#DeCoding the Gut–Brain Axis: from multimodal data to behavioural insights

How are our gut and brain connected, and how does this relationship influence the way we think, feel and behave? Within the HEREDITARY project, this question is at the heart of ongoing research in Work Package 2, where partners are exploring the complex interplay between microbiota, brain activity, and human behaviour. 

The Deliverable 2.4: Linkage and feature extraction from gut-brain, intermediate evaluation, led by Radboud University Medical Center, marks an important step forward in this journey. Building on earlier work, it provides new evidence that combining multimodal data with advanced AI can reveal meaningful patterns linking the gut and the brain, bringing us closer to understanding this intricate biological system.

From data to discovery: integrating the gut and the brain 

At the core of this research lies a simple but ambitious idea: to move beyond isolated measurements, and, instead, analyse the gut and brain as a connected system. 

To achieve this, HEREDITARY researchers worked with data from the Healthy Brain Study (HBS), a large cohort of deeply characterised individuals. By combining brain imaging (resting-state fMRI), gut microbiota profiles, and behavioural and physiological data, the team applied a supervised multimodal data integration method (an advanced AI method) known as SuperBigFLICA, an extension of Linked Independent Component Analysis (LICA), designed to work with large-scale, heterogeneous datasets. This approach allows researchers to identify latent components (shared patterns across different types of data), which, in this context, correspond to hidden structures in the data that capture how microbiome composition, brain connectivity, and individual behavioural are interrelated. 

One of the most relevant outcomes of this work is the identification of robust gut-brain components (multivariate patterns that simultaneously involve microbiome features and brain activity). In particular, one component revealed a strong interaction between gut microbial composition, brain networks linked to reward and emotion (such as limbic and default mode networks), and health and behavioural measures such as anxiety sensitivity, life stress, and Body Mass Index (BMI). 

In an independant validation, this component was also able to predict food-related behaviour from an independent task. This task reflects how individuals value unhealthy vs. healthy food, and their likelihood of choosing unhealthy options. These findings validate the feasibility of supervised multimodal integration and identify promising biological targets for follow-up analyses. It shows that gut–brain interactions are not only measurable, but also meaningfully linked to real-life behaviour. 

Advancing Use Cases 4 and 5 

These findings directly contribute to HEREDITARY’s Use Case 4 and lay the groundwork for Use Case 5. 

  • Use Case 4 focuses on understanding gut–brain interactions in healthy populations. The intermediate results confirm that it is possible to identify stable and biologically meaningful gut–brain patterns at population scale. 
  • Use Case 5 will extend this approach to clinical data, exploring whether similar patterns can explain variations in mental health conditions and maladaptive behaviours, with broader applications in the prediction of other gut (.eg. Ulcerative Colitis) and brain related conditions (e.g. depressive episodes).
What comes next?

The work does not stop here. The HEREDITARY project will evaluate the robustness of the discovered components, apply them to broader clinical data, and extend analyses to metabolic markers and future hypotheses-driven studies on stress, diet, and hedonic eating. It will directly address disease relevance by examining whether the same multivariate gut–brain components explain variation in psychopathology and maladaptive eating in psychiatric populations. 

The next phase of the research will deepen the analysis of how gut–brain interactions relate to behaviour. These analyses will continue to leverage multimodal datasets (including brain imaging, microbiome data, stress responses and behavioural tasks) to further explore how stress, anxiety and dietary factors influence food choices through gut–brain mechanisms, and how the interaction between the gut microbiome and reward-related brain connectivity contributes to stress-related eating patterns in daily life. In parallel, future work will focus on defining what constitutes a “healthy” gut–brain profile and ensuring robust clinical interpretation of results.

HEREDITARY is receiving EOSC EU Node funding to advance Federated Learning experiments

The HEREDITARY project has been granted 40,000 credits from the European Open Science Cloud (EOSC EU Node) to support, among other things, its federated learning activities within a secure European research infrastructure.

