Inclusion criteria for both groups were 1) age between 20 to 60 years old 2) Thai literate 3) had at least of 4 years of formal education 4) no major neurological diseases (e.g., stroke), and 5) no history of substance dependences (except for smoking). We recruited 50 patients diagnosed with schizophrenia, according to the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) criteria by our psychiatrists, at an outpatient mental health clinic, Thammasat University Hospital, Thailand from 1 June 2016 to 31 May 2017. All the schizophrenia patients had received medical treatment for at least 12 weeks and were clinically stable, defined by no significant changes in medical treatment and psychotic symptoms for the last 12 weeks. Schizophrenia patients were matched 1:1 for comparable education with a control group. These 50 normal controls were recruited from relatives or spouses of patients (78%) or patients attending the family medicine clinic (22%: health checkup 12% and other diagnoses 10%). They were free of any history of major psychiatric disorder (i.e., schizophrenia, bipolar disorder, major depressive disorder, and intellectual disabilities).

A paper version of the Eyes test was used which consists of 36 photos of eyes expressing a variety of complex emotional and mental states. For each photo, a participant has to decide which one of four words best describes what the person in the picture was feeling. Responses are scored 1 or 0 for correctness. Scores range from 0 to 36, with higher scores indicating better affective theory of mind [2,10]. Following Harkness et al., [9] the test items were classified into the three emotional valences of negative, positive and neutral. There were 8 items for positive valence (playful, fantasizing ×2, thoughtful, friendly, interested, flirtatious, and confident), 12 items for negative (upset, worried, regretful, accusing, doubtful, preoccupied, defiant, hostile, cautious, distrustful, nervous, and suspicious), and 16 items for neutral (desire, insisting, uneasy, despondent, preoccupied, cautious, skeptical, anticipating, contemplative, decisive, tentative, pensive, interested, reflective, serious, and concerned).

Characteristics and clinical data were collected from participants; medical records included: age, gender, educational level, duration of illness, number of psychotic episodes, and medications. The doses of antipsychotics were presented as haloperidol equivalents according to Inada and Inagaki [11]. The severity of symptoms in schizophrenia patients were rated prior to the Eyes test using the Scale for the Assessment of Positive Symptoms (SAPS) and the Scale for the Assessment of Negative Symptoms (SANS) by a psychiatrist. The SAPS and SANS have global rating scores of 0-20 and 0-25, respectively. Higher scores indicate more severe symptoms [12-15]. Subsequently, the Eyes test was administered by independent research assistants. All participants gave written inform consent, and the study was approved by the Human Ethics Committee of Thammasat University (MTU-EC-ES-6-043/59).

Data were analyzed with STATA v14. Performance of the Eyes test and the sub-score of emotional valences (positive, negative and neutral) were presented by percentages of the correct answers. Between groups’ differences were analyzed with independent t-test, and Cohen’s effect sized were calculated. Multivariable linear regression analysis was also carried out to examine and adjust the effects of various characteristics.

One strength of our study is we have used the Eyes test to examine a greater variety of subtle emotional states versus earlier research. We also included healthy controls that matched the schizophrenia patients’ level of education, which is usually considered as an important predictor of the Eyes test performance [5].

In terms of limitations, our patients are not homogeneous in episodes of psychosis; 48% of them were first-episode psychosis, while the rest were multiple. These two groups might demonstrate a different level of deficits. Secondly, all schizophrenia patients received antipsychotic drugs, so the effects of medication possibly influenced task performance. However, since our inclusion criteria was clinically stable schizophrenia who received treatment for at least 12 weeks, antipsychotic medication was unavoidable. Lastly, the proportion of females in our control group was significantly higher than in the patient group. A meta-analysis of the Eyes test in a healthy population demonstrated a trend of female superiority; therefore, there was a possibility that the higher proportion of females in the control group of our study might magnify the result. However, the difference between gender was small, with an effect size of only 0.17 [22]. Moreover, studies comparing schizophrenia patients and healthy controls usually found that there were no significant effects of gender for the Eyes test performances [4,5,23]. This could be because the effect of disease status is much stronger and might override the effect of gender.

In conclusion, there is a variability in the impairments of affective theory of mind according to emotion valence. The deficits seem to specific for only negative and neutral emotions, but not positive emotions. Further research is needed to determine why exactly the impairment of positive emotion detection does not occur in schizophrenia patients.