As part of the pre-project activities to build visibility (Phase 0 of impact maximisation), the project published the following article to gather feedback from the broad community of healthcare and IT experts on HealthManagement.org, which reaches 30.000 stakeholders and is supported by MindByte Communications.
Introduction
Neurological disorders are among the most urgent challenges today. In terms of years lived with disability (YLDs), they have already surpassed cancer and cardiovascular disease in Europe, making them a top priority on the global health agenda. Over 300 million people worldwide are directly affected, and the economic and societal costs amount to hundreds of billions of euros each year. Alzheimer’s disease alone is estimated to cost Europe more than €250 billion annually (European Brain Council), while the overall burden of neurological diseases, including Parkinson’s, multiple sclerosis, and stroke, represents a significant portion of global healthcare expenditure. This burden is not just financial; it also profoundly impacts individuals’ quality of life, affecting patients, families, and healthcare providers alike.
Against this backdrop, the emergence of generative artificial intelligence (GenAI) introduces a paradigm shift in how neurological diseases are understood, diagnosed, and treated. NeuroGenAI is an ambitious, multidisciplinary European research initiative designed to unlock this potential. Its mission is to develop the first foundational GenAI model and federated data space dedicated specifically to neurology and psychiatry. In doing so, NeuroGenAI aims to position Europe as a global leader in health innovation and personalised medicine, ensuring that technological progress is balanced with ethical oversight and clinical impact.
The NeuroGenAI Initiative
At its core, NeuroGenAI is a collaborative project that unites fifteen of Europe’s most esteemed research centres, hospitals, and brain institutes. Importantly, the initiative also extends beyond Europe, including clinical sites in America, Africa, and Asia. This global scope ensures that models developed within the project are not restricted to a narrow patient demographic but are validated across diverse populations and healthcare systems. Such diversity improves generalisability and increases confidence in the reliability of AI-driven insights.
The project is structured around three main clinical pillars. The first concentrates on neurodegenerative disorders such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (ALS). Here, NeuroGenAI will employ personalised progression modelling to support earlier interventions and more precise treatment planning. The second pillar addresses neuroinflammatory and pain-related disorders, including multiple sclerosis and chronic headache syndromes, where digital twin technologies enable patient-specific simulations and more accurate response predictions. The third pillar focuses on neuropsychiatric and neurodevelopmental conditions, including attention deficit hyperactivity disorder (ADHD) and bipolar disorder, where explainable and bias-aware generative models can help navigate fragmented and heterogeneous data landscapes.
Within these pillars, four high-value use cases have been prioritised:
- Disease trajectory simulation in neurodegenerative disorders using multimodal longitudinal data.
- Synthetic data generation for low-resource or sensitive domains to safeguard privacy whilst enhancing model training.
- Digital twin–based in silico clinical trials for multiple sclerosis and bipolar disorder.
- Multimodal data interpretation where large language models integrate imaging data with unstructured text for more cohesive decision-making.
By addressing these use cases, NeuroGenAI is establishing a versatile and scalable foundation for transforming patient care. Its goal is not only to develop new algorithms, but also to design systems capable of implementation in real-world hospital environments, improving diagnostic speed and accuracy, and enabling clinicians to personalise treatments with greater confidence.
Survey Insights: Understanding Stakeholder Perspectives
Before starting full-scale development, NeuroGenAI carried out a comprehensive survey to better understand the opportunities, challenges, and expectations related to the use of generative AI in neurology. The survey invited 10,000 stakeholders, including clinicians, researchers, policymakers, and technology experts. Of these, 198 completed the survey, offering a valuable range of perspectives that continue to influence the project’s direction.
Major Barriers to Adoption
The survey found that regulatory approval is seen as the biggest obstacle to adoption, mentioned by sixty percent of respondents. Concerns about trust in AI outputs (54%) and ethical and privacy issues (54%) also ranked highly. These results highlight the need to align AI initiatives with transparent regulatory frameworks and to build trust through clearly demonstrated safe, explainable, and ethical practices.
Figure 1. Major barriers to GenAI adoption in biomedical research and clinical practice
Opportunities in Neurology
When asked to identify areas with the greatest potential, respondents highlighted clinical documentation, early diagnosis, and drug discovery. These are fields where GenAI can deliver immediate and transformative benefits:
- Automating clinical documentation could reduce administrative tasks, freeing up to 20% of physicians’ time each week, according to OECD benchmarks.
- Early diagnosis, especially in Alzheimer’s and Parkinson’s, could add years of healthy living by enabling earlier interventions.
- Drug discovery aided by synthetic data and molecular modelling could reduce the typical 10–15 year development cycle by several years.
Figure 2. Potential impact of GenAI in neurology use cases
Integration Challenges
The survey also highlighted the practical challenges of integrating multimodal data, especially imaging, genomics, and electronic health records (EHRs). Two-thirds of respondents identified data standardisation as a key difficulty, while over half mentioned the problem of missing or incomplete datasets. This underscores the urgent need for federated data infrastructures and harmonisation strategies, which are central to the NeuroGenAI design.
