Proportional intracranial volume correction differentially biases behavioral predictions across neuroanatomical features and populations

Abstract

Individual differences in brain anatomy can be used to predict variability in cognitive function. Most studies to date have focused on broad population-level trends, but the extent to which the observed predictive features are shared across sexes and age groups remains to be established. While it is standard practice to account for intracranial volume (ICV) using proportion correction in both regional and whole-brain morphometric analyses, in the context of brain-behavior predictions the possible differential impact of ICV correction on anatomical features and subgroups within the population has yet to be systematically investigated. In this work, we evaluate the effect of proportional ICV correction on sex-independent and sex-specific predictive models of individual cognitive abilities across multiple anatomical properties (surface area, gray matter volume, and cortical thickness) in healthy young adults (Human Connectome Project; n=1013, 548 females) and typically developing children (Adolescent Brain Cognitive Development study; n=1823, 979 females). We demonstrate that ICV correction generally reduces predictive accuracies derived from surface area and gray matter volume, while increasing predictive accuracies based on cortical thickness in both adults and children. Furthermore, the extent to which predictive models generalize across sexes and age groups depends on ICV correction: models based on surface area and gray matter volume are more generalizable without ICV correction, while models based on cortical thickness are more generalizable with ICV correction. Finally, the observed neuroanatomical features predictive of cognitive abilities are unique across age groups regardless of ICV correction, but whether they are shared or unique across sexes (within age groups) depends on ICV correction. These findings highlight the importance of considering individual differences in ICV, and show that proportional ICV correction does not remove the effects of cranium volumes from anatomical measurements and can introduce ICV bias where previously there was none. ICV correction choices affect not just the strength of the relationships captured, but also the conclusions drawn regarding the neuroanatomical features that underlie those relationships.

Brain-based predictions of psychiatric illness-linked behaviors across the sexes

“…Sex differences in neurobiology and behavior are well established (2,(5)(6)(7)(64)(65)(66)(67)(68)(69)(70)(71)(72)(73)(74)(75)(76)(77). More recently, researchers have also begun to look at sex differences in brain-behavior relationships (20,33,34,43,78). To explain the underlying factors driving these differences in clinical populations, sex-based and gender-based theories have been proposed.…”

“…Linear regression models and deep learning algorithms achieve comparable accuracies for brain-based behavioral predictions (23), but linear models avoid overfitting, are more interpretable, and are less computationally expensive (18). The predictive models used here rely on a similar framework as those previously described (19,20,43) to perform novel analyses addressing cross-behavioral model generalization within and across the sexes in the context of psychiatric illness-linked behaviors. We used linear ridge regression models to predict each behavioral score based on functional connectivity data (Figure 1E).…”

Confronting racially exclusionary practices in the acquisition and analyses of neuroimaging data

“…Therefore, brain-behavior associations captured by these models are representative of the specific populations they were trained on as opposed to the general relationships across the populations 16,68 . Even the use of field-standard covariates, such as the proportional intracranial volume correction for anatomical analyses, can differentially bias observed brain-behavior relationships across the sexes 69 . Given these findings, it is not unreasonable to expect that issues in recruitment and methodology will continue to impact our analyses and findings until they are addressed.…”

Reliable and generalizable brain-based predictions of cognitive functioning across common psychiatric illness

“…Here we demonstrate that the meta-matching model generalizes not only across diagnostic categories, but also between independent datasets relying on different measures of cognition, neuroimaging protocols, and data processing strategies. Usually, models trained in one dataset lose much of their predictive capacity when applied to an independent dataset, even when the two datasets are diagnostically or demographically similar 2,38,[41][42][43] . The meta-matching approach likely achieves this high level of generalizability by exploiting correlations amongst phenotypes, relying on a common set of neurobiological features which predict a broad range of behaviors that underlie an individual's global cognitive performance, independent of diagnosis or measurement methods.…”

“…This requirement far exceeds the vast majority of samples available to psychiatric research groups, calling into question both the utility and feasibility of developing clinically focused predictive models. Moreover, even brain-cognition predictive models derived from consortia-level samples can fail to generalize or show substantially reduced accuracy when applied to different datasets 2,38,[41][42][43] , greatly diminishing the scope of their potential applications. In large population-based cohorts, the functioning of specific brain systems can be leveraged to predict a broad variety of phenotypes, ranging from demographic factors to physical and mental health-related variables [44][45][46][47] .…”