Thesis
Personalised medicine for psychotic symptoms in Alzheimer's disease : elucidating mechanisms and improving assessment
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2026
- Thesis identifier
- T18211
- Person Identifier (Local)
- 202182376
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- Psychotic symptoms (delusions and hallucinations) are among the most prevalent and impactful neuropsychiatric symptoms of Alzheimer’s disease (AD), being associated with serious adverse outcomes which make them important targets of investigation. This thesis aims to explore the underlying mechanisms of different manifestations of psychotic symptoms in AD, and to propose a novel digital tool to improve their monitoring. Addressing the first aim, two systematic reviews and a series of studies, using secondary data made available from the ADNI database (adni.loni.usc.edu), were performed, proposing novel methodological and analytic approaches. Study 1 investigated the stability of a neuropsychiatric symptoms clusters: results indicated that while a psychotic cluster can be identified, this is not stable, and the patterns of association with other neuropsychiatric symptoms are dependent upon disease progression. Through the first systematic review and the related Study 2, neuroanatomical correlates and predictors of psychotic symptoms in AD were investigated, broken down into delusions and hallucinations, and the thematic subtypes misidentification and paranoia. These indicated that while certain brain areas appear to be linked to all psychotic symptoms, distinct patterns appear in association with specific manifestations. The second systematic review and the related Study 3 identified comparable findings from a cognitive perspective, as psychosis as a cluster was associated with overall greater cognitive dysfunction while clearer profiles emerged looking at the individual symptoms and themes. On the basis of the findings of the previous studies and systematic reviews, Study 4 employed a machine learning technique, Random Forest analysis, to propose potential multi-modal markers of a psychosis phenotype in AD. Lastly, addressing the second aim, Study 5 introduced a new digital tool which allows for a frequent monitoring of psychotic symptoms in AD, adding novel elements to current assessment strategies, including capturing daily symptoms fluctuations, and timely identification of potential triggers and symptoms co-occurrences.
- Advisor / supervisor
- McGeown William J.
- Moshfeghi, Yashar
- Resource Type
- DOI
- Funder
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T18211 | 2026-10-01 | Public | Download |