Psychiatric neuroimaging designs for individualised, cohort, and population studies

Abstract

Psychiatric neuroimaging faces challenges to rigour and reproducibility that prompt reconsideration of the relative strengths and limitations of study designs. Owing to high resource demands and varying inferential goals, current designs differentially emphasise sample size, measurement breadth, and longitudinal assessments. In this overview and perspective, we provide a guide to the current landscape of psychiatric neuroimaging study designs with respect to this balance of scientific goals and resource constraints. Through a heuristic data cube contrasting key design features, we discuss a resulting trade-off among small sample, precision longitudinal studies (e.g., individualised studies and cohorts) and large sample, minimally longitudinal, population studies. Precision studies support tests of within-person mechanisms, via intervention and tracking of longitudinal course. Population studies support tests of generalisation across multifaceted individual differences. A proposed reciprocal validation model (RVM) aims to recursively leverage these complementary designs in sequence to accumulate evidence, optimise relative strengths, and build towards improved long-term clinical utility.

Fig. 1. Common psychiatric neuroimaging designs via heuristic data cube.

A balance of scientific goals and resource constraints leads common psychiatric neuroimaging studies to differentially emphasise design features. Prototypical examples of psychiatric neuroimaging designs are displayed according to the dimensions of sample size (y-axis), the number of different measures collected (x-axis), and the number of time points assessed (z-axis).

Fig. 2. Trade-offs among common psychiatric neuroimaging designs.

Due to finite resources and varying scientific goals, prototypical examples of psychiatric neuroimaging designs differentially emphasise within-person precision (e.g., measurement reliability, internal validity, and the potential for experimental control) compared to between-person generalisability (e.g., the potential for a sample to capture “real-world complexity”). Note the y-axis denotes a heuristic relative scale, where lower ‘values’ convey relatively lower (not an absolute absence of) within-person precision or between-person generalisability.

Fig. 3. Reciprocal validation model for psychiatric neuroimaging.

Common psychiatric neuroimaging designs have inherent methodological trade-offs (due to practical resource constraints and varying scientific goals) that can, nevertheless, be recursively sequenced to leverage relative strengths. A given psychiatric neuroimaging result may first emerge from a small sample, intensive longitudinal individualised or cohort study (e.g., a given brain region “A” changes with depression treatment) and be independently replicated with the same or similar design (left). The reciprocal validation model (RVM) emphasises the sequential testing of a conceptually related result (e.g., the link between depression symptoms and brain region “A”) to be tested for generalisability across individuals with a population study. Conversely, a psychiatric neuroimaging observation may start as an inter-individual difference neural correlate developed in a population sample (e.g., whole-brain connectivity correlate of depression) and be independently replicated with the same or similar population design. RVM emphasises testing this neural correlate for within-person “mechanisms” via interventions and tracking of precise longitudinal courses with individualised and cohort studies. We note that individualised and cohort studies are grouped in this figure based on the proposed shared study goal of within-person mechanisms. We refer the reader to earlier sections for further distinctions among these designs.