The Fung Model for Soft Tissue Biomechanics

Living biological tissues are masterworks of material engineering, capable of remarkable strength, flexibility, and resilience. However, describing their mechanical behavior with mathematical precision presents a significant challenge. Unlike simple steel springs that obey Hooke's linear law, tissues like skin, muscle, and arteries exhibit a complex, non-linear response: they are initially soft and compliant but become dramatically stiffer as they are stretched. This J-shaped stress-strain curve is a fundamental characteristic that linear models fail to capture, creating a knowledge gap between simple mechanics and complex biology.

This article explores the seminal work of bioengineer Y.C. Fung, who developed a powerful mathematical framework to describe this behavior. The Fung model, based on an exponential law of elasticity, has become a cornerstone of modern biomechanics. By reading, you will gain a deep understanding of this essential theory and its far-reaching consequences. First, in the "Principles and Mechanisms" chapter, we will dissect the mathematical heart of the model, exploring its exponential form, the physical significance of its parameters, its extension to account for the fibrous, anisotropic nature of tissues, and the inclusion of viscoelastic effects. Following this theoretical foundation, the "Applications and Interdisciplinary Connections" chapter will demonstrate how the model is put into practice, from interpreting experimental data to powering complex computer simulations that are revolutionizing medicine and bioengineering.

So, we've had our introduction. We've agreed that living tissue is a remarkable material, and we want to understand its mechanical secrets. But how do we go from a philosophical appreciation to a real, predictive science? We need a mathematical description, a "law" that tells us how a piece of tissue responds when we push or pull on it.

This is where things get interesting. A simple steel spring follows Hooke's Law: the force is proportional to the stretch. Double the stretch, you double the force. It's a beautifully simple, linear relationship. But pull on a piece of skin or a muscle, and you'll find it tells a very different story. At first, it's quite compliant, stretching easily. But as you pull further, it rapidly becomes much, much stiffer. It fights back with a vengeance. This non-linear, stiffening behavior is the hallmark of most soft biological tissues. Our first job is to capture this essential character.

The great bioengineer Y.C. Fung proposed a beautifully simple yet powerful idea. Instead of a linear relationship, he suggested that the stiffness of tissue grows exponentially with strain. The mathematical heart of this idea is the ​​strain-energy density function​​, which we call $W$. Think of $W$ as the amount of energy you store in a cubic centimeter of material when you deform it. All the information about the material's elastic response is packed into this one function.

For a simple material that looks the same in all directions (​​isotropic​​), the Fung model for the strain-energy density looks like this:

Don't be intimidated by the symbols. Let's break it down. $I_1$ is a strain invariant that measures the overall amount of deformation the material is experiencing. For an undeformed material, $I_1 = 3$, so the term in the exponent is zero, $\exp(0)=1$, and the stored energy $W$ is zero. This is just a fancy way of saying a relaxed tissue has no stored energy.

Now, what about the parameters $c$ and $k$?

• The parameter $c$ has units of stress (like Pascals). It sets the overall stiffness scale of the material. A tissue with a larger $c$ is fundamentally tougher, like the difference between a flimsy jellyfish and a tough piece of cartilage. It's the baseline strength.
• The parameter $k$ is dimensionless. This is the star of the show! It controls the nonlinearity of the response. A small $k$ means the material behaves more like a linear spring, while a large $k$ means the stiffness ramps up dramatically with very little strain. It dictates how quickly the tissue "fights back."

For instance, if we imagine a simple shear deformation, like sliding the top of a block of gelatin sideways, the resulting shear stress turns out to be $\sigma_{12} = c k \gamma \exp(k\gamma^2)$, where $\gamma$ is the amount of shear. You can see it right there in the formula: the stress isn't just proportional to the strain $\gamma$; it's multiplied by an exponential term that grows very, very fast. This exponential stiffening is precisely what we see in the lab.

