The human cortex is often measured by its total volume, yet volume is a composite of two biologically distinct dimensions: cortical thickness and surface area. Research across diverse clinical domains—from Alzheimer’s disease and depression to autism and chronic pain—increasingly demonstrates that these two metrics are governed by different genetic, developmental, and cellular processes. By "unmixing" these signals, researchers can identify specific neuroanatomical signatures that aggregate volume measurements often obscure.
The Biological Independence of Thickness and Area
The primary intuition behind morphometric decomposition is that the brain’s "depth" and "expansion" tell different stories. According to the radial unit hypothesis, cortical surface area is determined by the number of radial columns produced during early embryonic development, while cortical thickness is influenced by the number of neurons within those columns and subsequent processes like myelination and synaptic pruning.
Because these features have distinct origins, they respond differently to disease and environment. For example, Sean R. McWhinney and colleagues in the ENIGMA study found that while surface area is largely established in early life, cortical thickness remains plastic and sensitive to metabolic factors like BMI. Similarly, Tao Wang and Feng Shi showed that in amnestic Mild Cognitive Impairment, thickness and area exhibit unique patterns of degradation that correlate with different plasma biomarkers (Wang et al., 2016).
Why Aggregate Volume Masks Pathology
When researchers rely solely on gray matter volume, they risk a "cancellation effect." In early-stage Major Depressive Disorder, Daihui Peng and Gang Li observed that cortical thickness might decrease in a region while surface area simultaneously increases (Peng et al., 2015). A volumetric analysis would sum these changes and report "no significant difference," missing the structural "scar" entirely.
This decomposition is also vital for understanding comorbidity. Nanyu Kuang and Benjamin Becker found that in children with comorbid mental disorders, surface area reductions are additive (worse in combined cases), but cortical thickness follows a non-linear pattern where comorbid cases look deceptively like healthy controls (Kuang et al., 2022).
Beyond Thickness: Gyrification and Tension
The push for independence extends to the brain's folding patterns. Eduardo Caverzasi and Howie Rosen identified that carriers of the c9orf72 mutation show reduced gyrification (folding) decades before symptom onset, even in regions where cortical thickness is still normal (Caverzasi et al., 2019).
Furthermore, Yujiang Wang and Tobias Ludwig have proposed using scaling laws from statistical physics to further decouple these measures into independent components like "tension" and "shape" (Wang et al., 2020). This allows for the differentiation of conditions that look similar on the surface—such as healthy aging and epilepsy—by revealing that they have opposite effects on underlying white matter tension.
Disentangling Overlapping Signals
The ultimate goal of this research direction is to separate "nuisance" variance from "interest" variance. In A model of brain morphological changes, Raphaël Sivera and Nicholas Ayache demonstrate how to disentangle the signals of normal aging (ventricular expansion) from Alzheimer's pathology (hippocampal atrophy).
New computational tools like sisPCA, developed by Jiayu Su and David A. Knowles, provide a mathematical framework for this "unmixing." By decomposing high-dimensional data into independent subspaces, researchers can isolate the specific structural drivers of clinical outcomes, moving from broad descriptions of "brain change" to precise, interpretable models of neurobiological progression (Su & Knowles, 2024).