Subject recruitment and assessment

Patients were recruited from the outpatient psychiatry clinics. People who experienced psychotic symptoms for the first time in life were submitted to the Valencia Clinic Hospital First-Episode Unit for evaluation and treatment. Patients diagnosed with schizophrenia and persistent hallucinations (according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10]) were recruited from the Valencia Clinic Hospital and considered chronic patients. All patients and control participants gave written informed consent to participate in this research, which was approved by the local ethics committee. All subjects, patients, and controls were older than 18 years old.

The inclusion criteria for the patients were as follows: (a) first-episode samples, individuals who had been diagnosed with a psychotic disorder for the first time in their lives within the first 18 months after symptom onset; (b) chronic patient samples, a diagnosis of chronic auditory hallucinations that met the selection criteria for persistent hallucinations [Reference Kay and Opler22]; (c) the ability to speak Spanish fluently; and (d) informed consent for the study signed by the patient.

The exclusion criteria for the patients were as follows: (a) individuals who had previous contact with mental health services for psychosis, (b) evidence of psychotic symptoms precipitated by an organic cause, (c) transient psychotic symptoms resulting from acute intoxication, (d) substance abuse (except tobacco), (e) a metal prosthesis, or (f) claustrophobia. Two weeks after their diagnosis, all of the patients underwent a semistructured interview including sociological information and clinical scales. After the clinical evaluation, they were subjected to an fMRI scan, with a minimum interval of 6 months after their diagnosis.

A sample of HCs matched for age, sex, education level, and work status was recruited from the same geographic areas as the patients. The inclusion criteria for the HCs were the same as those for the patients, except for past or present psychotic symptoms. The exclusion criteria were the same as those for the patients but also included (a) past or present psychotic symptoms or past criteria for psychoactive drug abuse or dependence (except for tobacco) and (b) having a first-degree relative with a history of a diagnosed psychotic disorder. No individual in either group suffered from hearing loss.

The patients’ sociodemographic data are shown in Table 1. There were significant differences in age (0.04), work status (0.032), and coexistence (0.001) between the FEP and chronic groups (0.04).

Table 1.

Sociodemographic characteristics of subjects.

Abbreviation: FEP, first-episode psychosis.

* Tukey HSD (p < 0.05): Control ≠ Chronic patients; FEP ≠ Chronic patients.

The final sample consisted of 85 subjects, which included 31 FEP patients with nonaffective psychotic disorders (a manic episode with psychotic features, schizoaffective disorder, transient psychotic disorders, persistent delusional disorder, bipolar affective disorder with psychotic features, and schizophreniform disorder), 23 chronic patients with schizophrenia and persistent auditory hallucinations, and 31 HCs.

Psychopathological measurements

Patients were assessed by personnel trained in FEP evaluation. Interviews were conducted during the first episode and a few days after the start of treatment (JS). All FEP patients underwent several psychopathological measurements, including the Clinical Global Impression [Reference Haro, Kamath, Ochoa, Novick, Rele, Fargas, Rodríguez, Rele, Orta, Kharbeng, Araya, Gervin, Alonso, Mavreas, Lavrentzou, Liontos, Gregor and Jones20], to provide a global rating of illness severity, improvement, and response to treatment; the Global Assessment of Function [Reference Jones, Thornicroft, Coffey and Dunn21], to provide understanding about how well FEP patients could perform everyday activities; and the PANSS [Reference Kay and Opler22], which is a well-established scale that has been used to assess schizophrenia symptoms objectively. The results of the clinical measurements are summarized in Table 2. No significant clinical differences were found in any of these scales between the FEP and chronic groups.

Table 2.

Clinical characteristics of subjects.