The simple isotropic model is a great start, but we can generalize it to describe more complex materials. The more general Fung model looks like this:

Here, $Q(E)$ is a quadratic function of the strain tensor $E$. Think of it as a more sophisticated way of measuring "how much" strain there is, accounting for stretches and shears in all directions. It's a basket where we collect all the different components of strain, each potentially weighted differently.

Now, we must impose a crucial condition on this function $Q(E)$: it must be ​​positive-definite​​. This sounds like an abstract mathematical requirement, but it’s rooted in deep physical principles. It means that $Q(E)$ must be positive for any possible strain $E$ (unless $E$ is zero, in which case $Q(E)=0$). Why is this so important?

1. ​​A Stable Home:​​ Because $Q(E) > 0$ for any non-zero strain, the strain energy $W(E)$ is also greater than zero. The only state with zero energy is the undeformed state ($E=0$). This guarantees that the material has a unique, stable, stress-free reference state. The material "wants" to be in this state and you have to do work on it to move it away. Without this, a material could spontaneously deform and release energy, which would be a very strange world indeed!

2. ​​A Well-Behaved Start:​​ At very small strains, any smooth curve looks like a straight line. The positive-definite nature of $Q$ ensures that our fancy nonlinear model gracefully simplifies to the familiar, stable linear theory of elasticity for tiny deformations. The material doesn't do anything crazy when you just barely poke it.

3. ​​No Getting Lost:​​ This condition also ensures that the energy function is ​​convex​​. Geometrically, this means the graph of the energy versus strain looks like a bowl. This is a mathematician's dream, because it guarantees that for a given load, there's a unique, stable deformation state. The material's response is predictable and well-behaved.

4. ​​Infinite Resistance:​​ Finally, it implies that as the strain becomes infinitely large, the energy required also goes to infinity. You can't just stretch the material indefinitely; it becomes infinitely stiff. This property, known as ​​coercivity​​, is a basic physical requirement for a solid.

So, this one "simple" mathematical condition—positive-definiteness—is the linchpin that ensures our model is not just a bunch of equations, but a physically and thermodynamically sound description of a real material.

So far, we've talked about isotropic materials, which have the same properties in all directions. But very few biological tissues are like that. A tendon, an artery wall, a muscle—they all have preferred directions due to embedded collagen or muscle fibers. They are ​​anisotropic​​.

The Fung model's elegance shines here. We can incorporate this directionality by simply adding terms to our $Q(E)$ function. For a tissue with a single family of fibers, like a tendon, we can write:

Here's the beautiful interpretation:

• The first term, with the parameter $\alpha$, represents the contribution from the soft, isotropic ​​ground matrix​​—the "gushy" stuff in between the fibers.
• The second term, with the parameter $\beta$, represents the contribution from the stiff ​​fibers​​. The new invariant, $I_4$, is nothing more than the square of the stretch along the fiber direction! So, this term only "wakes up" and adds stiffness when the fibers themselves are stretched ($I_4 > 1$) or compressed ($I_4 1$).

This is wonderfully intuitive. If you pull the tissue along its fibers, you engage the powerful $\beta$ term and the material is very stiff. If you shear the tissue in a plane perpendicular to the fibers, the fiber length doesn't change ($I_4=1$), the $\beta$ term is zero, and the response is governed only by the softer matrix through the $\alpha$ term. The model naturally captures this anisotropic behavior.

And why stop at one fiber family? An artery wall might have two families of fibers wrapped helically around it. No problem. We just add another term for the second family:

The framework is a modular and powerful "language" for describing the complex architecture of living tissues.

Two more realities of biology need to enter our picture: water and goo.

