Sequence Search: Automated Sequence Design using Neural Architecture Search

2026-04-16

AI Summary

This work proposes a novel, fully automated framework for magnetic resonance imaging (MRI) pulse sequence design by introducing neural architecture search (NAS) for the first time in this domain. Moving beyond reliance on human expertise, predefined templates, or large training datasets, the method integrates a differentiable Bloch simulator with a task-specific loss function to directly optimize pulse sequences via gradient-based search. The framework successfully reproduces established sequences—including spin echo, T2-weighted, and inversion recovery—while also discovering new, highly efficient sequences featuring reduced radiofrequency energy and non-standard refocusing phases that are difficult for human designers to conceive. This approach transcends traditional empirical limitations and demonstrates the potential of data-driven, end-to-end optimization in MRI protocol development.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Neuro-Symbolic Learning
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations

Abstract

Developing an MR sequence is challenging and remains largely constrained by human intuition. Recently, AI-driven approaches have been proposed; however, most require an initial sequence for parameter optimization or extensive training datasets, limiting their general applicability. In this study, we propose "Sequence Search," an automated sequence design framework based on neural architecture search. The method takes tissue properties, imaging parameters, and design objectives as inputs and generates pulse sequences satisfying the design objectives, without requiring prior knowledge of conventional sequence structures. Sequence Search iteratively generates candidate sequences through neural architecture search and optimizes them via a differentiable Bloch simulator and objective-specific loss functions using gradient-based learning. The framework successfully replicated conventional spin-echo, T2-weighted spin-echo, and inversion recovery sequences. Less intuitive solutions were also discovered, such as three-RF spin-echo-like sequences with reduced RF energy and refocusing phases deviating from the conventional Hahn-echo. This work establishes a generalizable framework for automated MR sequence design, highlighting the potential to explore configurations beyond conventional designs based on human intuition.

Problem

• MR sequence design
• automated design
• neural architecture search
• pulse sequence
• generalizability

Innovation

• Neural Architecture Search
• Automated MR Sequence Design
• Differentiable Bloch Simulator
• Gradient-based Optimization
• Pulse Sequence Discovery