Individual differences in aesthetic engagement are reflected in resting-state fMRI connectivity: Implications for stress resilience

Abstract

Objective: Individual differences in aesthetic engagement-the propensity to be moved by art, nature, and beauty-are associated with positive health outcomes, as well as stress resilience. The purpose of the current study was to identify potential neural substrate mechanisms underlying individual differences in aesthetic engagement and reported proneness to aesthetic chill.

Methods: Data from the Human Connectome Project (HCP) 1200 Subjects Release were utilized. Resting-state fMRI connectivity was extracted for 361 regions in the brain including cortical, subcortical and cerebellar regions for each participant, using participant-specific segmentation and parcellation of subcortical gray matter nuclei and a network-based statistics analytical approach. The Aesthetic Interests subcluster of the Openness to Experience scale (NEO-Five Factor Inventory; NEO-FFI) was used to characterize individual differences in aesthetic engagement and chill.

Results: Participants reporting higher aesthetic engagement, particularly proneness to aesthetic chill responses, exhibited significantly higher connectivity between the default network and sensory and motor cortices, higher connectivity between the ventral default and salience networks, and decreased connectivity between the cerebellum and somatomotor cortex.

Conclusions: Current findings suggest that greater integration of the default mode network, involving processing of internal narrative, with neural representations of sensory perception and salience detection may be a mechanism underlying individual differences in aesthetic engagement. Thus, these individual differences may reflect general integration of environmental perception with internal emotional experience, which in turn may facilitate comfort with novelty, self-regulation, and positive adaptation to potentially stressful experiences.

Keywords: Aesthetics; Openness to experience; Resting state functional connectivity; Stress resilience; fMRI.

Figure 1:

Functional network connectivity between pairs of 17 brain networks and personality metrics. Color scale shows t-statistic, with significant network pairs showing p<0.05, corrected, for full sample analyses and p<0.05, uncorrected, for split sample analyses (bottom). Age, sex, and mean head motion were included as subject level covariates in all analyses. Black squares show connections in common across samples. Network names are descriptive based on functional regions present. The networks are listed in the same order as in Yeo et al. (2011).

Figure 2:

Partial correlation of functional network connectivity between 5 pairs of networks, personality metrics, and twin status. Color scale shows partial correlation, with significant associations showing p<0.05, corrected. Age, sex, and mean head motion were included as subject level covariates in all analyses and personality factors (left) and twin status (right) were additionally included. Functional connectivity results were included one at a time in a model with other factors because of similarity of functional connectivity results to each other across subjects.

Figure 3:

Functional connectivity between pairs of 361 gray matter ROIs was compared to sensitivity to chills, with colored squares showing connections with p<0.01. Only connections belonging to a maximal connected component are colored above, and all connections are shown with p<0.05 for female and male cohorts below. Color scale represents t-statistic for functional connectivity vs. chills, with age and head motion as subject-level covariates.

Figure 4:

Clustering of maximally connected component. For all ROIs included in the maximally connected component showing differences in connectivity with respect to aesthetic chills, a clustering was performed that grouped ROIs showing similar t-statistics of connectivity vs. chills with respect to the other ROIs. The brain regions comprising each of the 8 nontrivial components are shown in the images below, and t-statistic of functional connectivity vs. chills for each pair of ROIs is shown in the pseudocolor plot above, grouped by component.