Universität zu Köln

Scharfenberg, Daniel ORCID: 0000-0003-2781-3143 (2026). Evaluation of cognitive domains in people with Parkinson's disease - A network approach. PhD thesis, Universität zu Köln.

PDF DissertationScharfenbergCognitiveDomainsPD.pdf Bereitstellung unter der CC-Lizenz: Creative Commons Attribution. Download (7MB)

Abstract

Background & Objective: Cognitive impairment is a frequent non-motor symptom in people with Parkinson’s disease (PwPD), potentially impacting patients’ quality of life and affecting further therapeutic decision. Valid and reliable cognitive diagnostics are thus highly relevant. Cognitive domains are an essential feature of cognitive diagnostics in PwPD, yet they are inconsistently defined by different diagnostic guidelines and lack empirical justification. These issues may lead to unreliable cognitive subtype diagnoses, that is, cognitive profiles based on observed cognitive domain impairments, challenging reliability and validity of cognitive diagnostics. Psychometric network models of cognitive test scores might be a promising alternative for conceptualizing cognitive functioning as a complex, interconnected network. The present thesis project aims to evaluate the concept of cognitive domains in PwPD and to complement it by taking a network perspective on cognitive functioning in PwPD. Specifically, it aims to i) give an overview of existing empirical findings regarding cognitive domains in PwPD, ii) identify the cross-sectional structure of cognitive functioning in empirical data from PwPD, and iii) identify temporal dynamics of cognitive functioning in empirical data from PwPD. Within the empirical analyses, we aim to explore the usefulness of network perspectives on cognitive functioning as an alternative to categorical cognitive domain systems. Methods: The present thesis project comprises three studies: one scoping review and two empirical analyses. The scoping review (study I) featured a preregistered literature search aiming to identify sources of evidence that applied dimensionality reduction methods to cognitive test data from PwPD. The two empirical studies used data from the DEMPARK/LANDSCAPE study, a completed, observational, prospective multi-center cohort study that initially aimed to characterize the natural course of cognitive decline in PwPD. For study II, we applied exploratory graph analysis (EGA) based on psychometric network analysis to extensive cognitive test data (19 cognitive test scores) of N = 698 PwPD with different cognitive statuses to investigate the dimensionality structure of cognitive functioning and compare it to a theoretically assumed cognitive domain structure. In addition, we performed a conceptual replication of our analyses with data from a sample of over 60,000 people with healthy cognition. For study III, we applied dynamic EGA (dynEGA) to the longitudinal DEMPARK/LANDSCAPE cognitive test data, including data from N = 355 PwPD with different cognitive statuses and available repeated, yearly follow-up assessments up to three years after baseline (i.e., four time points of yearly assessments). By this, we investigated dimensions of cognitive test scores that showed similar rates of change over time and compared these dimensions to previous findings of study II and the theoretically assumed domain structure. Further, we applied graphical vector autoregression (GVAR) models to explore the sequential cognitive decline, that is, which longitudinal dimensions predict each other over time. Results: The scoping review included 22 studies that performed dimensionality reduction methods to cognitive test data from PwPD. Synthesis of findings revealed that most studies employed exploratory approaches. Methods and results of these studies were highly inconsistent and none of the studies reproduced their previously assumed cognitive domain structure. In study II, EGA revealed five dimensions of cognitive test scores that differed from theoretically assumed domains and strong associations between cognitive test scores from the same test paradigms. After reducing redundancy in the data, EGA identified a unidimensional structure of cognitive test scores. Findings were conceptually replicated with data from cognitively healthy people. DynEGA from study III found five longitudinal dimensions that again differed from both the cross-sectional findings of study II and the theoretically assumed domains. GVAR models showed two clusters of longitudinal dimensions and indicated a temporal dissociation between card sorting/cognitive flexibility and visuoconstruction. Discussion & Conclusion: All three studies revealed a considerable gap between theoretical assumptions about cognitive domain structures and empirical findings in PwPD. Interpretation of existing empirical studies is limited by inadequate choice and high heterogeneity of methods. Results of the empirical studies position network models as a promising alternative conceptualization of cognitive functioning compared to categorical cognitive domain systems, as they might better reflect shared cognitive processes and influences of method-specific variances. Additionally, incorporating information about longitudinal trajectories of cognitive decline might enhance diagnostic and prognostic accuracy of cognitive diagnostics in PwPD. Future research will have to develop new measures and indices of cognitive performance within the network framework to initiate the shift towards a holistic view of cognitive functioning.

Item Type:	Thesis (PhD thesis)
Translated title:	Title Language Evaluation kognitiver Domänen in Menschen mit Parkinson-Krankheit - Ein Netzwerk-Ansatz German
Creators:	Creators Email ORCID ORCID Put Code Scharfenberg, Daniel daniel.scharfenberg@uk-koeln.de https://orcid.org/0000-0003-2781-3143 UNSPECIFIED
URN:	urn:nbn:de:hbz:38-813925
Date:	2026
Language:	English
Faculty:	Faculty of Human Sciences
Divisions:	Faculty of Human Sciences > Department Psychologie
Subjects:	Psychology Medical sciences Medicine
Uncontrolled Keywords:	Keywords Language Parkinson's disease; cognitive domains; network analysis; cognitive diagnostics English Parkinson-Krankheit, kognitive Domänen; Netzwerkanalyse; kognitive Diagnostik German
Date of oral exam:	16 September 2026
Referee:	Name Academic Title Stahl, Jutta Prof. Dr. Kalbe, Elke Prof. Dr.
Funders:	Studienstiftung des deutschen Volkes
Refereed:	Yes
URI:	http://kups.ub.uni-koeln.de/id/eprint/81392