First, soft tissues are mostly water. For all practical purposes, they are ​​incompressible​​. You can change their shape easily, but you can't squeeze them into a smaller volume. In the language of mechanics, this means their Poisson's ratio is very close to $0.5$. Our model can handle this. We can split the strain energy into a part that governs shape change (isochoric) and a part that penalizes volume change (volumetric). By making the penalty for changing volume very high, we can model this near-incompressibility. A careful analysis shows that our model predicts an effective Poisson's ratio of $\nu_{\text{eff}} \approx \frac{1}{2} - \frac{\mu}{2K}$, where $\mu$ is the material's shear stiffness and $K$ is its bulk stiffness. As the material becomes harder to compress ($K$ gets very large), the Poisson's ratio gets closer and closer to the ideal value of $0.5$. It’s beautiful to see a familiar concept from freshman physics emerge from this sophisticated nonlinear theory.

Second, our story so far has been about pure elasticity—like a perfect spring that gives back all the energy you put into it. But real tissues are more like a spring combined with a gooey dashpot. They are ​​viscoelastic​​. If you stretch a piece of tissue and then release it, it doesn't trace the same path back. It loses some energy, primarily as heat. This energy loss over a cycle is called ​​hysteresis​​. A purely elastic model like the ones we've discussed so far cannot capture this, because for them, the energy change depends only on the start and end points, not the path taken. Thermodynamically, their dissipation is always zero.

To account for this, we need to add a "memory" to the material. The ​​Quasi-Linear Viscoelasticity (QLV)​​ theory, also pioneered by Fung, does this via a hereditary integral. The idea is that the stress you feel now depends not just on the strain now, but on the entire history of how it was strained, with a "fading memory" where more recent strains have a stronger effect. This extension allows the model to predict phenomena like stress relaxation (if you hold the tissue at a constant stretch, the force required slowly decreases) and creep (if you apply a constant force, the tissue slowly continues to stretch), which are defining features of real biological tissues.

We now have a sophisticated and powerful model. But a model is just a story. To make it a science, we need to connect it to the real world through experiments. This brings us to a final, profound point about the interplay between theory and experiment.

Imagine you have our transversely isotropic model with its matrix parameter $\alpha$ and fiber parameter $\beta$. How do you determine their values for a specific tissue? You might think, "Easy, I'll do two experiments. I'll pull it along the fibers to get one stress-stretch curve, and then I'll pull it across the fibers to get another."

It turns out, this is not enough! For this type of exponential model, both of these simple stretching tests mix up the effects of $\alpha$ and $\beta$ in a way that creates a "conspiracy." Different combinations of $\alpha$ and $\beta$ can produce almost identical stress-stretch curves in these tests, making it impossible to disentangle them. The parameters are not uniquely identifiable.

The key is to perform a test that probes the material in a fundamentally different way. A simple shear test in the plane perpendicular to the fibers is perfect for this. As we saw, this deformation doesn't stretch the fibers at all, so the resulting shear stress depends only on the matrix parameter $\alpha$. This test isolates the matrix response, letting us pin down $\alpha$. Once we know $\alpha$, we can go back to our simple stretching data and easily figure out the fiber parameter $\beta$.

This is a beautiful lesson in the scientific method. Having a good theoretical model is only half the battle. You must also be a clever experimentalist, designing tests that can isolate different physical effects and break the conspiracies between your parameters. The true understanding of nature lies in this elegant dance between a beautiful theory and a discerning experiment.

Now that we have explored the mathematical heart of the Fung model, you might be asking a fair question: “This is all very elegant, but what is it good for?” A physical law or a mathematical model is like a tool. A box of shiny, unused tools is a sad sight; their true beauty is only revealed when they are put to work, shaping our world and our understanding of it. In this chapter, we will take the Fung model out of the toolbox and see how it helps us understand, predict, and even engineer the world of living tissues. This journey will take us from the lab bench to the supercomputer, and from the fundamental mechanics of materials to the rhythm of our own breath.

A model, no matter how sophisticated, is an empty shell until it is informed by reality. The first and most crucial application of the Fung model is its use in a dialogue with a piece of actual biological tissue. Imagine we take a small sample of an artery wall or skin and place it in a machine that can pull on it. As we stretch the sample, we measure the force it resists with. What do we see? We don't see a simple straight line like a common spring. Instead, we see a curve that is lazy at first and then rises with astonishing steepness. This is the famous “J-shaped” stress-strain curve, a mechanical signature of soft biological tissues.

