Equitable modelling of brain imaging by counterfactual augmentation with morphologically constrained 3D deep generative models

We describe CounterSynth, a conditional generative model of diffeomorphic deformations that induce label driven, biologically plausible changes in volumetric brain images. The model is intended to synthesise counterfactual training data augmentations for downstream discriminative modelling tasks where fidelity is limited by data imbalance, distributional instability, confounding, or underspecification, and exhibits inequitable performance across distinct subpopulations.Focusing on demographic attributes, we evaluate the quality of synthesised counterfactuals with voxelbased morphometry, classification and regression of the conditioning attributes, and the Frechet inception distance. Examining downstream discriminative performance in the context of engineered demographic imbalance and confounding, we use UK Biobank and OASIS magnetic resonance imaging data to benchmark CounterSynth augmentation against current solutions to these problems. We achieve state-of-the-art improvements, both in overall fidelity and equity. The source code for CounterSynth is available at https: //github.com/guilherme-pombo/CounterSynth.

Automated summary

CounterSynth is a conditional generative model that uses diffeomorphic deformations to induce label driven, biologically plausible changes in volumetric brain images. The model is designed to synthesise counterfactual training data augmentations for discriminative modelling tasks affected by data imbalance, distributional instability, confounding, or underspecification, and exhibit unfair performance across different subpopulations. Through various evaluations, we demonstrate that CounterSynth achieves state-of-the-art improvements in fidelity and equity compared to current solutions. The source code for CounterSynth is available at https://github.com/guilherme-pombo/CounterSynth.