Mixed distribution analysis identifies saltation and stasis growth.

Area of Science:

• Developmental Biology
• Biostatistics
• Pediatric Growth Analysis

Background:

• Infant growth is often described as saltatory, characterized by periods of rapid growth (saltation) and no growth (stasis).
• Traditional methods for identifying these phases rely on specific algorithms applied to serial growth measurements.
• A model-independent statistical approach is needed to validate the saltatory growth hypothesis.

Purpose of the Study:

• To investigate the utility of finite mixed distribution analysis (MDA) for identifying saltation and stasis in infant longitudinal growth data.
• To determine if MDA can objectively identify distinct Gaussian populations within incremental growth data, supporting the saltatory growth model.
• To provide a growth model-independent statistical test for saltation and stasis.

Main Methods:

• Applied maximum likelihood method of finite mixed distribution analysis (MDA) to daily infant incremental growth data.
• Tested the null hypothesis that a single Gaussian distribution describes the data (smooth growth).
• Explored the presence of two distinct Gaussian populations, consistent with saltation and stasis phases.

Main Results:

• MDA identified that each individual's incremental growth data is best described as a mixture of two populations (chi-square, p < 0.05).
• One population centered around zero increment, consistent with stasis intervals.
• The second population showed unique distributions for each infant, reflecting individual saltation patterns in amplitude and frequency.

Conclusions:

• Finite mixed distribution analysis supports the concept of infant growth as a saltatory process with two distinct states (stasis and saltation).
• MDA provides a statistically robust, model-independent method for analyzing growth patterns.
• Individual growth patterns are unique and cannot be accurately reconstructed from group data.