Why this shape? At small stretches, the tissue is floppy; its tangled network of protein fibers, mainly collagen, is just beginning to unfold. But as the stretch increases, these fibers pull taut, and suddenly the tissue becomes incredibly stiff and resistant, like a rope snapping tight. This prevents the tissue from over-stretching and failing.

Herein lies the first glimpse of the Fung model's beauty. Simpler models, like the neo-Hookean model, predict a stress that grows linearly with strain at small deformations, failing to capture this dramatic stiffening. Other models, like the Ogden model, use polynomial terms and can be made to fit the curve, but often require many parameters. The Fung model, with its exponential form, $W = \frac{c}{2}(\exp(k(I_1-3)) - 1)$, is exquisitely suited for this task. The exponential function naturally produces this J-shaped curve, capturing the gentle initial response and the subsequent explosive stiffening with just two parameters, $c$ and $k$.

But how do we find the right values for $c$ and $k$? We ask the material! We perform the experiment, record the data points of stress versus stretch, and then use numerical methods—a sort of sophisticated, automated guess-and-check—to find the values of $c$ and $k$ that make the model's curve pass as closely as possible through our experimental data. This process, known as parameter fitting or calibration, gives a voice to the tissue, allowing it to tell the model how it behaves.

Of course, a tissue in the body is rarely just pulled in one direction. An artery inflates like a cylindrical balloon, stretching both around its circumference and along its length. To build a truly robust model, we must listen to the tissue’s response to more complex deformations. By performing multiple tests, such as simple tension and equibiaxial inflation (stretching equally in two directions), and fitting the model to all the data simultaneously, we create a far more reliable "digital twin" of the material. We can even establish precise mathematical relationships to see how a model like Fung's compares to an Ogden model, matching their behavior not just at the beginning of stretching but also at larger, more physiologically relevant stretches.

To a first approximation, one might think of a piece of tissue as a uniform, isotropic "blob"—a material whose properties are the same in all directions. The basic Fung model we've discussed so far makes this assumption. But look closely at a steak, and you see the grain of the muscle. Look at the wall of an artery, and you'll find tough collagen fibers wrapped around it like the hoops on a barrel. Most biological tissues are anisotropic; they are fiber-reinforced composites, exquisitely designed to be strong in the specific directions where they face the highest loads.

Is our model ruined? Not at all! The true power of the continuum mechanics framework that underpins the Fung model is its flexibility. We don't have to throw out the model; we simply enrich it. We can introduce new terms into the strain-energy function, terms that activate only when the material is stretched along the direction of a specific fiber family. For a tissue with two families of fibers oriented at angles $+\theta$ and $-\theta$, the energy function might look something like this:

Here, the terms $I_{4}^{(1)}$ and $I_{4}^{(2)}$ are special invariants that measure the square of the stretch within each of the two fiber families. This is a wonderfully intuitive idea: the total energy is the sum of the energy stored in the general ground substance and the energy stored by stretching the individual fiber populations. This approach allows us to model a vast array of tissues, from arteries and heart valves to skin and ligaments, with far greater fidelity.

The framework is also adaptable to geometry. Many tissues, like the pericardium sac around the heart or the leaves of a plant, are very thin. Treating them as full three-dimensional solids can be computationally wasteful. Instead, we can use the same principles to develop a two-dimensional membrane theory. By defining special 2D invariants, we can create a Fung-type model that works specifically for these thin structures, capturing their in-plane stretching and fiber reinforcement under a state of plane stress. This is a common and powerful strategy in physics and engineering: identify the essential characteristics of a problem and simplify the mathematics accordingly.

