Research and writing on AI adoption and beyond

Introduces the MAP model — Measurement, Algorithm, Presentation — to describe how clinical observations are gathered, transformed, and returned to clinicians by AI decision-support systems, and the human-centred challenges at each stage.

A field study of adoption opportunities for a predictive Alzheimer's decision-support tool, conducted before it was built — combining workflow analysis, user profiles, and a design-reality gap model to surface adoption risks early.

Field research with a UK critical care unit's quality improvement team, translating how staff interpret longitudinal pain data into a set of dashboard visualisations for the hospital's electronic health record.

Explores what it takes to bring a predictive Alzheimer's model into clinical use — from field research on clinical needs to designing and testing icompass, a decision-support tool concept — and identifies which technical features and clinician profiles are best placed to drive early adoption.

A personal review of the Interaction Design Foundation's UX and interaction design courses, written partway through her PhD — what made the platform and its community genuinely learner-oriented, plus a few tips for anyone starting the same path.

A perspective paper outlining the idiosyncrasies of cognitive test data used in Alzheimer's data-sharing initiatives, written for computational researchers unfamiliar with how these tests are built and scored — with guidance on avoiding common misinterpretations.

Presents and tests icompass, a decision-support tool concept for guiding clinicians through the evolving pattern of Alzheimer's disease markers — the applied strand of the wider PhD research.

A multi-cohort validation of two data-driven models for sequencing the biological and clinical events of Alzheimer's disease, testing models trained on research data against real-world clinical cohorts.

Notes from the Technology and Dementia preconference at the Alzheimer's Association International Conference — a snapshot of how technology for early detection, monitoring, and clinical translation is being designed, and whether it's built to support clinicians and patients rather than just showcase what's technically possible.

A position paper — first presented at a UCL Grand Challenge Workshop on Human-Centred Explainable AI — arguing that explainability in healthcare AI is a human-centred requirement, not just a technical one: ML explanations only build clinician trust when they map to how clinicians actually reason, not to what's technically expressive.

A conference abstract presenting an early user study on the opportunities and barriers to translating disease progression models into clinical practice — the seed of the adoption questions later expanded in the PhD thesis.

An earlier collaborative study on the molecular effects of meditation and relaxation practice on coronary blood flow and inflammatory gene expression in patients recovering from myocardial infarction.

Part of the EU-funded MindSee project, this paper describes a symbiotic system that adapts literature search results to a user's ongoing relevance and satisfaction using physiological and eye-tracking data.