Published online on 21 October 2026 in Vitality, Medicine & Engineering, the 2026 Field Guide to Scientific Operating Systems for Longitudinal Disease Research: Architectures, Workflows, Validation Frameworks, and Emerging Directions in ML-Enabled Biomedical Research examines a fundamental challenge facing modern biomedical science: how to transform vast amounts of heterogeneous research data into reliable, reproducible and clinically meaningful knowledge.
Co-authored by Ara S. Khachaturian, Václav Snášel, Jean-Marie Bouteiller, Zaven S. Khachaturian, Ornit Chiba-Falek, Jacques Touchon, among other international experts, members of the INDRC (International Neurodegenerative Disorders Research Center) and the CLARA (The Center for Artificial Intelligence and Quantum Computing in System Brain Research) Scientific Advisory Board.
Beyond individual AI models: an integrated scientific approach
Despite decades of investment in biomedical research, the ability to predict how diseases will develop in individual patients remains limited. Medical imaging, molecular biomarkers, genomic information, digital measurements and longitudinal clinical records provide complementary insights, but these data are often fragmented across institutions, collected using different methods and difficult to integrate.
The authors argue that progress requires more than increasingly powerful AI algorithms. It calls for integrated Scientific Operating Systems (SOS): coordinated scientific and computational frameworks that connect data, knowledge, AI models, research workflows, validation and governance throughout the research lifecycle.
The publication explores several complementary approaches, including multimodal foundation models, knowledge graphs, neuro-symbolic AI and agentic AI systems. Rather than advocating a single technology, it highlights the importance of combining their strengths within modular, interoperable and scientifically accountable systems.
Why this matters for neurodegenerative disease research
Alzheimer’s disease and other neurodegenerative disorders illustrate the complexity of predicting long-term disease trajectories. Biological changes may begin years or decades before clinical symptoms become apparent, while individuals with similar initial profiles may experience very different outcomes.
Building trustworthy and reproducible AI research
A central message of the field guide is that trustworthy AI must be designed into the entire scientific process. Data harmonization, transparent documentation, traceability of analytical steps, uncertainty estimation and independent validation are essential for assessing whether findings can be reproduced and whether models remain reliable across populations, institutions and time.
The publication also highlights the importance of human expert oversight, responsible data governance, international collaboration and alignment with regulatory expectations. AI-generated analyses must be critically evaluated rather than accepted solely because their outputs appear plausible.
A shared direction for the future of brain research
The field guide presents an evolving framework for researchers, clinicians, engineers and policy stakeholders working at the intersection of machine learning and longitudinal disease research. Its central proposition is that the future of predictive medicine will depend not only on advances in individual AI models, but also on the scientific infrastructure that enables evidence to be integrated, evaluated and progressively refined.
This systems-level perspective is relevant to CLARA’s interdisciplinary mission to combine artificial intelligence, advanced computational methods and neuroscience in the study of complex brain disorders.
Publication details
Khachaturian AS, Pronk NP, Sivakumaran S, et al. 2026 Field Guide to Scientific Operating Systems for Longitudinal Disease Research: Architectures, Workflows, Validation Frameworks, and Emerging Directions in ML-Enabled Biomedical Research. Vitality, Medicine & Engineering. 2026;9:1–27. Published online 21 October 2026.