This repository presents an exploratory, data-driven characterization of clinical heterogeneity in functional neurological disorder (FND).
The project combines:
- clinical phenotyping across symptom, trauma, affective, pain, and functional dimensions;
- three-dimensional UMAP to represent participants within a continuous, nonlinear clinical manifold;
- HDBSCAN to identify density-based profiles without requiring every participant to belong to a cluster; and
- natural language processing (NLP) and semantic modelling to organize clinical features into a structured, interpretable ontology.
- minimum spanning tree (MST) analysis to summarize neighbourhood relationships and visualize the topology of the clinical manifold.
The resulting clusters are treated as clinical profiles within a continuous symptom space, rather than as fixed or mutually exclusive disease subtypes. The repository is intended to support transparent reporting, reproducibility, and further methodological development.
This project is under active development. Code, documentation, figures, and citation details may change as the accompanying research is refined. Materials should not be used for clinical diagnosis or individual treatment decisions.
The manuscript associated with this project is currently under submission to Neurology:
Monteiro, S., Maillard, A., Louis, E., Hentzen, C., Al Chare, I., Baltasis, S., Teng, M., Adrien, V., & Garcin, B. (2026). Towards a Multidimensional Exploration of Functional Neurological Disorder. Manuscript submitted to Neurology.
If you use the code, documentation, figures, or other materials from this repository, please also cite the repository together with the article:
Monteiro, S. (2026). URO–FND Clustering Study Repo [Research repository].
A DOI, repository URL, and final publication details will be added when available.
You may share and adapt the repository materials for any purpose, including commercial use, provided that you:
- give appropriate credit;
- acknowledge to the CC BY 4.0 licence;
- indicate whether changes were made; and
- do not imply endorsement by the authors or participating institutions.
Participant-level clinical data are not automatically covered by this permission. Data access, sharing, and secondary use remain subject to the relevant ethical approvals, consent conditions, institutional policies, and data-use agreements.
Unless otherwise stated, the original materials in this repository are licensed under the Creative Commons Attribution 4.0 International Licence.
Copyright © 2026 Sara Monteiro and contributors.
