Below, we describe the characteristics of our participants, pareidolia tests, and statistical analyses. The study was approved by the Hungarian Scientific and Research Committee of the Medical Research Council ethics board (Budapest, Hungary).
Participants
We enrolled 50 patients with schizophrenia, 50 patients with type I bipolar disorder with a history of psychotic symptoms, and 50 control volunteers without any history of psychiatric disorders. Participants were matched for age, sex, education, Intelligence Quotient (IQ), and general psychosocial functions (Table 1). The study was coordinated in the Nyírő Gyula National Institute of Psychiatry and Addictions and was approved by the Hungarian Scientific and Research Committee of the Medical Research Council ethics board (Budapest, Hungary). All participants gave written informed consents. The inclusion criteria are as follows: Diagnostic and Statistical Manual of Mental Disorders (DSM-5) diagnosis of bipolar disorder or schizophrenia, ability and willingness to participate, age between 18-65 years, and lack of acute psychosis. All of our participants were outpatients. All patients lived in the community and were in clinical remission according to the Andreasen-criteria for schizophrenia (Andreasen et al., 2005) and to the Systematic Treatment Enhancement Program for Bipolar Disorder criteria for bipolar disorder (Perlis et al., 2006).
The exclusion criteria were as follows: neurological disorders and other general medical conditions affecting the central nervous system, evidence of head injury, electroconvulsive therapy, and psychoactive substance misuse confirmed by clinical history or by a urine test. In total, we have excluded only 5 patients with comorbid substance and alcohol misuse. Forty-seven patients with schizophrenia and 40 patients with bipolar disorder received either second-generation antipsychotic medications (amisulpride, olanzapine, quetiapine, and risperidone) or third-generation medication (e.g., aripiprazole) at the time of testing. In total, 32 schizophrenia patients, 34 bipolar disorder patients, and 30 controls regularly smoked tobacco. The chlorpromazine-equivalent doses, calculated by a standard method (Leucht et al., 2015), are shown in Table 1. Twelve patients with schizophrenia and 45 patients with bipolar disorder also received mood stabilizers (lithium, valproate, or lamotrigine).
Pareidolia Tests and Clinical Measures
The pareidolia tests were based on the exact adoption of a previously published protocol (Kaji, Kitamura, & Kitagawa, 2015; Mamiya et al., 2016; Uchiyama et al., 2012; Yokoi et al., 2014). The scene test consisted of 10 blurred images of natural scenes that frequently produced illusory face responses in a previous study (Mamiya et al., 2016; Uchiyama et al., 2012). The task goal was to point to and describe the objects on each image in as much detail as possible. The noise pareidolia test included 32 black and white images consisting of visual noise (spatial frequency: 1/f3) and 8 images depicting human faces. Participants were asked to respond whether a face was present on these images and to point to the location where they saw the face. The maximum exposure time was 60 sec in the scene task and 30 sec in the face task. Participants did not receive feedback on the appropriateness of their responses, and they were not informed that in the face task only noise was presented. Illusory responses were defined as answers when observers falsely identified objects that were not on the images in the scene task (maximum illusory score: 10), and the number of noise images in which they reported the presence of a face (maximum illusory score: 32). We also calculated the total pareidolia score for each task (the sum number of images with illusory responses in the scene and noise tests). The responses were scored by two independent raters with an excellent congruence (kappa > 0.9).
We used the following instruments for clinical evaluation: Structured Clinical Interview for DSM-5 Disorders—Clinician Version (SCID-5-CV) (First, Williams, Karg, & Spitzer, 2016), Positive and Negative Syndrome Scale (PANSS) (Kay, Fiszbein, & Opler, 1987), Hamilton Rating Scale for Depression (HAM-D) (Hamilton, 1960), Young Mania Rating Scale (YMRS) (Young, Biggs, Ziegler, & Meyer, 1978), and the World Health Organization Disability Assessment Schedule (WHODAS 2.0) of the DSM-5 (APA, 2013).
General intellectual and cognitive functions were measured with the Wechsler Adult Intelligence Scale - IV (WAIS-IV) (Wechsler, 2008). The DSM-5 structured clinical interview and the rating scales were administered by trained and supervised clinical psychologists or psychiatrists. Below, we provide a brief description of WHODAS 2.0, HAM-D, and PANSS scales
WHODAS 2.0.