So we have a mathematical model, finely tuned with experimental data, that accounts for the tissue's structure. What can we do with it? We can build a virtual world. One of the most profound applications of these constitutive models is in ​​Finite Element Analysis (FEA)​​.

The idea behind FEA is simple in principle: to analyze the mechanics of a complex object, like a human heart or an artery with a stent inside it, you break it down digitally into a huge number of tiny, simple shapes (the "finite elements") connected at nodes. The computer can then solve the fundamental equations of force balance for this interconnected mesh. The Fung model plays the role of the "constitutive law" or "material property" for each and every one of those tiny elements.

To make these colossal calculations work, the computer needs to know not only the stress in an element for a given strain, but also how the stress changes as the strain changes. This quantity, a fourth-order tensor known as the ​​consistent tangent modulus​​, is the key to the rapid and stable convergence of the numerical solution. Deriving this tangent modulus is a challenging but essential exercise in tensor calculus. The final expression, when translated into a matrix form suitable for a computer, allows for the efficient simulation of incredibly complex biological events. It is this machinery that allows biomedical engineers to design better medical devices, surgeons to plan interventions, and researchers to understand disease. When you see a stunning computer simulation of a beating heart, you are witnessing the Fung model—or one of its descendants—in action.

There is one last piece of the puzzle, and it may be the most important. Living tissues are not just elastic like a rubber band; they are viscoelastic. If you apply a sudden stretch and hold it, the stress won't stay constant—it will gradually relax. If you apply a constant load, the tissue will slowly "creep". This is because tissues are filled with water and long, tangled polymer chains. Resisting deformation involves not just stretching atomic bonds, but also pushing fluid through a porous matrix and untangling molecular spaghetti. These processes take time and dissipate energy.

How can we capture this time-dependent nature? A beautifully simple starting point is to model the tissue with a combination of springs (representing elasticity) and dashpots (representing viscosity). For instance, a simple spring-and-dashpot arrangement can remarkably predict a real physiological phenomenon: the fact that the compliance of your lungs (how easy they are to inflate) depends on how fast you breathe. At slow breathing rates, the viscous elements have time to relax, and the lung seems more compliant. At rapid rates, the dashpots don't have time to move, they feel "stiff," and the lung seems less compliant.

This simple linear model is insightful, but it cannot handle the large strains and nonlinear elasticity of real tissues. This is where Y.C. Fung made his most celebrated contribution: the theory of ​​Quasi-Linear Viscoelasticity (QLV)​​. The name sounds intimidating, but the core idea is one of breathtaking elegance and simplicity.

Fung recognized that the stress response of a tissue is not linear in strain, but he hypothesized that it might be linear with respect to its own elastic response. The Boltzmann superposition principle of linear viscoelasticity states that the total stress is the sum (or integral) of responses to all past strain increments. QLV proposes something similar, but different: the total stress is the sum of responses to all past increments of the instantaneous elastic stress.

Mathematically, this means the stress is given by an integral like:

Here, $\sigma^{(e)}(\epsilon)$ is the nonlinear elastic stress we have been discussing (which could come from our exponential Fung model), and $G(t)$ is a "reduced relaxation function" that describes how that elastic response fades in time. The model cleverly separates the nonlinear dependence on strain from the time-dependent relaxation. It's "quasi-linear" because superposition still works, not on the strain $\epsilon$, but on the transformed variable $\sigma^{(e)}$.

This QLV theory was a revolution. It provided a thermodynamically consistent framework that could accurately describe the stress relaxation, creep, and hysteresis observed in virtually all soft tissues. While the underlying assumption of separability has its limits—it works best for simple loading types and may break down under complex, rotating stretches—its power and utility have made it an indispensable tool in biomechanics for decades.

From the fitting of experimental data to the simulation of virtual surgeries and the modeling of the very rhythm of life, the Fung model and its QLV extension form a coherent and powerful intellectual framework. It is a testament to the idea that even the complex, "messy" world of biology is governed by principles of profound mathematical beauty and unity.