Subclinical paranoid beliefs and enhanced neural response during processing of unattractive faces

Abstract

The perception of faces and consequent social inferences are fundamental for interpersonal communication. While facial expression is important for interindividual communication, constitutional and acquired features are crucial for basic emotions of attraction or repulsion. An emotional bias in face processing has been shown in schizophrenia, but the neurobiological mechanisms are unclear. Studies on the interaction between face processing and the emotional state of healthy individuals may help to elucidate the pathogenesis of the paranoid syndrome in psychosis. This study addressed facial attractiveness and paranoid ideas in a non-clinical population. Using functional magnetic resonance imaging (fMRI), we investigated neural activation patterns of 99 healthy subjects during the passive perception of a dynamic presentation of faces with different attractiveness. We found that the perceived attractiveness of faces was linked to the activity of face processing and limbic regions including the fusiform gyrus, amygdala, and prefrontal areas. Paranoid beliefs interacted with perceived attractiveness in these regions resulting in a higher response range and increased activation after the presentation of unattractive faces. However, no behavioral interactions between reported subjective attractiveness and paranoid beliefs were found. The results showed that increased activation of limbic brain regions is linked to paranoid beliefs. Since similar correlations were found in clinical populations with paranoid syndromes, we suggest a dimension of emotional dysregulation ranging from subclinical paranoid beliefs to paranoid schizophrenia.

Keywords: Amygdala; Dynamic Faces; Face Perception; Paranoid Beliefs; SyNoPsis; fMRI.

Fig. 1

Experimental design of the face perception task. (A) An illustration of all stimuli and events with the total number and duration in seconds of the task. Stimuli were presented in a pseudo-randomized order and were split into two MRI scan sessions. (B) All possible 16 facial characteristic category combinations of the dynamic face stimuli at the end of the animation. Dynamic faces varied regarding, attractiveness (unattractive, attractive) and gender (male, female), head movement (up, down), and gaze direction (direct, averted).

Fig 2

Schematic overview of the imaging data analysis. (A1) fMRI data acquisition for two MRI sessions using the face perception task. (B1) Pre-processing of the acquired fMRI data. (B2) 1st level analysis by including dynamic face stimuli, individual face ratings of attractiveness as parametric regressor and 6 motion parameters into the model. (B3) 2nd level analysis for the parametric regressor (attractiveness), applying FWE-correction and extracting masks based on activation clusters. (C1) Extraction of beta values from the GLM for all masks and calculating an average beta value for each participant and region of interest. (C2) Extraction of HRF values from the GLM for all masks and calculating an average HRF response for each participant, attractiveness rating level and ROI. (D1) Using multiple regression analysis to explain the degree of neural modulation (mean beta value) by perceived attractiveness with demographics (age, gender, handedness) and paranoid belief scores (conviction, preoccupation, distress). (D2) Visualization of the average parametric modulation by perceived attractiveness for each ROI. FWE, Family-Wise Error; GLM, General Linear Model; HRF, Hemodynamic Response Function.

Fig 3

Display of interindividual attractiveness ratings of every participant as boxplots. Attractiveness ratings for every participant (x-axis) split by the dynamic face characteristics (unattractive & attractive) are shown and arranged according to attractiveness for the attractive stimuli. Each boxplot represents a participant, and each participant is displayed twice.

Fig 4

Brain activations pattern for the parametric modulation by perceived attractiveness in both hemispheres of three brain areas. Fusiform gyrus (left: x  = −34, y = −62, z = −16; right: x  = 32, y = −64, z = −14). Amygdala (left: x  = −20, y = −6, z = −18; right: x  = 20, y = −6, z = −18) and frontal middle gyrus (left: x  = −36, y = 30, z = 32; right: x  = 40, y = 44, z = 18). Results of the event-related parametric modulation analysis for the fusiform gyrus, amygdala, and middle frontal gyrus are displayed. The lines represent the average neural response by dynamic faces rated from −3 (unattractive) to +3 (attractive).