Causation, trait correlation, and translation: Developmental brain imaging in research on neuropsychiatric conditions of childhood

Abstract

An irresistible but elusive promise of the field of developmental neuroimaging is to advance mechanistic understanding of neuropsychiatric conditions of childhood, toward translation to higher-impact intervention. In this article we wish to address a diversity of perspectives on that promise, which were expressed in a summarizing forum of the Fetal, Infant, and Toddler Neuroimaging Group (FIT'NG) conference in Santa Rosa, CA in September 2023. We organize our remarks according to three contemporary paradoxes: (1) the contrasting implications of neural correlates of development that reflect causes versus effects (or epiphenomena) of behavioral atypicality; (2) the interpretation of transient deviations in brain development that are associated with enduring developmental traits; and (3) the intensifying pursuit of discovery of neural correlates of behavior in an era of still-limited capacity to manipulate the course of early brain and behavioral development. In the article we leverage examples of recent advances in brain and behavioral science that help reconcile progress, skepticism, and hope as an emerging field matures and attracts new scientists into its ranks.

Keywords: Behavior; Brain; Development; Early childhood; Neuroimaging.

Fig. 1

Competing models of causation (A-C) for an observed (laboratory) association between a heritable developmental trait (black square) and a heritable neural signature (white square). Circles represent genetic influences, typically additive or polygenic in natures. Study outcomes (1−7) are organized according to their consistency / non-consistency with each respective model of causation for the association:(A) The genetic cause of the neural signature may partially overlap with the genetic cause of the developmental trait; e.g. a neural signature specific to autism and the developmental trait of inattention; (B) the neural signature arises secondary to the emergence of the behavioral trait (an epiphenomenon); (C) the neural signature arises from its own independent cause and contributes to variation in severity of a genetically-unrelated behavior via consequences of variation in neural function, (e.g. effect of genetically-induced cortical hyperplasia on amplification of familial epilepsy risk, i and iii), or via pleiotropic genetic influences that affect both neural development and variation in the behavioral phenotype (e.g. effect of inherited NF1 mutation on optic glioma formation and inattention, i and ii). Note that case-control studies alone (1) are not informative in resolving competing hypotheses for direction of causation of an observed association.

Fig. 2

Genetic and environmental influences on the transmission of neuropsychiatric conditions across generations. Genetic and environmental pathways across three generations: G1 (grand-parental generation), G2 (parental generation), and G3 (offspring generation). A1, A2, and A3: additive genetic loci of influence. C1, C2, C3: common environmental influence. E3: unique (unshared) environmental influence. Dashed arrows represent environmental pathways of inter generational transmission, whereas solid arrows represent pathways of genetic influence. Gene-environment correlation refers to environmental influences (eg. C2) on offspring (G2) that are themselves influenced by the genotype of the parent (A2), such as rearing behaviors that are affected by a heritable (or partially heritable) mental health condition of a parent, or adverse life events for a child that occur as a function of a heritable (or partially heritable) conditions of a parent conferring elevated risk for such events (eg. substance use disorder). Gene environment interaction is present when the joint influence of a given genetic factor and a given environmental factor is either more or less than the sum of the magnitude of their respective influences (therefore non-additive).