Metastability and Eye Movements: A Dynamical‑Systems Interpretation

-
  1. Why metastability is a natural language for gaze behavior The idea that perception, attention, and neural activity evolve through transiently stable states has deep roots in cognitive science and neuroscience. Concepts such as attractors , basins , noise‑induced transitions , and escape times appear in work on perceptual switching, decision‑making, and neural population dynamics (Kelso, 1995; Rabinovich et al., 2008; Deco & Jirsa, 2012). Eye movements, especially the alternation between fixations and saccades ,   fit remarkably well into this metastable picture. The Freidlin–Wentzell theory of rare events (Freidlin & Wentzell, 2012) provides a rigorous mathematical language for these intuitions. 2. Fixations as metastable states  Consider the gaze position X t as the state of a stochastic dynamical system. During a fixation the  gaze remains confined to a small region, microsaccades and noise generate small fluctuations, and the gaze tends to return...
Post a Comment

Freidlin–Wentzell Theory and Rare Transitions in Diffusion Processes

-

 

After introducing Sanov’s theorem, Girsanov transformations, and their applications to diffusion processes, we now turn to the broader framework provided by Freidlin–Wentzell theory. This theory describes how stochastic systems with weak noise escape from basins of attraction, offering a geometric and variational perspective on rare transitions, metastability, and escape times.

Notation

For clarity, we summarize the variables and symbols used throughout this post:

  • \(X_t^\varepsilon\): diffusion process with small noise intensity \(\varepsilon\).
  • \(\varepsilon\): noise amplitude, assumed to satisfy \(0 < \varepsilon \ll 1\).
  • \(b(x)\): deterministic drift field defining the noiseless dynamics \(\dot{x} = b(x)\).
  • \(\sigma\): diffusion coefficient (scalar or matrix), assumed non-degenerate.
  • \(W_t\): standard Brownian motion.
  • \(\phi(t)\): candidate path used in the variational formulation of large deviations.
  • \(\dot{\phi}(t)\): time derivative of the path \(\phi(t)\).
  • \(I_{0,T}(\phi)\): Freidlin–Wentzell action functional over the interval \([0,T]\).
  • \(\|\cdot\|_{\sigma^{-1}}\): norm induced by the inverse diffusion matrix, \(\|v\|_{\sigma^{-1}}^2 = v^\top (\sigma\sigma^\top)^{-1} v\).
  • \(V(x,y)\): quasi-potential, i.e., minimal action required to move the system from \(x\) to \(y\).
  • \(\tau_{\text{escape}}\): first exit time from a basin of attraction.

1. Small-noise diffusions

Consider a diffusion process with small noise:

\[ dX_t^\varepsilon = b(X_t^\varepsilon)\,dt + \sqrt{\varepsilon}\,\sigma\,dW_t, \]

where \(0 < \varepsilon \ll 1\). When the noise is weak, the process typically follows the deterministic flow \(\dot{x} = b(x)\), but occasionally performs rare transitions between attractors.

2. The Freidlin–Wentzell action functional

The probability of observing an atypical trajectory \(\phi\) on \([0,T]\) is governed by the action functional:

\[ I_{0,T}(\phi) = \frac{1}{2} \int_0^T \|\dot{\phi}(t) - b(\phi(t))\|^2_{\sigma^{-1}}\,dt. \]

This functional is the continuous analogue of the rate functional in Sanov’s theorem: it quantifies the “energetic cost” required to force the system to follow a non-typical path.

3. Quasi-potential and metastability

The quasi-potential between two points \(x\) and \(y\) is defined as:

\[ V(x,y) = \inf_{\phi(0)=x,\;\phi(T)=y} I_{0,T}(\phi). \]

It represents the minimal cost needed to push the system from \(x\) to \(y\). When \(x\) and \(y\) belong to different basins of attraction, the quasi-potential determines the escape time:

\[ \mathbb{E}[\tau_{\text{escape}}] \asymp \exp\!\left(\frac{V(x, \partial A)}{\varepsilon}\right). \]

This exponential scaling is the hallmark of metastability.

