Disorganized speech, language and communication, also called 'formal thought disorder,' is a key part of severe mental illnesses like psychosis and mood disorders. When someone's communication is disorganized, it makes social interactions difficult, increases stigma and affect educational and employment opportunities. However, we do not know much about why this happens. This project, called DIALOG, aims to understand the brain's role in disorganization by studying everyday language use instead of traditional clinical ratings. The study will look at how our brain creates predictions during interactions and how these processes break down in psychosis. This international project also includes experts with personal experience of mental illness. The study will look at speech, thinking patterns, symptoms, and brain waves. The goal of the study is to see if brain waves are disrupted in psychosis, especially in language-related problems. Speech tasks, like describing pictures, talking about a significant event, and telling a story are administered. These tasks will be audio-recorded for analysis. Non-invasive brain imaging technologies such as Magnetoencephalography (MEG) and Magnetic Resonance Imaging (MRI) are utilized. MRI creates images of the brain's structure, while MEG records magnetic activity from neurons, shown as brain waves. The MRI machine uses a large magnet to create images, and MEG captures small magnetic field changes from brain activity. Participants will also undergo clinical and neurocognitive assessments. The study will combine Large Language Models (LLM) applied to speech recordings with large scale participant data from neuroimaging tools (MRI/MEG). The goal of DIALOG is to pioneer a computationally informed, molecular-to systems-level account of disorganisation, identifying the precise mechanisms that can be targeted with novel treatments. This project aims to gather speech and neuroimaging data from Montreal \[100 healthy volunteers and 50 patients with psychosis\], Groningen \[17 synaptic density PET scans\], Cardiff \[600 participants\] and Marburg \[1600 participants\] with schizophrenia, schizoaffective disorder or mood disorders and user acceptability data at Pavia and Melbourne.