This instrument enables assessment of health and disability in 6 domains (cognition, mobility, self-care, interacting with other people, life activities, participation in communities). The administration time is short (5 to 20 minutes), and WHODAS 2.0. is valid in clinical and general populations across cultures. The concept of WHODAS 2.0. is based on ICF (International Classification of Functioning, Disability and Health) principles.
HAM-D
We used the structured version of the 17-item HAM-D to assess the severity of depressive symptoms (e.g., depressed mood, feelings of guilt, suicide, and insomnia). Each item was rated on a 3-point (items 4-6, 12-14, 17) or 5-point Likert-scale (0 – absent or no difficulty). A score in the range of 0–7 is normal. Patients scoring 20 or higher regularly require clinical attention.
PANSS
The patient is scored from 1 to 7 on 30 items classified according to positive, negative, and general symptoms. The positive symptom scale contains 7 items (minimum score = 7, maximum score = 49), such as delusions, conceptual disorganization, hallucinations, excitement, and grandiosity. The negative scale also consists of 7 items (e.g., blunted affect, emotional withdrawal, and poor rapport). Finally, the general psychopathology scale consists of 16 items (minimum score = 16, maximum score = 112) (e.g., somatic concerns, anxiety, guilt, and tension).
YMRS
The scale includes 11 items. Four items are graded on a 0 to 8 scale (irritability, speech, thought content, and disruptive or aggressive behavior), whereas seven items are graded on a 0 to 4 scale (e.g. elevated mood, increased motor activity, and sexual interest). A score of 20 or higher indicate severe mania.
Interview for the assessment of anomalous visual and auditory experiences.
We used the Structured Interview for Assessing Perceptual Anomalies (SIAPA) (Bunney et al., 1999; Kiss, Fabian, Benedek, & Keri, 2010). The SIAPA focuses on three aspects of anomalous perceptual experiences: sensory intensity (hypersensitivity), inundation or flooding, and selective attention to external stimuli on a scale of 0 (absent) – 4 (pervasive). The interview begins with open-ended and then structured questions regarding subjective experiences in each sensory modality. For example, to assess hypersensitivity in the auditory modality, the following questions are presented: “Have you ever had the feeling or sensation that sounds were particularly loud? Or louder than usual? Or that your sense of hearing was particularly keen or sensitive? Or that your ears were picking up the slightest detail of sounds?" To evaluate inundation and flooding, participants are asked: "Have you ever had the experience or felt like you were being flooded or inundated by sounds? Or that you couldn't block out sounds? Or that it seemed as if your ears were picking up everything going on around you?" Finally, questions for selective attention are as follows: "Have you ever had the experience or felt like you couldn't pay attention to one sound, or a conversation, because of interference from other sounds, like background noise? Do you find that your attention is captured by irrelevant sounds, like traffic noises, even though they are of no interest to you?" (Bunney et al., 1999). Similarly, in the visual modality, participants are asked whether lights seemed much brighter, colors were unusually vivid, the environment was bothersome, and whether they were overwhelmed by multiple objects in the scene and could not attend to one of many simultaneous visual inputs. We observed a good congruence between two independent raters (kappa > 0.7), and the internal consistency was good (Cronbach alpha > 0.8). The SIAPA scores correlated with objective psychophysical measures of sensory perception (Kiss et al., 2010).
Statistical Analysis
The STATISTICA 13.1 (Tibco, Palo Alto) software package was used for data analysis. First, we tested data distribution and homogeneity of variance with Lilliefors and Levene’s tests, respectively. Measures with normal distributions were entered into analyses of variance (ANOVAs) and two-tailed Student’s t tests. Dichotomous variables were analysed with chi-square tests. The SIAPA and the pareidolia scores were not normally distributed (p < 0.01), and therefore we used Kruskal-Wallis analyses of variance (ANOVA) followed by multiple comparisons for mean ranks tests. Spearman’s rank correlation coefficients were calculated between the pareidolia scores and the clinical measures. We used a receiver operating characteristic (ROC) analysis to test the sensitivity and specificity of the pareidolia test to differentiate schizophrenia from bipolar disorder or controls. The schizophrenia-control and schizophrenia- bipolar disorder differentiation was also investigated with discriminant function analyses. The level of statistical significance was set at alpha < 0.05, corrected for multiple comparisons with the Bonferroni method.