4. Rare transitions and optimal paths

Rare transitions occur along optimal paths, i.e., trajectories that minimize the Freidlin–Wentzell action. These paths are the continuous analogue of the optimal empirical measures in Sanov’s theorem and the drift-adjusted dynamics obtained via Girsanov transformations.

In many systems, the optimal path satisfies the reverse-time equation:

\[ \dot{\phi}(t) = b(\phi(t)) - \sigma^2 \nabla V(\phi(t)). \]

This describes the most probable route by which the system crosses an energy barrier.

5. Connections to Schrödinger bridges and optimal transport

The Freidlin–Wentzell framework is deeply connected to several modern mathematical structures:

  • Schrödinger bridges: stochastic optimal transport with entropic regularization;
  • Optimal transport: small-noise limits of quadratic transport costs;
  • Girsanov transformations: the change of measure that realizes the optimal path;
  • Importance sampling: constructing tilted measures for efficient rare-event simulation.

These connections make Freidlin–Wentzell theory a natural bridge between stochastic analysis, statistical physics, and modern machine learning.

6. Concluding remarks

Freidlin–Wentzell theory provides a geometric and variational description of rare transitions in weak-noise dynamical systems. Together with Sanov’s theorem and Girsanov transformations, it forms a unified framework for understanding empirical deviations, pathwise deviations, and optimal changes of measure in diffusion processes.

Comments

Post a Comment

Popular posts from this blog

Understanding Anaerobic Threshold (VT2) and VO2 Max in Endurance Training

-
Image
  Introduction: The Science Behind Ventilatory Thresholds VT1 vs. VT2: Defining Energy Transitions p (Aerobic Threshold) : The transition from lipid metawwwpp (fat-burning) to a mixed metabolism, where both carbohydrates and fats fuel muscle activity. V0T2 (Anaerobic Threshold) : The point where the  shifts to a predominantly carbohydrate-based wmetabolism , leading to rapid lactate waccumpulation and increased reliance on anaerobic energy pathways. www w In p VT1 and VT2 Wwww ewAgep & Sex Younger athletes generally exhibit higher VT2 values , while aging naturally reduces aerobicw capacity. Men tend to have a higher VO2 Max due to greater lung capacity and muscle mass, but women can achieve similar endurance levels through optimized training. Body Composition & Health Status High body fat percentage may reduce VT2 efficiency, as excess weight increases ...
Read more

Diagnosing Singular Fits in Linear Mixed Models: Practical Examples, Code, and Alternatives

-
Image
  Introduction This post is the practical follow‑up to When Linear Mixed Models Don’t Converge , where I discussed the theoretical reasons why convergence issues often arise in linear mixed models. Here, we move from theory to practice: simulated examples, diagnostic tools, interpretation of singular fits, and practical modeling alternatives. If you haven’t read the first part yet, you may want to start there for the conceptual foundations. Figure. Conceptual map showing how model complexity and available information interact Legend of Figure 1. Green = identifiable models; yellow = risk of singularity; red = non‑identifiable models. A Concrete Example Using R: When a Random-Slope Model Becomes Singular Let’s simulate a simple dataset with only 8 groups and a very small random‑slope variance. This is a classic scenario where the model converges numerically but is statistically non‑identifiable. ```r set.seed(123) n_groups <- 8 n_per_group <- 20 group <- factor(rep(1:n_gro...
Read more

Owen's Function: A Simple Solution to Complex Problems

-
Definition   If you are interested in statistics or probability, you may have encountered the Owen's function , a special function that arises in various applications involving bivariate normal distributions. Let’s explain what Owen's function is, how it is defined and notated, and why it is useful.  The Owen's function is defined as follows: for any real numbers h and a, Owen's function T(h,a) is given by T ( h , a ) = 1 2 π · ∫ 0 a exp ( - h 2 2 · ( 1 ...
Read more