Methodological considerations: Integrating measures across assessment modalities

Large-scale multimodal data have never been more accessible in neuroscience, providing an unprecedented opportunity to study interindividual differences. Integrating these data across assessment modalities, as opposed to treating each modality in isolation, requires careful consideration of both theoretical and practical nature. This chapter provides an overview of key concepts essential for truly integrative clinical neuroscience. Successful translation of basic research into clinical knowledge requires a common reference framework that accounts for dimensional phenotypes organized at multiple levels in a dynamically interacting system. At the same time, clinical care relies on relatively simple categories to enable efficient prevention, diagnosis, and treatment of pathological conditions. An appropriate framework must bridge both, with particular emphasis on aspects that are important for understanding the occurrence and developing treatment of dysfunctional states. Neuroscience methodology is embedded in this hierarchical organization, with different recording and stimulation methods targeting processes at different levels. Challenges remain in the actual integration of data obtained at different biological levels, on different timescales, at different spatial resolutions, and often in different species. In particular, cross-species translation of (clinical) research can only succeed if both the measurement methodology and the process to be measured are well conceived and have clear counterparts in humans and the respective animal models. Finally, we show how “deep multimodal phenotyping” represents an ideal case of neuroscience research crossing multiple levels of organization in single individuals. Emerging tools for complex multivariate data analysis should aim to successfully integrate these data into real-world knowledge.