At SimulaMet (Simula Metropolitan Centre for Digital Engineering), a leading Norwegian research center in Oslo, developing next-generation digital health solutions is at the core of our mission. As a partner in SEARCH, SimulaMet brings world-class expertise in AI and synthetic data generation, building the technical backbone for creating realistic, privacy-preserving healthcare data.

From digital theory to clinical reality

In digital health, the quality of data determines the quality of innovation. While real patient data is the gold standard, its use is rightly restricted by privacy concerns. SimulaMet addresses this challenge head-on by pioneering deep generative models that create high-fidelity synthetic data. This allows researchers and developers to build and test powerful AI tools without ever accessing sensitive patient information.

SimulaMet’s role in SEARCH is to lead the charge in creating these foundational synthetic datasets, turning complex medical information into a secure, usable resource for innovation.

Pioneering Synthetic Medical Data

SimulaMet contributes its deep expertise in generating synthetic medical signals and images, an area where its researchers have already made groundbreaking contributions. Our team developed the first "DeepFake ECG", a generative model that produces an unlimited amount of realistic electrocardiogram data, and co-created HyperKvasir, the world's largest dataset of gastrointestinal images and videos.

This experience is critical to SEARCH. SimulaMet is leading the work focused on generating structured and signal-based synthetic data (like ECGs) and is a key contributor to creating synthetic optical and radiological images (like cardiac magnetic resonance imaging). Our work ensures the synthetic data is not only realistic but also robust, diverse, and interpretable.

Building the engine for innovation

SimulaMet’s researchers are directly involved in creating and validating the core AI models that power the SEARCH platform. Key questions we help answer include:

• Can we generate synthetic data that is statistically indistinguishable from real patient data?
• How can we create personalized synthetic datasets that reflect individual patient characteristics?
• What methods can make our generative AI models more explainable and trustworthy for clinicians?

By co-leading development of these advanced generative methods with SEARCH partners, SimulaMet provides essential tools to validate clinical use cases in cardiology, focused on the chronic cardiovascular conditions atrial fibrillation and ischemic heart disease.

Trust through technical excellence

In SEARCH, SimulaMet ensures that the synthetic data foundation is scientifically sound and technically robust. This includes:

• Developing and refining deep generative models (GANs, VAEs, etc.) for various medical data types.
• Leading the generation of synthetic ECGs and other structured medical data.
• Contributing to methods for creating personalized synthetic data.
• Ensuring the quality, diversity, and privacy-preserving nature of all generated datasets.
• Collaborating on standards for validating the technical and clinical quality of synthetic data.

By providing this core technical expertise, SimulaMet helps the project achieve its goal of creating a trusted, ethical, and effective framework for synthetic data in healthcare.

Looking ahead

SEARCH is a vital collaboration between technical pioneers, clinical experts, and industry leaders. SimulaMet is proud to help drive the core technological innovation at the heart of this effort. With our commitment to open science and cutting-edge research, we are ensuring that the synthetic data generated is not just a digital copy but a powerful tool for discovery.

As healthcare becomes increasingly data-driven, SimulaMet’s work in SEARCH is paving the way for a future where secure, high-quality synthetic data accelerates medical breakthroughs and improves patient care for all.

"At SimulaMet, we build the algorithms that turn digital noise into clinically meaningful synthetic data. Our goal is to make data accessible for innovation while putting patient privacy first."

SimulaMet SEARCH Team:

• Dr. Vajira Thambawita, Research Scientist
• Prof. Molly Maleckar, Research Professor
• Thomas Dreibholz, Chief Research Engineer
• Steven Hicks, Senior Research Scientist
• Giulia Monopoli, PhD candidate