Trait positive and negative emotionality differentially associate with diurnal cortisol activity

Abstract

Inter-individual variability in metrics of hypothalamic-pituitary-adrenocortical (HPA) activity, such as the slope of the diurnal decline in cortisol, cortisol awakening response (CAR), and total cortisol output, have been found to associate inversely with trait ratings of extraversion and positive affect (E/PA) and positively with neuroticism and negative affect (N/NA) in some, but not all, investigations. These inconsistencies may partly reflect varied intensity of cortisol sampling among studies and reliance on self-rated traits, which are subject to reporting biases and limitations of introspection. Here, we further examined dispositional correlates of HPA activity in 490 healthy, employed midlife volunteers (M age=43 years; 54% Female; 86% white). Trait ratings were requested from participants and 2 participant-elected informants using the Positive and Negative Affect Schedule (PANAS) and Extraversion and Neuroticism dimensions of NEO personality inventories. CAR was assessed as percent increase in cortisol levels from awakening to 30min after awakening; and the diurnal slope and total output of cortisol [Area Under the Curve (AUC)] were determined from cortisol measurements taken at awakening, +4 and +9h later, and bedtime, across 3 workdays. Structural equation modeling was used to estimate multi-informant E/PA and N/NA factors. We used 3days of measurement as indicators to model each of the three latent cortisol factors (slope, CAR, and AUC). With the two latent emotionality and three latent cortisol indices included there was good fit to the data (χ2(200)=278.38, p=0.0002; RMSEA=0.028, 90% CI=0.02–0.04; CFI/TLI=0.97/0.96; SRMR=0.04). After controlling for covariates (age, sex, race), results showed higher latent E/PA associated with a steeper diurnal slope (Standardized β=−0.19, p=0.02) and smaller CAR (Standardized β=−0.26, p=0.004), whereas N/NA did not associate with any cortisol metric (Standardized β’s=−0.12 to 0.13, p’s=0.10 to 0.53). These findings suggest that positive emotionality may be more closely associated with indices of diurnal cortisol release than negative emotionality.