Brain & Cognition / Psychology
Faculty of Social and Behavioural Sciences
University of Amsterdam
Phone: +31 020 525 6430
Email Joost Agelink van Rentergem Zandvliet
Personal webpage Joost Agelink van Rentergem Zandvliet
prof.dr. Ben Schmand (UvA, AMC)
prof.dr Hilde Huizenga (UvA)
prof.dr. Jaap Murre (UvA)
On March 27th, 2018, Joost Agelink van Rentergem Zandvliet will defend his thesis entitled:
Statistical Advances in Clinical Neuropsychology
The goal of this thesis was to improve the reliability of neuropsychological assessment, specifically by improving the normative comparison procedure. The first goal was to provide multivariate normative comparisons, which test the patient’s whole profile of scores. The second goal was to provide normative comparisons that are corrected for age, sex, and level of education. These goals had two requirements. First, a normative database had to be established with many, demographically diverse, healthy participants. Second, a statistical framework had to be developed that allows for demographically corrected multivariate normative comparisons with this new normative database. The statistical framework was the focus of this thesis.
First, we described how an aggregate normative database can be constructed by combining data from healthy people from multiple studies. Second, we described how multivariate normative comparisons can be made using an aggregate database with a multivariate multilevel regression model. It was shown that although the model can be fitted with missing data, it cannot be if there is missing overlap between tests. Third, we described how the model can be extended to accommodate missing overlap between tests, by incorporating a factor model. Fourth, the fit of different factor models for neuropsychological tests was compared in two studies. Fifth, the methods developed in this thesis were put to an empirical test, by using the results to predict the development of Parkinson’s Disease Dementia at a second measurement occassion. Sixth, we compared various correction methods for univariate normative comparisons with an aggregate normative database in clinical neuropsychology.