🧠 neuroscience

Theory Note: Emergence of a Proportional-Derivative Control Law from Two Coupled Oscillating Brain Circuits Near Synchrony

This paper analytically demonstrates that two coupled oscillating brain circuits with delayed, odd-function interactions inherently implement a proportional-derivative (PD) control law near synchrony, establishing a direct link between oscillatory synchronization and feedback control without requiring additional mechanisms.

Original authors: Refy, O.

Published 2026-07-20

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Refy, O.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The Rhythm of Control: How Brains Might Dance to Move

Imagine your body as a giant, intricate orchestra. In the motor system—the part of your brain and nerves that tells your muscles when to move—there is a constant, rhythmic humming. Scientists call these "oscillations." You can think of them like the steady beat of a drum or the swaying of a pendulum. For a long time, researchers knew these rhythms were great for keeping time, like a conductor keeping an orchestra in sync, or for helping different parts of the brain talk to each other. But there was a big mystery: how does this rhythmic dancing actually help you stop your hand from shaking when you reach for a cup, or correct your balance when you trip? We know the brain uses "feedback control" (like a thermostat adjusting the heat) to fix mistakes, but we didn't know how the brain's natural rhythms fit into that job.

This paper steps into that gap by looking at two specific ideas that scientists already accept. First, the brain often uses "coupled oscillators," which are just two rhythmic things that influence each other, like two pendulums swinging on the same wall. Second, these rhythms often have a "delay," meaning it takes a tiny fraction of a second for a signal to travel from one to the other. The big question is: if you have two rhythms that are almost in sync, but with a tiny delay, does that setup accidentally create a perfect system for correcting mistakes?

The Discovery: When Synchronization Becomes a Controller

The author, Omar Refy, suggests a surprising answer: yes, it does. The paper shows that if you have two brain circuits (oscillators) that are nearly in sync with each other, and they talk to each other with a slight delay, they automatically turn into a "Proportional-Derivative" (PD) controller. That sounds like a mouthful of engineering jargon, but let's break it down with a simple metaphor.

Imagine you are trying to walk in perfect step with a friend (the "lead" oscillator) while you are both walking on a treadmill (the "plant" or the thing being moved). You are the "follower." If you fall a little behind, your brain needs to do two things to catch up:

Proportional: It needs to push harder based on how far behind you are. If you are just a tiny step behind, you take a small step. If you are a huge step behind, you take a giant leap.
Derivative: It needs to push harder based on how fast you are falling behind. If you are lagging but slowing down, you don't need to panic. But if you are lagging and speeding away, you need to sprint.

The paper proves that when two oscillators are nearly synchronized and have a delay in their connection, the math of their interaction naturally creates exactly these two rules. The "how far" part comes from the shape of how they talk to each other, and the "how fast" part comes directly from the time delay. It's as if the delay itself acts as a built-in speedometer for the correction.

How the Magic Happens (Without the Math)

The author uses a bit of "math magic" (linearization) to show this. Imagine the two oscillators are almost perfectly in step. The paper assumes the lead oscillator knows where you want to be, and the follower oscillator knows where you actually are. Because they are so close to being in sync, the difference between them is tiny.

When the lead oscillator sends a signal to the follower, it arrives a tiny bit late (the delay). Because of this delay, the signal the follower receives is actually a mix of where the lead was a moment ago and where it is going. When you crunch the numbers, this mix turns into a perfect recipe for a PD controller. The "Proportional" gain (how hard you push) depends on how sensitive the connection is, and the "Derivative" gain (how much you care about speed) depends entirely on the length of the delay.

In fact, the paper shows a neat trick: if you measure how strong the "speed" correction is compared to the "distance" correction, you can actually figure out exactly how long the delay is. The ratio of these two forces is just the delay time itself.

The Simulation: Does it Work in Practice?

To check if this theory holds up, the author ran computer simulations. They created a virtual system where a "follower" oscillator tried to drive a machine (a plant) to follow a target. They used a simple, rhythmic connection (a sine wave) with a delay of 0.12 seconds.

The results were striking. When the system was near synchrony, the machine moved exactly like it was being controlled by a perfect, ideal PD controller. The step response (how it reacted to a sudden change) matched the theoretical prediction perfectly.

However, the paper also shows the limits of this magic. The author tested what happens when the oscillators are not in sync. They gave the system a little push to throw it off rhythm. When the oscillators were close to being in sync, the system corrected itself perfectly, and the machine returned to its target. But when the oscillators were far out of sync, the system failed to correct the machine's position properly, leaving a permanent error. This confirms that the "PD controller" effect is a special property that only emerges when the rhythms are nearly aligned.

What This Means for the Real World

This paper doesn't claim to have found a new machine in the brain, but it offers a new way of looking at what we already see. It suggests that the rhythmic synchronization we see in motor circuits might not just be for keeping time; it might be the mechanism for controlling movement.

The author proposes three things to look for in real experiments to see if this is true in living brains:

The error in movement should show up as an error in the phase (timing) between the brain signals.
If you mess up the synchronization, the ability to correct movements should get worse, even if the timing of the rhythm stays the same.
The relationship between the "distance" correction and the "speed" correction should match the delay in the signal.