About the EOSC EU Node

The European Open Science Cloud (EOSC EU Node) is the operational platform of the EOSC Federation, designed to facilitate open, collaborative and data-driven research in Europe. It supports multidisciplinary scientific work by providing access to digital research services such as computing and storage resources, containerized environments, and collaborative tools through institutional credentials. The platform promotes the sharing and reuse of research outputs in a secure, GDPR-compliant cloud ecosystem based on FAIR data principles and a credit-based access model.

The awarded credits will be used to deploy and maintain the central server required for federated learning experiments on EOSC virtual machines. Federated learning allows multiple institutions to collaboratively train machine learning models without sharing sensitive data. By using EOSC infrastructure, HEREDITARY can efficiently manage firewall configurations and incoming connections, overcoming common technical barriers associated with institutional IT restrictions.

The EOSC EU Node credits will support several weeks of experimentation, with server configurations adapted to different model sizes and algorithmic requirements. In parallel, HEREDITARY is exploring the development of an API-based solution that would allow researchers to deploy experiment-specific software containers through the EOSC Cloud Container platform. This approach aims to streamline workflows, facilitate testing and deployment, and potentially deliver open-source tools that could benefit the broader research community.

#DeCoding Federated Learning: from Infrastructure design to secure collaboration across Europe

How can hospitals collaborate on sensitive medical data without ever sharing the data itself? This is the core question behind Federated Learning (FL), and one of the key technological pillars of the HEREDITARY project.

Over the past two years, HEREDITARY has progressively designed, deployed and tested a federated learning infrastructure capable of connecting medical centres across Europe while ensuring that raw patient data never leaves its original location. What began as a technical design challenge has now evolved into a secure network supporting distributed machine learning experiments across heterogeneous datasets.

Building the Foundations: Computing Infrastructures

Federated Learning only works if each participating centre has the technical capacity to train models locally and communicate securely with the rest of the network. The first step was ensuring this. Under Deliverable D2.14 in Month 9 and lead by SURF, partners established secure computing infrastructures capable of handling sensitive clinical and genomic data, equipping centres with appropriate storage, processing power and secure communication channels. Thanks to this, data owners can process data locally, train models without centralising records and exchange model updates securely within the federation.

With local infrastructures in place, the next step was to design and validate the full federated learning architecture. Deliverable D2.11 in Month 18 presents a federated infrastructure that is secure, flexible and deployable across heterogeneous environments, including high-performance computing systems and cloud platforms. Encrypted communication via gRPC/TLS was implemented to protect model exchanges, while Secure Aggregation mechanisms (SecAgg/SecAgg+) were integrated to prevent the central server from accessing individual model updates.

The system was engineered to support both horizontal federated learning (same data types across centres) and vertical federated learning (different data modalities distributed across centres). Dedicated project workshops demonstrated that both approaches could run successfully across geographically distributed nodes, even when accounting for network latency between countries. By Month 18, HEREDITARY had a federated network capable of running both horizontal and vertical learning experiments on ALS data, without moving any raw records.

Securing the Communication: Communication Protocols

Security does not stop at this point. Deliverable D2.15 in Month 22 dives deeper into how model updates are protected during training. SURF analysed and validated advanced communication protocols within the federated learning framework. Three key mechanisms were the driving force behind this:

  • Secure Aggregation ensures that the server can combine model updates without seeing any individual contribution. Clients (Medical Centres) mask their updates using cryptographic techniques so that when all updates are aggregated, the masks cancel out, but no single update can be inspected independently. Tests showed no significant decrease in model performance, with only a modest increase in runtime due to additional communication steps.
  • Differential Privacy was also evaluated, introducing controlled noise to model updates to further reduce the risk of information leakage, again with minimal performance degradation.
  • Trusted Execution Environments were explored as an additional layer of security, though their hardware requirements make them less practical in heterogeneous clinical environments.