Figure 3. Challenges tointegrating multimodal data
Explainability and Stakeholder Engagement
- Eighty-one percent of respondents emphasised that explainability is highly important in GenAI systems designed for clinical use.
- Seventy-nine percent highlighted the necessity for strong stakeholder engagement throughout all stages of development and validation, particularly in ethical governance.
From Survey to Strategy
The survey results were seen not just as background data but as a foundation for action. The NeuroGenAI strategy directly addresses stakeholders' concerns and expectations. Trust and explainability are built into the model design from the outset, with deliberate investment in frameworks for bias detection and transparent reasoning. Challenges related to data quality and integration are handled through federated data spaces that comply with privacy regulations while reducing fragmentation. Regulatory hurdles are proactively managed through engagement with the European Medicines Agency (EMA) and policymakers responsible for the AI Act and medical device regulation. Stakeholder involvement is formalised, with governance structures ensuring that patients, clinicians, researchers, and policymakers stay central to the initiative rather than just observers.
Health System Implications
For clinicians, the implications are immediate and practical. Generative AI tools developed through NeuroGenAI will assist in earlier diagnosis, providing more precise trajectories for conditions such as Parkinson’s or multiple sclerosis. They will also support personalised treatment recommendations that adapt over time, aligning with patient-specific data rather than broad generalisations. In Alzheimer’s disease, for example, AI systems could combine imaging, speech pattern analysis, and genetic data to predict progression years in advance, offering clinicians a window of opportunity for earlier intervention.
Hospitals and healthcare organisations will benefit from reduced inefficiencies, particularly in clinical trials and documentation. Digital twin simulations could decrease recruitment costs and shorten trial timelines by up to 30%, based on prior experience with oncology AI trials. Automated documentation systems may free clinicians to spend more time with patients, improving both satisfaction and outcomes. For health systems under strain, these efficiencies translate into tangible cost savings and enhanced patient flow.
Patients themselves stand to gain the most significantly. With models trained on diverse, multimodal datasets, they will receive more precise diagnoses, more personalised treatments, and ultimately a better quality of life. Significantly, AI-generated synthetic data can facilitate research into rare conditions like ALS, which often suffer from limited available data. This means patients with rare diseases will no longer be excluded from AI benefits simply because of small datasets. Meanwhile, policymakers and payers will have stronger evidence to inform funding decisions and long-term health planning. If AI systems demonstrate improved outcomes at lower costs, they will support stronger arguments for reimbursement and investment in digital health infrastructure. Health economics becomes central to the value proposition of NeuroGenAI, offering potential reductions in disability-adjusted life years (DALYs) and related economic losses.
Global Positioning: Europe as a Leader
One of NeuroGenAI's key features is its European identity. Although the United States and China are also making rapid advancements in AI for healthcare, Europe has chosen a path that prioritises ethics, transparency, and inclusivity. By embedding these values into the very foundation of its AI models, NeuroGenAI helps the European Union not only to compete technologically but also to set global standards. In doing so, Europe can shape international benchmarks for trustworthy AI, influencing not just how the technology is used within its own borders but worldwide.
The inclusion of clinical sites outside Europe bolsters this leadership role. By testing models in America, Africa, and Asia, NeuroGenAI ensures that its tools are relevant beyond a European patient base, making them more robust and globally influential. This international validation enhances credibility and strengthens the EU’s position as a leader in responsible AI globally.
Comparisons highlight Europe’s unique role. In the US, private industry leads development, often outpacing regulatory adaptation. In China, government-led initiatives focus on centralisation and scale, sometimes raising concerns about transparency. Europe, however, bases its competitiveness on trust, ethics, and patient inclusion. For international stakeholders, this approach increases reliability and encourages collaboration.
Future Outlook
Looking ahead, the potential of multimodal generative models in neurology is vast. Over the next five to ten years, we can anticipate systems that:
- Integrate real-time wearable data with hospital records for continuous monitoring.
- Support prevention by recognising risk profiles years before symptoms emerge.
- Foster cross-sector collaboration between pharma, biotech, and healthcare providers to accelerate therapy development.
- Train clinicians to utilise AI tools effectively, with curricula combining medical expertise and digital literacy. Incorporating these advancements into clinical practice could lead to a significant shift in how neurological disorders are managed worldwide within the next decade.
Conclusion
Generative AI has the potential to revolutionise neurology by creating new opportunities for diagnosis, modelling, and treatment. NeuroGenAI shows how Europe can responsibly use this technology by incorporating trust, explainability, and ethical governance at every development stage. Supported by survey insights, clinical expertise, and international collaboration, it is not just a research project but also a blueprint for the future of personalised medicine in neurology.
As the burden of neurological disease continues to grow, the significance of NeuroGenAI becomes even more vital. By advancing scientific understanding and clinical practice, and by addressing the concerns of regulators, clinicians, and patients, it represents a vital step towards a healthcare system that is more precise, more efficient, and above all, more compassionate.
Disclaimer
All survey data, graphs, and figures presented in this article are the intellectual property of HealthManagement.org and its partners. Reproduction, redistribution, or reuse in any form is strictly forbidden without prior written permission.
