Effects of depression status and genetic risk on emotional word processing in a large cohort

Depression has been linked to a negative bias in emotional perception, with patients frequently showing more negative valence and higher arousal ratings of emotional stimuli, such as words and images. Nevertheless, studies in large depression cohorts which also incorporate longitudinal and genetic data are lacking. We investigated the effects of depression on valence and arousal ratings of emotional words at two time points in 1392 participants of the BiDirect Study, a German cohort enriched in patients that were hospitalized due to an acute depressive episode at the time of recruitment. Data from the baseline (y0) and 2nd-wave assessments (three-year interval, y3) were used to test for associations of valence and arousal ratings of words with depression status, depressive symptoms and polygenic risk score for major depressive disorder (MDD-PRS). We found negative and positive effects of CES-D score-defined depression status and symptoms on valence and arousal ratings, respectively, at both time points. Changes in depressive symptoms and remission status at y3 related to valence ratings. MDD-PRS showed a negative effect on valence ratings (beta = -0.64, p = 0.0004). We further explored the latter finding by testing associations of valence ratings with genetic variants previously linked to depression in genome-wide association studies. This exploratory analysis suggested potential candidate genes (e.g., AREL1, LINGO1, LTBP2, PROX2, YLPM1), as well as links to neuroticism and anxiety. In conclusion, using a large cohort, our study confirmed that depression relates to more negative emotional valence and suggested that genetic risk of depression may confer certain susceptibility to cognitive bias.