Pathomechanisms and Signatures in the Longitudinal Course of Psychosis



005_ Polygenic risk score analysis of trajectories of cognitive performance in psychiatric patients

Research Question and Aims

Bipolar disorder (BD), schizoaffective disorder (SZA) and schizophrenia (SZ) are severe mental illnesses with a broad phenotypic as well as genotypic overlap. Cognitive deficits are a common symptom in patients suffering from these disorders, even though there are differences regarding domain as well as intensity (1,2). These cognitive deficits have a major impact on both quality of life and overall functioning of affected individuals (3). Existing research on cognition in these disorders is predominantly cross-sectionally and focuses on measuring different cognitive domains in the individual disorders. The course of cognitive abilities in severe mental disorders is less studied (1,4,5). Furthermore, recent studies hypothesize that higher schizophrenia polygenic risk scores (SZ-PRS) are associated with poorer cognitive performance (6,7,8). Using a novel approach to cluster longitudinal data, we would like to investigate whether this technique has the potential to identify novel trans-diagnostic patient groups that differ in cognitive performance and correlate them with SZ-PRS. Our study has two main aims:
1) To allocate participants, independently of diagnosis, to different short-term trajectories of cognitive performance, and
2) Analyze a possible association between the cluster affiliation and SZ-PRS and further explore whether the polygenic load for SZ differs between the groups identified in 1

Analytic Plan

H10: A higher polygenic risk score for schizophrenia is associated with a poorer cognitive performance in the course of the disease for specific cognitive domains.
H20: A higher polygenic risk score for schizophrenia is associated with a poorer cognitive performance in the course of the disease on dimensional level.
Data will be used from participants who have completed all cognitive tests at every follow-up (FU2-4) based on PsyCourse 3.0 or later.
Cognitive data: TMT A/B, digit symbol test, verbal digit span forward and backward, and the verbal learning and memory test. The results of the MWT-B test of crystallized intelligence will be compared between groups.
Genomic data:
- SZ- and BD-PRS
- Ancestry principal components
Analytic methods:
- Principal component analysis of cognitive data
- FlexMix for the clustering of cognitive trajectories after adjusting for age, study center and education
- ANOVA/?2-tests to compare the different clusters regarding diagnosis, GAF, employment status, symptoms, duration of illness
- Multinomial regression model to test for a possible association between cluster membership and SZ-PRS at several p-value thresholds

Resources needed

Statistical analysis for longitudinal clustering.
Ancestry principal components.