Welcome to the Digital Health Unit
We are biomedical engineers, physicists, and mathematicians working closely with doctors, clinicians, and nurses to improve health through data, AI, and computational science
AI for Healthcare
We develop robust, clinically meaningful AI methods for prediction, decision support, and health data analysis.
Multimodal Data
We work with multimodal data including clinical records, behaviour, wearable devices, imaging, genomics and multi-omics, microscopy, and real-world data to develop impactful digital health solutions.
Translation to Practice
We work closely with hospitals, startups, healthcare technology companies, and pharmaceutical partners to move innovation into real clinical and societal impact.
Featured Projects
CardioGPT
Cardiac function, Cine MRI, Diagnosis
RAPID-AIM
Cardiac function, ECG, Diagnosis
ZeCardioAI
Cardiomyopathies, Drug discovery, Widefield fluorescence microscopy, Artificial intelligence, Knowledge graph
LiverAIM
Liver fibrosis, Screening, AI tools
AI for Mental Health Diagnosis
Mental health, Machine learning, Diagnosis
Past Events
IV Congreso de personas con cáncer y familiares de Cataluña 2026
Barcelona · June 13, 2026
Taula rodona sobre IA, Humanització i Càncer
De la imagen al diagnóstico: IA aplicada a cardiología y ultrasonido 2026
Barcelona · March 2, 2026
Speaker
Our institutional home
The Digital Health Unit is part of the Life Sciences Department at the Barcelona Supercomputing Center (BSC-CNS) in Barcelona.
BSC hosts MareNostrum 5, one of Europe’s leading pre-exascale supercomputers and a key infrastructure for scientific research. MareNostrum supports large-scale computational work across many fields, including life sciences, artificial intelligence, biomedical research, and digital health.
Visit our BSC pageIn the News
What is Neuralese and Why it is Bad?
September 3, 2026
Chain of thought is the sequence of intermediate reasoning steps an AI model uses to work through a problem before producing an answer. Until recently, those steps were often expressed in human-readable language, which gave researchers at least some visibility into how a model was reasoning. Neuralese changes that: instead of “thinking” in words, a model can reason through dense internal numerical representations that humans cannot easily interpret.
The concern is that this could make powerful AI systems much harder to monitor. If a model is reasoning in an unexpected way, researchers may no longer be able to spot those warning signs by inspecting its intermediate reasoning.
AI in pediatrics: a 2026 guide
August 16, 2026
While some medical AI tools show real promise for children, like algorithms that help detect heart defects on fetal ultrasounds, pediatric AI is actually lagging far behind adult applications. This article explains the reasons why: it covers the lack of pediatric data used in AI training, the regulatory hurdles, and why you can't simply repurpose adult models for growing children whose anatomy and baselines are constantly changing.
Read more
Privacy risks from medical AI tools are not shared equally
August 4, 2026
Recent research published in Nature reveals that sophisticated "privacy attacks" can identify if someone's personal medical data was used to train a specific AI model. The study finds that these privacy risks are not shared equally: patients from minority demographic groups are significantly more vulnerable to having their data exposed than those in the majority.
Read more