the susceptible clusters, thereby generating the semi-directed geometry in time.

Simulation Techniques

• Monte Carlo Simulations: Performed on lattices up to $L = 4096$ with $10^3$ independent trials.
• Initial Conditions:
  • Fully Active: All sites start as state 2.
  • Single Seed: One site starts as state 2, others as 0.
• Analysis:
  • Order Parameters: Density of active sites ($\rho$) and survival probability ($P_s$).
  • Critical Exponents: Determined via power-law fits of $\rho(t)$, $P_s(t)$, and the number of active sites $N(t)$.
  • Correlations: Spatial ($g_\perp$) and temporal ($g_\parallel$) correlation functions.
  • Dynamic Susceptibility ($\chi$): Calculated as the variance of the activity density.

3. Key Contributions

1. Dynamical Formulation of SDP: Successfully mapped the static SDP problem to a 1D three-species dynamical process, proving that the time evolution of this system naturally generates SDP clusters.
2. Universality Class Confirmation: Numerically verified that the dynamical SDP model belongs to the Directed Percolation (DP) universality class, confirming earlier theoretical predictions.
3. Discovery of Dual Pseudo-Thresholds: In the presence of spontaneous activity ($\epsilon > 0$), the authors identified a novel phenomenon where the system exhibits two distinct pseudo-thresholds:
   • One where the dynamic susceptibility ($\chi$) is maximized (maximum response).
   • Another where spatial and temporal correlations exhibit scale-free (power-law) behavior.
   • These two points coincide only at low $\epsilon$ but diverge as $\epsilon$ increases.

4. Key Results

A. Critical Behavior ($\epsilon = 0$)

• Critical Threshold: The phase transition occurs at $p_c = 0.6320(2)$. This aligns closely with previous theoretical values ($0.6317$ to $0.631985$).
• Critical Exponents: The calculated exponents match the (1+1)-dimensional DP universality class:
  • Decay exponents: $\alpha \approx 0.16$, $\delta \approx 0.16$.
  • Growth exponent: $\theta \approx 0.31$.
  • Order parameter exponents: $\beta = \beta' \approx 0.276$.
  • Correlation ratios: $\beta/\nu_\perp \approx 0.252$, $\beta/\nu_\parallel \approx 0.159$.
• Scaling: Data collapse was achieved for density, survival probability, and correlation functions using standard DP scaling relations, confirming the universality class.

B. Quasicritical Behavior ($\epsilon > 0$)

• Widom Line: The presence of spontaneous activity ($\epsilon$) eliminates the true absorbing state (activity never dies out). However, the dynamic susceptibility $\chi$ shows a distinct peak at a specific $p$ (pseudo-threshold), defining a non-equilibrium Widom line.
• Peak Broadening: As $\epsilon$ increases, the susceptibility peak broadens and decreases in height, indicating a transition from sharp criticality to a broader critical-like region.
• Divergence of Thresholds:
  • At low $\epsilon$, the $p$ value for the susceptibility peak and the $p$ value for power-law correlations are nearly identical.
  • At higher $\epsilon$ (e.g., $\epsilon = 0.1$), the $p$ value where correlations show power-law decay ($p \approx 0.611$) is lower than the $p$ value where susceptibility peaks ($p \approx 0.617$).
• Exponent Variation: The slope of the power-law correlations ($\beta/\nu$) increases with $\epsilon$ before saturating at higher values.

5. Significance and Implications

• Theoretical Validation: The study provides robust numerical evidence that SDP, despite its mixed geometry, falls strictly within the DP universality class, reinforcing the robustness of DP as the generic class for absorbing state transitions.
• Insight into Biological Systems: The findings on "quasicriticality" and the existence of dual thresholds are highly relevant for modeling biological systems (e.g., neuronal networks, cardiac tissue) where spontaneous activity (noise) is inherent. The separation of the "response maximum" from the "correlation scale-free point" suggests that biological systems operating near criticality may exhibit complex, multi-faceted critical behaviors depending on the metric used.
• Methodological Advancement: The proposed 3-species dynamical framework offers a new, computationally efficient way to study semi-directed percolation and can be adapted to study other variants of percolation with spontaneous noise.

In conclusion, the paper successfully bridges static percolation theory and non-equilibrium dynamics, revealing that the introduction of spontaneous activity in SDP models creates a rich landscape of quasicritical behavior characterized by the splitting of critical indicators into distinct pseudo-thresholds.