Deriving the Sigmoidal Hill Maturation Function

From a Saturating Age Response to a Maturation Fraction

A maturation function represents how a physiological capacity changes with age relative to a reference level. In a clearance model, for example, the function can express the fraction of a reference clearance that is present at a given age. The earlier material on maturation functions introduced this role; here we derive a common bounded, sigmoidal form from a saturating age-response curve and calculate its values.

The derivation is a modeling choice, not a claim that all biological maturation follows one universal curve. It is useful when the response is nonnegative, starts at zero (or close enough that a zero-baseline approximation is reasonable), rises with age, and approaches a finite plateau.

Writing the function in terms of \(a/A_{50}\) makes its units clear: age is divided by an age measured in the same units, so the ratio and its power are dimensionless. The equivalent ratio form is often convenient in calculations. For \(a=0\), the fraction is zero; for positive ages, it is strictly between zero and one under the stated parameter conditions.

Deriving the Hill Form

Suppose a nonnegative age-response \(R(a)\) approaches a finite upper limit \(R_{\max}>0\). To construct a Hill curve, define the dimensionless age-response ratio \(x=(a/A_{50})^h\). A saturating response has the form \(R(a)=R_{\max}x/(1+x)\): the numerator increases with age, while the denominator includes the same increasing term plus a positive baseline term. Dividing by the plateau gives the maturation fraction.

This construction has a simple interpretation: the response is expressed as a fraction of its limiting value. The power term allows the age-response ratio to increase nonlinearly before saturation. The resulting equation is a particular saturating curve, not the only possible mathematical form for a response that approaches a plateau.

The parameter conditions matter. If \(A_{50}=0\), the expression is undefined at age zero; if \(A_{50}<0\), it no longer represents a positive age scale. If \(h=0\), the positive-age fraction is constant rather than a developmental rise. Thus the maturation interpretation uses \(A_{50}>0\), \(h>0\), and ages \(a\geq0\). Age and \(A_{50}\) must use matching units, such as both years or both days.

These properties follow directly from the formula. At \(a=A_{50}\), the numerator and the two terms in the denominator are equal, giving \(A_{50}^h/(2A_{50}^h)=1/2\). At positive ages, the denominator exceeds the positive numerator, so the fraction is less than one. At age zero, the numerator is zero and the denominator is positive. Finally, as \(a\) becomes large, dividing numerator and denominator by \(a^h\) gives \(1/[1+(A_{50}/a)^h]\), which tends to one.

Checking the Curve’s Shape

For positive ages, differentiation provides a direct check that the curve rises rather than falls:

Every factor in the derivative is positive for \(a>0\), so the function is strictly increasing over positive ages. Its bounded range and increasing shape make it suitable for a response that approaches a plateau. The term “sigmoidal” refers to the S-shaped pattern that the curve can display; the exact curvature depends on the shape parameter and the age range being viewed. The derivative result does not establish that a particular observed biological process follows this curve.

Using the Fraction in a Clearance Calculation

When a maturation fraction is used as a multiplier on a reference clearance, the product has the same units as clearance. This is appropriate only when the model assumes that the age-dependent component is proportional to this fraction. Other effects or a nonzero immature baseline would require a different specification.

Diagnostics, Alternatives, and Limitations

Before interpreting calculations or using the curve in a model, check that its mathematical behavior matches the intended use. For selected ages across the data range, evaluate the fraction and verify that it is nonnegative, no greater than one, and nondecreasing. Confirm that ages and \(A_{50}\) use the same units. A value above one, a negative value, or a decrease with age signals an implementation or parameterization error when \(A_{50}>0\) and \(h>0\).

The curve’s boundary behavior is also a substantive assumption. It gives zero maturation at age zero and approaches a plateau of one. If the process has a meaningful positive response at the lower boundary, this zero-baseline form may not describe it well. Likewise, if observations over the relevant ages do not show a plateau, the curve’s saturating structure may be poorly supported. These concerns cannot be resolved by checking the equation alone; the modeled values and the observed data must be examined together.

The Hill function can be compared with other broad approaches without assuming one is universally preferable. A linear age effect changes at a constant rate over its specified range and does not automatically remain bounded outside that range. An exponential saturating curve also approaches a plateau, but has a different shape and parameterization. The Hill form is useful when a bounded rise with an adjustable nonlinear shape is appropriate. Choosing among these forms requires attention to the data range, the biological assumptions, and the adequacy of model predictions.

A common mistake is to treat \(A_{50}\) as an arbitrary age unit and substitute a value expressed in different units from \(a\). For example, entering age in months while entering \(A_{50}\) in years makes the ratio numerically wrong and can produce a misleading fraction. Convert one quantity so that both use the same units before calculating \(a/A_{50}\). Another error is to interpret \(F_{\mathrm{mat}}(a)\) as an observed proportion without qualification: it is the fraction specified by the model at that age, not automatically an empirical estimate from an individual observation.