Beyond Simulation: paving the way for actual implementation

One key lesson emerging from this work is that federated learning is relatively straightforward in simulation, but deploying it across real institutions introduces new challenges: hardware variability, network latency across countries, IT coordination and regulatory compliance. Through interactive workshops and live experiments, HEREDITARY has moved beyond theoretical experimentation to operational deployment.

Today, the project operates a federated network linking multimodal clinical data without centralising any raw records. Advanced AI models can be trained across distributed datasets and privacy-enhancing technologies can be implemented with limited performance trade-offs. The infrastructure is reliable, secure and resilient. This “data stays at source” approach aligns closely with the principles of the European Health Data Space, demonstrating that privacy-preserving, cross-border health data collaboration is technically feasible.

The next step will arrive in June 2026, when the project moves from validated design to consolidated implementation. Deliverable D2.12 will formalise the full implementation of the federated infrastructure, while Federated Learning will demonstrate its clinical relevance through Deliverable D2.17, presenting intermediate results from the neurodegenerative use cases. Together, these upcoming milestone will mark a transition from infrastructure validation to scientific and clinical impact.

Aligning strategy, technology and clinical value: HEREDITARY’s 5th Plenary Meeting in Lisbon

On 5–6 February 2026, the HEREDITARY consortium gathered at Universidade Nova de Lisboa (UNL), Portugal, for its 5th Plenary Meeting and the first in-person meeting of the project’s third year. Over two intensive days, partners reviewed progress, aligned on strategic priorities, and advanced key technical developments that will shape the next phase of the project. 

The meeting followed directly after the Federated Learning Workshop (3–4 February), creating strong momentum around HEREDITARY’s core mission: enabling privacy-preserving, multimodal data analysis across European medical centres. 

Opening the meeting, Project Coordinator Gianmaria Silvello (UNIPD) provided a comprehensive overview of the project’s current status. With the first review completed and 41 Deliverables successfully delivered, the consortium is now fully focused on addressing reviewers’ recommendations and consolidating technical achievements into high-impact results. 

Throughout the first day, each Work Package presented its latest developments and next steps, demonstrating strong cross-WP integration and alignment with the project’s strategic objectives. The review of ongoing activities confirmed steady technical progress across data infrastructuresemantic integrationanalyticsvisualizationlegal frameworkandcitizen science, which reinforces the coordination between clinical, technical, social and legal dimensions. 

A central highlight of the meeting was the Federated Learning and Federated Analytics sessions. On the second day, SURF reported on the Federated Learning workshop and the evolution of infrastructure leadership. Discussions explored the idea of creating a living document to guide institutions in setting up secure federated learning environments. On the Federated Analytics side, the Hereditary Data Network (HDN) architecture and deployment roadmap were presented by UNIPD, ensuring a real HDN query system running by December 2026, with a clear maintenance plan, and preparing a demonstrator for reviewers in early 2027. These developments mark a decisive step towards operational federated workflow execution across heterogeneous clinical and genomic datasets. 

After this, the five HEREDITARY use cases were reviewed in detail, with particular emphasis on: data storage and sources clarification, strengthening the causal interpretation of results and ensuring robust legal alignment. The consortium reaffirmed that clinical relevance and methodological rigour must be a cenral topic in the project. 

Looking Ahead 

With federated learning infrastructure maturing, HDN endpoints being installed, FAIRification progressing, and use cases consolidating clinical relevance, the consortium is moving decisively towards delivering a scalable, privacy-preserving framework for multimodal health data analysis in Europe. 

The meeting concluded with a clear set of next action points: 

  • Online Plenary Meeting planned for June 2026. 
  • Steering Committee meeting planned for April 2026. 
  • Federated Learning Workshop at AAU (May 2026). 

The next two years will be key to the project’s results and impact, and HEREDITARY is aligned, coordinated and ready. Check out some photos from the event here: