This research involves the examination of speech, thinking patterns and symptoms and recordings from brain's activity in patients with schizophrenia. The reserchers are investigating if the electrical waves (recorded indirectly using the resulting magnetic field around our head) are disrupted in psychosis, especially in the language related problems seen in this disorder. To examine language, participants will be asked to complete a speech tasks using the DISCOURSE speech elicitation script in which they will be requested to share their thoughts. For this purpose, the participants will be shown some pictures and asked to describe what they see, discuss a significant event from their life and tell a story in their own words. The description of the pictures and other responses to the tasks will be audio-recorded for later transcription into a written format.

The researches will examine the brain regions using Magnetoencephalography (MEG) and Magnetic Resonance Imaging (MRI). MEG and MRI are both non-invasive functional brain imaging technologies. MRI collects structural information and creates images of the brain while MEG collects magnetic activity from neurons, visualized as brain waves. The MRI machine uses a large magnet to help create images while MEG captures the fluctuations in the small magnetic fields resulting from brain's activity. Participants will also undergo a clinical and neurocognitive assessments.

This project design involves 50 individuals who will take part in the study; 25 affected and 25 healthy volunteers aged 18 to 50 years.

Schizophrenia is a common and severe mental illness affecting millions of people worldwide. With a lifetime prevalence of about 1%, it is estimated to generate annual costs of $10 billion in Canada alone, with cognitive and functional impairments as major contributors. One of the main and most debilitating symptoms of schizophrenia is related to the disorganization and impoverishment of speech, also known as Formal Thought Disorder (FTD). The presence of linguistic disorganization and impoverishment seen in FTD is recognized as an early marker of disease; it can be present in otherwise healthy adolescents or those in prodromal or at-risk stages and predict the onset of a first psychotic episode. In the context of a well-established diagnosis, it serves as an indicator of decompensation severity and disease prognosis and is associated with occupational and social functioning. Yet, FTD remains poorly understood and largely understudied, hindered by difficulties in clinical assessment. The recent emergence of objective computer-based analysis of speech (Natural Language Processing) to generate markers of disruptions in syntax, meaning and parts-of-speech has revived the field, promising to offer novel insights into how communication becomes disrupted in the presence of psychosis.

Pathophysiological enquiries into schizophrenia to date suggest that it is a disorder of dysconnectivity in both spatial and temporal dimensions, with both being related to neurochemical aberrations such as subcortical dopamine excess. In other words, the relationship among various parts of the brain constituting networks (a spatial dimension) appears to be notably affected. This occurs in conjunction with a disruption in the moment-to-moment physiological continuity, or temporal irregularity, in neuronal activity. The spatial disconnect is well captured using functional magnetic resonance imaging (fMRI) while the temporal irregularity can be studied using oscillatory patterns in the electroencephalography (EEG). In more recent times, the capability to concurrently study both spatial and temporal dysconnectivity has emerged from the use of magnetoencephalography (MEG), while non-invasive proxy measurements of dopamine function are feasible from short anatomical scans.

In the era of precision medicine, individual characterisation of the disease can support both diagnosis and treatment response. Magnetoencephalography (MEG) has shown to be a powerful tool to study the underpinnings of verbal production using task-free and task-based paradigms. Recent data published by the Baillet lab have shed light on novel approaches to spectral brain mapping and speech impairment quantification to characterize the functional neural pathology contributing to speech impairment in patients with Parkinson's disease. A previous study has demonstrated the feasibility of identifying individual specific patterns, termed 'brain fingerprints', in healthy participants using MEG for as short as 30 seconds. Recovering specific neural patterns from brief non-invasive recordings is very promising for the study of psychosis, however, no prior studies have investigated the spectral and connectome fingerprint of patients with schizophrenia. In addition to characterizing the speech abnormalities prevalent in patients with schizophrenia, identifying individual markers of temporal connectivity will enable us to understand the mechanism behind disorganised language in psychosis.

It has been shown that healthy individuals exposed to asynchronous audiovisual stimuli (i.e., out of the expected similar timing) undergo a rapid neural tuning or resetting (temporal recalibration). This relevant physiological mechanism is linked to our response to dynamic changing environments, such as speech comprehension. Previous research has shown that multisensory integration is impaired in schizophrenia; however, no previous studies have tried to dissect the temporal and spatial oscillatory mechanisms through which this phenomenon occurs and whether the degree of desynchrony correlates to psychotic symptoms.

Previous studies have also identified synchronized neural oscillatory activity as one of the physiological mechanisms underlying the hierarchical and integrative process of language production in healthy state. In line with these findings, research studies have suggested that beta oscillatory activity are involved in language production. Brain oscillations in the alpha and beta frequency ranges are considered as indices of sensorimotor integration in speech perception and production. Previous findings in healthy participants with high schizotypy (a trait that is related to the construct of schizophrenia) and patients with schizophrenia have led to the hypothesis that beta oscillatory dysfunction are key contributors to the mechanism of symptom formation and persistence in schizophrenia. Beta oscillations are thought to play a key role in the maintenance of prior context that is essential for coherent motor action and speech production; when this is disrupted, speech may contain words that are 'out-of-context' i.e., lower predictability for a given context . In particular, beta-band oscillations appear to 'carry the predictions' about the ongoing context, thus mediating top-down interactions between motor and frontotemporal language system during auditory listening tasks , by suppressing cortical responses to predicted sensory input. Inappropriate language area activity (resulting in disorganised speech) can thus be expected to arise from beta band deficits in schizophrenia.

