Understanding and producing language depends on smooth, two-way communication between brain regions in the frontal and temporal lobes. In people with disorganised thinking, this flow of information may be disrupted. While traditional brain scans can show which areas are active, they don’t reveal how signals move between regions or how efficiently they’re processed. To address this, we use advanced computational models that estimate effective connectivity—the direction and strength of information flow in the brain.

By combining these models with resting-state brain scans and speech analysis using Large Language Models (LLMs), we aim to uncover how disruptions in brain communication relate to disorganised thinking. We’re also exploring how these patterns differ across mental health conditions like schizophrenia and depression, and how they relate to people’s lived experiences. This work could help identify the specific brain mechanisms behind disorganisation and guide more targeted, personalised treatments.