The paper concludes that synchronization and feedback control, which scientists usually treat as two separate ideas, might actually be two sides of the same coin. Near synchrony, the act of syncing up is the act of controlling. It's a playful, elegant idea: that the brain might not need a separate "controller" to fix mistakes; it just needs to keep its rhythms dancing in step, and the corrections happen automatically.

Technical Summary: Emergence of a Proportional-Derivative Control Law from Two Coupled Oscillating Brain Circuits Near Synchrony

Problem Statement
Oscillatory activity and synchronization are ubiquitous features of the motor system, observed in cortical motor areas, corticomuscular coherence, and central pattern generators. While the roles of these rhythms in rhythm generation, temporal coordination, and inter-areal communication are well-established, their specific contribution to the moment-to-moment feedback control of movement—specifically the regulation of limb states against perturbations—remains unclear. Traditionally, synchronization dynamics and feedback control are treated as separate principles in motor neuroscience. This paper addresses the gap by investigating whether the synchronization dynamics of coupled oscillators can inherently implement a feedback control law.

Methodology
The author employs an analytical derivation to examine the interaction between two coupled oscillators: a "lead" oscillator encoding a desired plant state ( $x_d$ ) and a "follower" oscillator driving a plant with state $x(t)$ . The analysis relies on three core assumptions:

Phase-State Encoding: Both oscillators share a locally linear encoding map $g$ where phase $\phi$ relates to state $x$ via a constant slope $g'$ .
Odd Coupling: The interaction is governed by a coupling function $\Gamma$ that is an odd function of the phase difference $\Delta\phi = \phi_L - \phi_F$ (e.g., sinusoidal coupling as in Kuramoto models), ensuring the follower is advanced when lagging and retarded when leading.
Delayed Interaction: The coupling includes a fixed delay $\delta$ on the reference (lead) pathway, such that the interaction at time $t$ compares the delayed lead phase $\phi_L(t-\delta)$ with the current follower phase $\phi_F(t)$ .

The derivation linearizes the system near the synchrony limit ( $\Delta\phi \approx 0$ ) to approximate the coupling function. This approach is model-independent, applying to any oscillator type provided the near-synchrony and odd-coupling conditions are met.

Key Contributions and Derivation
The central theoretical contribution is the demonstration that two near-synchronous oscillators interacting through delayed, odd-function coupling necessarily implement a Proportional-Derivative (PD) control law on the plant they drive.

By expanding the delayed lead phase to the first order ( $\phi_L(t-\delta) \approx \phi_L(t) - \delta\dot{\phi}_L(t)$ ) and substituting the phase-state encodings, the coupling term $\Gamma(t)$ is shown to approximate:
$\Gamma(t) \approx K_p e(t) + K_d \dot{e}(t) - \delta \dot{x}(t)$
Where:

$e(t) = x_d(t) - x(t)$ is the state error.
$K_p = \kappa / g'$ is the proportional gain, determined by the coupling strength and the slope of the encoding map.
$K_d = \kappa\delta / g'$ is the derivative gain, determined by the coupling strength, the encoding slope, and critically, the interaction delay $\delta$ .
The term $-\delta \dot{x}(t)$ represents a residual damping term proportional to the plant velocity.

A significant finding is the direct relationship between the control gains and the physical delay: the ratio of derivative to proportional gain ( $K_d / K_p$ ) equals the interaction delay $\delta$ . Thus, PD control emerges with no additional machinery beyond the synchronization dynamics themselves.

Results
Numerical simulations were conducted using a second-order plant driven by sinusoidal coupling ( $\Gamma = \kappa \sin[\phi_L(t-\delta) - \phi_F(t)]$ ).

Step Response: The coupled-oscillator system's response to a step target closely matched an ideal PD controller using the analytically predicted gains ( $K_p = \kappa$ , $K_d = \kappa\delta$ ).
Perturbation Recovery: The system was tested with brief perturbations under two conditions: near synchrony ( $\Delta\phi \approx 0$ $Δ ϕ \approx 0$ ) and away from synchrony ( $\Delta\phi \approx -0.15$ $Δ ϕ \approx - 0.15$ rad).
- In both cases, the phase difference recovered.
- However, the plant state $x(t)$ returned to baseline only when the system was near synchrony.
- Far from synchrony, the plant state retained a sustained offset, indicating that control quality degrades systematically as the system moves away from the linearized near-synchrony limit.

Significance and Claims
The paper establishes a direct, model-independent bridge between oscillatory synchronization and feedback control. It posits that synchronization dynamics and feedback control are not separate principles but two aspects of a single mechanism: near synchrony, the synchronization dynamics is the controller implementation.

The author suggests that this framework reframes pervasive motor oscillations as a plausible substrate for feedback control. The paper proposes three specific experimental signatures to identify such systems:

Regulation error should be encoded in the phase error between coupled populations.
Disrupting synchrony should specifically degrade control performance while leaving timing largely intact.
The effective derivative-to-proportional gain ratio should track the interaction delay ( $K_d / K_p \approx \delta$ ).

The author notes that these predictions apply specifically to systems of two coupled oscillators and may not hold for more complex networks. The work offers a concrete theoretical basis for testing the underlying mechanisms of motor control in candidate biological systems.