2023-07-06

068_ Contributions of Polygenic Risk Scores for Genomic Common EF- and P-Factors to Longitudinal Variation in Cognitive Performance

Research Question and Aims

Executive Functions (EFs) are meta-cognitive abilities that control and coordinate mental processes (e.g., PMID: 34408280). In a previous analysis of PsyCourse data (PsyCourse Proposal 055_ Contributions of Polygenic Risk Scores for Genomic Common EF- and P-Factors to Cognitive Performance), we were able show that polygenic scores (PS) for a Common EF factor (cEF) contribute a sizable amount of variation (up to 4.5%) to cognitive performance, and that this effect is larger in cognitive test with a more pronounced executive component (e.g., the TMT-B). Conversely, PS for a “general psychopathology” p-Factor, negatively correlated to cEF PS, showed an inverse but otherwise similar relationship, with stronger effects in cognitive tests tapping the executive domain. Here, we aim to extend this research to the longitudinal course, beyond the baseline data previously analyzed. Since cognitive test performance generally improves over time/measurement points (PMID: 35232513) and, at least in neurotypic PsyCourse participants (PMID: 34247186), appears to reflect commonly assessed characteristics of cognitive tests, such as familiarity or strategy, we do not expect this longitudinal relationship to be influenced by PS for cEF or the p-factor. Also, an earlier twin study (PMID:26619323) found EFs to be quite stable over time, mainly attributable to high genetic correlations across time. We would like to study all longitudinally measured cognitive tests available in PsyCourse. As in the previously proposal, we also plan to include a latent phenotypic EF factor into the analysis. Exploratory analyses will also be carried out taking case/control status into account.

Analytic Plan

Where possible, we will use linear mixed models (implemented by the R lmer package) to determine the influence of PS and time on the cognitive phenotypes. The basic model syntax will include ancestry principal components and age and sex:

Cognitive Variable ~ v1_age + v1_age2 + v1_sex + PC1 + PC2 + PC3 + PC4 + time*PS
For each cognitive test, we expect a significant time effect. We also expect significant effects of both the cEF and p-factor PS, like the baseline analyses carried out in the previous proposal. Importantly, our hypothesis is that the time × PS interaction will be non-significant in all cognitive variables. For an analysis of the latent phenotypic EF factor, we will use a longitudinal extension of Confirmatory Factor Analysis, such as a Latent Growth Model, implemented by the R lavaan package.

Resources needed

v1_stat
v1_center
v1_sex
v1_ed_status
v1_scid_dsm_dx
age
gsa_id
Antidepressants
Antipsychotics
Mood_stabilizers
Tranquilizers
Other_psychiatric
curr_paid_empl
panss_sum_pos
panss_sum_neg
panss_sum_tot
idsc_sum
cgi_s
gaf
nrpsy_mtv
nrpsy_tmt_A_rt
nrpsy_tmt_A_err
nrpsy_tmt_B_rt
nrpsy_tmt_B_err
nrpsy_dgt_sp_frw
nrpsy_dgt_sp_bck
nrpsy_dg_sym
nrpsy_vlmt_check
nrpsy_vlmt_corr
nrpsy_vlmt_lss_d
nrpsy_vlmt_lss_t