Evidence for Embracing Normative Modeling

Abstract

In this work, we expand the normative model repository introduced in (Rutherford, Fraza, et al., 2022) to include normative models charting lifespan trajectories of structural surface area and brain functional connectivity, measured using two unique resting-state network atlases (Yeo17 and Smith-10), and an updated online platform for transferring these models to new data sources. We showcase the value of these models with a head to head comparison between the features output by normative modeling and raw data features in several benchmarking tasks: mass univariate group difference testing (schizophrenia versus control), classification (schizophrenia versus control), and regression (predicting general cognitive ability). Across all benchmarks, we confirm the advantage (i.e., stronger effect sizes, more accurate classification and prediction) of using normative modeling features. We intend for these accessible resources to facilitate wider adoption of normative modeling across the neuroimaging community.

Normative Modeling with the Predictive Clinical Neuroscience Toolkit (PCNtoolkit)

“…We refer to Rutherford et al (Rutherford, Barkema, et al, 2022) for a detailed overview of the possibilities of downstream analyses that can be conducted using normative models and an indepth comparison between normative modeling outputs (deviation scores) and raw data using different data modalities (structural and functional MRI) across several tasks (multivariate prediction (regression and classification) and case-control group different testing). Owing to this flexibility, and the ability to move beyond group level inferences to individual prediction, we consider that normative modeling is a promising method for understanding variation in large datasets.…”

Unraveling the Link between CNVs, General Cognition, and Individual Neuroimaging Deviation Scores from a Reference Cohort

“…By employing this approach, z-scores can be calculated for each individual across various neuroimaging modalities, quantifying individual variations against the mean and centiles of population reference norms. Normative models thereby shift focus from group-level to individual-level inferences [22][23][24][25] and allow us to quantify atypical developmental trajectories 26,27 . Normative modeling has proven its efficacy in correlating individual behavioral phenotypes with deviations from reference cohorts, spanning disorders like schizophrenia, autism, and ADHD [28][29][30] and mapping disease progression in Alzheimer's disease 31 .…”