Abbreviations: CGI, Clinical Global Impression; GAF, Global Assessment of Function.

fMRI experimental paradigm

An emotional auditory fMRI stimulation paradigm designed by our group [Reference Sanjuán, Lull, Martí-Bonmatí, Aguilar, Gadea and Moratal-Pérez15, Reference Sanjuan, Lull, Aguilar, Martí-Bonmatí, Moratal and Gonzalez16] was employed. The emotional response paradigm was designed to replicate the emotions related to hallucinatory experiences. Eighty-two patients with schizophrenia meeting DSM-IV criteria with AH were selected to choose words of emotional content specific to their psychoses. All patients were administered the PSYRATS, and their discourses about the content of their AH were recorded on tape. The recordings underwent transcription. Hallucinations based on complex phrases or with neutral content were ruled out. A qualitative analysis of their content in five categories was conducted: negative content and imperative tone, insults, imperative tone, and exclamations related to emotional states and positive content. The final paradigm included 13 emotional Spanish words considering their frequency of presentation in the patient’s hallucinations and 13 neutral words with a similar grammar complexity and frequency of use. For the recording procedure, a professional actor from a specialized center was hired to pronounce the words. He pronounced neutral words using a neutral tone and emotional words using an emotional tone but maintained a constant voice intensity (65 dB) [Reference Algarabel González23]. A set of earphones connected by a pair of air tubes to an external audio player was used. Two different fMRI acquisitions, neutral and emotional, were conducted for each subject. The neutral words were presented first using a block design approach. Four blocks of rest and four blocks of stimulation were interleaved and presented to each participant. Each block lasted 20 s. Subjects were informed before the MR acquisition about the experiment and were asked to focus their attention on the words.

Neuroimaging acquisition and data processing

The mean time between clinical treatment and fMRI acquisition was 4 weeks (sd +/− 6). A 3-T magnet (Achieva, Philips Medical Systems, Best, The Netherlands) was used to acquire the fMRI images. A 32-channel head coil and a dynamic echo-planar imaging T2*-weighted sequence were used (repetition time = 2.000 ms; echo time = 30 ms; flip angle 90; slice thickness = 3.50 mm with no interslice gap; acquisition matrix = 128 × 128; voxel size = 1.80 × 1.80 × 3.50 mm), covering the whole brain with 40 contiguous slices. Additionally, a 3-dimensional spoiled gradient-echo pulse T1 image was also acquired for anatomic assessment (repetition time = 13.18 ms; echo time = 7.38 ms; flip angle = 8°; NEX = 1; slice thickness = 1 with no interslice gap; acquisition matrix = 256 × 256; voxel size = 0.90 × 0.90 × 1 mm; 160 slices, thickness = 1 mm). Each fMRI session (emotional and nonemotional) had a total duration of 160 s. After the acquisition, a radiologist and a computer engineer checked the images to ensure data quality. During the acquisition, the patients were under direct observation and they asked about their experiences immediately after the fMRI procedure. Data were then anonymized and transferred to a workstation for postprocessing analysis.

Functional images were realigned to correct for involuntary movement of the patient’s head during the test. Those subjects with translation movements (x, y, z) >1.80 mm (voxel size) were discarded, resulting in a final sample of 31 HCs, 31 FEPs, and 23 chronic patients. Afterwards, a slice timing correction to adjust for temporal delays between the first and the last slices in each fMRI dynamic was applied. Then, a normalization algorithm to standardize the brains to a referenced template (MNI350, Montreal Neurological Institute) was used. Finally, the normalized images were smoothed with a 6-mm FWHM Gaussian smoothing kernel. These analyses were conducted using the Statistical Parametric Mapping SPM software (SPM8, Wellcome Institute, London, United Kingdom) on MATLAB R2017a (The MathWorks, Natick, MA).

Statistics

To describe the fMRI response in each group of patients (or in HCs), we used one-sample t-tests. To assess whether patients with FEP (or with chronic schizophrenia) showed abnormal fMRI responses, we used two-sample t-tests (i.e., comparing their response to the HCs). Finally, to compare the fMRI responses between FEP and chronic schizophrenia, we also used two-sample t-tests. We did not use ANOVAs because we were not interested in knowing if there were differences between the three groups per se; we only used the HC data to provide a reference of normality. To correct for multiple testing, the familywise error (FWE) was applied. Only uncorrected tests for exploratory purposes were applied when no clusters survived the FWE rate correction. Additionally, an extent threshold was applied to consider only those clusters with the minimum number (k) of contiguous voxels as the expected number of voxels per cluster provided by SPM. Finally, the Automatic Area Labeling [Reference Tzourio-Mazoyer, Landeau, Papathanassiou, Crivello, Etard and Delcroix24] atlas was used to report the areas of activation.