Two key proofs are needed to further advance the hypothesis of beta deficits in schizophrenia: (1) Demonstrating beta-oscillatory deficits in patients with an established diagnosis of schizophrenia, especially in the language network that is thought to underlie FTD, and (2) Relating the specific patterns of temporal connectivity (spectral connectome) and irregularity (beta-oscillatory deficits) in the brain of affected individuals to objective measures of thought disorder using computerised analysis of speech.

PURPOSE The primary purpose of this study is to demonstrate that deficits in beta-band power in the motor cortex (elicited via a simple motor task of responding to a tone) and spectral similarity of language network (at rest) occur in schizophrenia. In patients with schizophrenia, the investigators expect these deficits to correlate with the degree of disorganisation measured as semantic predictability/coherence (by analysing recorded speech analysis).

HYPOTHESES Hypothesis 1: Beta-band oscillatory power (rebound from baseline after task execution) during a simple motor task in response to audiovisual stimuli is affected in schizophrenia.

Hypothesis 2: Across both patients and controls, the beta-rebound after motor task will correlate with coherent speech production (context-based semantic coherence or predictability) measured offline.

Hypothesis 3: Among patients with schizophrenia, the spectral maps for beta oscillatory frequency will show the greatest deviation from healthy controls in the frontotemporal language network.

The researchers also aim to explore if resting-state MEG connectomes across multiple frequency bands in schizophrenia display deviations from unaffected healthy subjects, if multisensory integration through rapid temporal recalibration is schizophrenia differs from that of healthy individuals, and if cortical structure (by measuring cortical thickness, and folding pattern) and subcortical profile of dopamine turnover (by measuring neuromelanin) relate to the MEG spectral deviations. The correlations obtained through the above-mentioned exploratory hypothesis will serve to estimate likely effect sizes to plan future larger studies.

1. Patients:

   Inclusion Criteria:

   • Clinical diagnosis of schizophrenia as per DSM-5 criteria.
   • Aged between 18 and 50 years.
   • In a stable phase of the illness (no change in Global Functioning > 10 units within 1 month before the MEG scan assessed using the Global Assessment of Functioning Scale).

   Exclusion Criteria:

   • Inability to speak English or French.
   • History of head trauma, major neurological disorders (e.g., epilepsy), somatic disorders with neurological complaints (e.g., multiple sclerosis), or intellectual disabilities.
   • Diagnosis of substance use disorder (current or within the last year) as per DSM-5.
   • Unsuitable for MRI due to safety concerns (e.g., metal in the body, pregnancy, tinnitus).

2. Healthy Controls

Inclusion Criteria:

• Healthy volunteers aged between 18 and 50 years.
• Matched to patients for age (within 2 years), sex, and parental occupational status (rated by NS-SEC).

Exclusion Criteria:

• Inability to speak English or French.
• History of head trauma, major neurological disorders (e.g., epilepsy), somatic disorders with neurological complaints (e.g., multiple sclerosis), or intellectual disabilities.
• Diagnosis of substance use disorder (current or within the last year) as per DSM-5.
• Personal or family history of psychotic disorder.
• Unsuitable for MRI due to safety concerns (e.g., metal in the body, pregnancy, tinnitus).

Patients will be recruited from 3 institutions; 1. Douglas Mental Health University Institute 2. Lakeshore General Hospital and 3. McGill University Health Centre. Initial contact and information about this study will be communicated by clinicians who are members of the patient's mental health care team at these locations.

Healthy Control participants will be recruited from the Greater Montreal Area using online study advertisements and research study posters placed on bulletin boards at the three Mental Health Institutions

The researchers propose to use a clinical linguistic archiving system called the TalkBank for sharing de-identified speech data (https://www.talkbank.org/). TalkBank registry/database is located at Carnegie Mellon University, Pittsburgh, Pennsylvania-USA. Only the de-deidentified audio (and transcriptions) data will be provided. These are audio files in wave format and text files, with no identifiable patient information or tags. The de-deidentified data will be coded and anonymized before being transferred.

Deidentified MEG data will be made available to qualified researchers via the Open MEG Archive (https://omega.bic.mni.mcgill.ca; Dr S. Baillet, co-principal investigator). This repository already contains resting state MEG data from healthy subjects available for re-use and comparative analysis. Open MEG Archive will not be provided with any study participant dataset information.

Data will be made available after all of the intended recruitment and the curation of data is completed. This will be no later than 2 years after the date of the last recruited subject. There are no plans for data destruction of the study information stored (de-identified audio and transcriptions) within TalkBank.

Access to the speechbank database will be password protected and restricted to scientific members of an international consortium of researchers who collaborate on this effort (the DISCOURSE group)