Coventry University Featured PhD Programmes
Sheffield Hallam University Featured PhD Programmes
Engineering and Physical Sciences Research Council Featured PhD Programmes
Norwich Research Park Featured PhD Programmes
Cardiff University Featured PhD Programmes

How to detect changes in cognition trajectories: longitudinal study design to efficiently estimate biomarker change-point outcomes and time-to-change-point

  • Full or part time
    Dr S White
    Dr B Tom
    Dr B Lehmann
  • Application Deadline
    Tuesday, January 07, 2020
  • Competition Funded PhD Project (Students Worldwide)
    Competition Funded PhD Project (Students Worldwide)

Project Description


It is common to make longitudinal observations, repeated measurements over time, of biomarkers (and other covariates). Rather than smoothly changing over time, some biomarkers may exhibit sharp changes in their behaviour; a change-point is an identifiable shift in the long-run biomarker level [1].

In specific situations, such as our interest in monitoring cognition, the abrupt change in the biomarker, rather than the specific level, is the primary outcome of interest. A rapid decline in cognitive ability is linked with dementia and other cognitive impairments, beyond the normal age-related decline [2,3].

The aim of this project is to help researchers answer one of the most important questions, how to design a study to detect such a change-point. There are many competiting factors, costs being a significant driver of study time, but also how to maximise the information from each participant while minimising the burden of participation in the study..

Estimation of the change-point in cognitive decline is well established in the cognition literature, using the Mini-Mental State Exam (MMSE) [4].

Change-point models are a non-linear longitudinal model of a biomarker, they make inference on the trajectory of the biomarker and its change-point, both of which may be a clinical outcome of interest by themselves or linked to increased risks of another disease or condition.

Study design of non-linear models is non-trivial, especially if we incorporate the issue of missing data due to participant attrition (drop out), and consider design in a Bayesian context (rather than the more familiar frequentist metrics) [5].

The project will investigate novel study designs that adapt the number and interval of observations for each individual and demonstrate whether this leads to an increased power to detect change-points within a biomarker.

As a distinct, but closely linked concept, is to consider change-points as the outcome of interest in a trial or comparison of treatment groups. A treatment might fail to prevent rapid decline, but could delay the time until onset of rapid decline; to test this involves analysing the time to the change-point, not just the existence of a change-point.

The project will be hosted in the MRC Biostatistics Unit in Cambridge, with visits and a potential short research visit to Oxford.

Funding Notes

The MRC Biostatistics Unit offers 4 fulltime PhDs funded by the Medical Research Council for commencement in April 2020 (UK applicants only) or October 2020 (all applicants). Academic and Residence eligibility criteria apply.

In order to be formally considered all applicants must complete a University of Cambridge application form. Informal enquiries are welcome to

Applications received via the University application system will all be considered as a gathered field after the closing date 7th January 2020

For all queries see our website for details View Website


1. Carlin, Gelfand, Smith (1992). Hierarchical Bayesian Analysis of Changepoint Problems. JRSSC.
2. Hall, Lipton, Sliwinski, Stewart (2000). A change point model for estimating the onset of cognitive decline in preclinical Alzheimer's disease. Stat Med.;2-3
3. Ji, Xiong, Grundman (2003). Hypothesis testing of a change point during cognitive decline among Alzheimer’s disease patients. J Alzheimer’s Dis.
4. Muniz Terrera, van den Hout, Matthews (2009). Random change point models: investigating cognitive decline in the presence of missing data. J Appl Stat.
5. White, Muniz-Terrera, Matthews (2016). Sample size and classification error for Bayesian change-point models with unlabelled sub-groups and incomplete follow-up. SMMR.

Related Subjects

Email Now

Insert previous message below for editing? 
You haven’t included a message. Providing a specific message means universities will take your enquiry more seriously and helps them provide the information you need.
Why not add a message here
* required field
Send a copy to me for my own records.

Your enquiry has been emailed successfully

FindAPhD. Copyright 2005-2019
All rights reserved.