Impaired Facial Emotion Recognition in Individuals at Ultra-High Risk for Psychosis and Associations With Schizotypy and Paranoia Level

Abstract

Background: Patients with schizophrenia and individuals at ultra-high risk for psychosis (UHR) have been reported to exhibit impaired recognition of facial emotion expressions. This impairment has involved both inaccuracy and negative bias of facial emotion recognition. The present study aimed to investigate whether UHR individuals display both types of impaired facial emotion recognition and to explore correlations between these impairments and schizotypy, as well as paranoia levels, in these individuals. Methods: A total of 43 UHR individuals and 57 healthy controls (HC) completed a facial emotion recognition task consisting of 60 standardized facial photographs. To explore correlations, we assessed schizotypy using the Revised Physical Anhedonia Scale and Magical Ideation Scale and paranoia level using the Paranoia Scale and persecution/ suspicious item of the Positive and Negative Syndrome Scale in UHR individuals. Results: Compared with HC, UHR individuals exhibited less accuracy for facial emotion recognition (70.6% vs. 75.6%, p=0.010) and a higher rate of "fear" responses for neutral faces (14.5% vs. 6.0%, p=0.003). In UHR individuals, inaccuracy was significantly correlated with schizotypy scores, but not with paranoia level. Conversely, "disgust" response for neutral faces was the only fear response correlated with paranoia level, and no threat-related emotion response correlated with schizotypy scores. Discussion: UHR individuals exhibited inaccuracy and negative bias of facial emotion recognition. Furthermore, schizotypy scores were associated with inaccuracy but not with negative bias of facial emotion recognition. Paranoia level was correlated with "disgust" responses for neutral faces but not with inaccuracy. These findings suggest that inaccuracy and negative bias of facial emotion recognition reflect different underlying processes, and that inaccuracy may be a vulnerability marker for schizophrenia.

Facial emotion recognition and schizotypal traits: A systematic review of behavioural studies

“…As regards studies focusing on total schizotypy scores, findings strongly indicate that high schizotypal individuals have poorer performance in FER tasks (Abbott & Byrne, 2013; Germine & Hooker, 2011; Statucka & Walder, 2017). This is in accordance with findings of studies on UHR for psychosis individuals (Seo et al, 2020; Wölwer et al, 2012), SPD (Dickey et al, 2011; Waldeck & Miller, 2000) as well as schizophrenia or schizoaffective disorder (Edwards et al, 2002; Kohler et al, 2010; Savla et al, 2013) patients. The lower accuracy of high schizotypal individuals in the identification of anger, has also been reported in SPD (Waldeck & Miller, 2000; Wickline et al, 2012) and schizophrenia (Won et al, 2019) patients.…”

“…In schizophrenia, it is well documented that FER is severely impaired, with meta-analyses reporting large effect sizes when comparing schizophrenia patients with controls (Kohler et al, 2010;Savla et al, 2013). Impairments (a) remain stable during the course of the illness (Comparelli et al, 2013;Green et al, 2012), (b) are observed in early onset and first-episode schizophrenia patients, (Barkl et al, 2014;Healey et al, 2016) as well as in unaffected first degree relatives of schizophrenia patients (Lavoie et al, 2013) and Ultra-High-risk (UHR) for psychosis individuals (Seo et al, 2020;Van Rijn et al, 2011) and (c) are also found in Schizotypal Personality Disorder (SPD) patients (Dickey et al, 2011;Waldeck & Miller, 2000). Based on these findings, FER could be a trait marker/endophenotype for schizophrenia, reflecting the risk for transition into the disorder (Allott et al, 2014;Martin et al, 2020).…”

Emotional anticipation for dynamic emotional faces is not modulated by schizotypal traits: A Representational Momentum study

“…Despite of this, there was a positive correlation between cognitive-perceptual SPQ factor and ambiguity threshold when faces were presented statically, with higher self-report of "positive" schizotypy (e.g., suspiciousness, ideas of reference, unusual perceptual experiences) being correlated to an ambiguity threshold slightly biased towards a happier expression (i.e., more happy cues are needed to judge a stimulus as ambiguous). This fits previous descriptions of a threat-related emotion bias for neutral faces in people at ultra-high risk (e.g., Seo et al, 2020;van Rijn et al, 2011) and with high schizotypal traits (e.g., Brown & Cohen, 2010), which also aligns with the hypothesis that psychotic experiences, namely persecutory delusions, are associated with an aberrant threat perception in ambiguous events (see Green & Phillips, 2004). Importantly, the fact that there a positive correlation between cognitive-perceptual SPQ and ambiguity threshold, but a null effect of schizotypy in RM, could suggest that "positive" schizotypy is related to an anger bias for ambiguous faces solely when it is explicitly required to discriminate the emotion, whereas this association ceases to exist when more automatic (instead of emotion recognition) mechanisms are recruited.…”

Comparison of Subjective Facial Emotion Recognition and “Facial Emotion Recognition Based on Multi-Task Cascaded Convolutional Network Face Detection” between Patients with Schizophrenia and Healthy Participants

“…Introducing patients with schizophrenia to robotic technology has proven difficult due to the following reasons: (1) these patients have communication-related issues and (2) unlike healthy people, they have a poor perception of facial emotional expressions [ 9 , 10 ]. This difficulty could also partly be attributed to the fact that robot-assisted psychiatric rehabilitation is not covered by health insurance.…”