A Neuroimaging Simulation Framework for Developing and Evaluating Causal AI

2026-06-26

AI Summary

This study addresses a critical limitation in causal AI for medical imaging—the scarcity of datasets with ground-truth causal structures—by introducing the first realistic 3D T1-weighted MRI synthesis framework that enables explicit causal control. The proposed method generates anatomically plausible brain images by sampling from a learned subspace derived from real data, combining template deformation with region-specific volumetric modulation guided by user-specified causal graphs to precisely manipulate target brain region volumes. Synthetic data produced by this approach exhibit remarkably low relative volume errors (0.3–2.66%) in targeted regions and minimal mean absolute errors (0.034–0.397 ml) in non-targeted regions, substantially outperforming existing methods. Furthermore, the framework exposes the tendency of current causal discovery algorithms to infer spurious associations, thereby establishing a verifiable benchmark dataset for advancing causal AI in neuroimaging.

Abstract

Causally linking disease-related factors to image-derived biomarkers provides a powerful pathway to understanding disease mechanisms. Despite growing interest in applying causal artificial intelligence (AI) approaches for this task, these methods still need to be adapted for complex medical images, and especially, neuroimaging. However, the lack of ground-truth data presents a barrier to development. To bridge this gap, we developed and tested a method for generating synthetic neuroimages, which adhere to a user-specified causal structure describing the non-image to image variable relationships, permitting the creation of ground-truth neuroimaging datasets. In the simulated T1-weighted magnetic resonance images, anatomical variability is modeled by sampling from a subspace estimated from real data and deforming a template image to create unique simulated subjects. Causal relationships are encoded via precise volumetric changes of any region-of-interest without unwanted global artifacts. We achieved relative volume errors of 0.3-2.66% for the targeted regions-of-interest and demonstrate their statistically significant causal relationships, while maintaining mean absolute errors for non-target brain regions between 0.034-0.397ml. An initial evaluation of causal discovery methods exposes their limited ability to suppress spurious connections, highlighting the need for image-appropriate methods. Our framework is the first to enable the generation of realistic synthetic 3D neuroimages with explicit causal control that can serve as the missing ground-truth data necessary for the objective benchmarking and development of causal AI methods.

Problem

• causal AI
• neuroimaging
• ground-truth data
• synthetic neuroimages
• causal discovery

Innovation

• causal AI
• neuroimaging simulation
• synthetic MRI
• ground-truth dataset
• anatomical variability