Functional Magnetic Resonance Imaging of Spatiotemporal Brain Dynamics in Patients with Schizophrenia

Abstract

Brain connectivity reflects the functional organization of the brain and serves as a critical indicator for evaluating neuropsychiatric disorders and treatment outcomes. Schizophrenia, marked by impaired functional connectivity, presents challenges in identifying and characterizing complex patterns of abnormal brain networks. Auditory hallucinations, a hallmark symptom, involve false perceptions without external stimuli, and cognitive models suggest that multiple processes contribute to their emergence. This study uses resting-state and task-based functional magnetic resonance imaging (fMRI) to assess connectivity in schizophrenia patients and healthy controls, applying CONN imaging software and energy landscape analysis (ELA) to identify disease-related connectivity patterns. The results of the resting-state fMRI show that abnormal energy landscape characteristics are significantly correlated with the severity of auditory and visual perceptual disturbances in schizophrenia. Task-based fMRI supports these findings, revealing reduced connectivity in patients with auditory verbal hallucinations (AVH+) compared to healthy controls, and in AVH+ patients compared to those without hallucinations (AVH-). The study also incorporates large language models (LLM) to examine the relationship between impaired brain connectivity and hallucination generation, proposing that degradation in the generative process corresponds to altered brain networks. These impairments manifest as altered auditory and visual production. By analyzing connectivity in regions of interest (ROI) that differ between patients and controls, the study offers insight into memory storage and retrieval mechanisms. Advanced electrical engineering techniques, including signal processing, computational modeling, and machine learning, underpin the analysis. ELA models dynamic brain states and transitions, quantifies network stability, and identifies biomarkers linked to perceptual disturbances. Hierarchical clustering further helps to classify connectivity patterns. By integrating engineering methodologies with cognitive neuroscience, the study provides new insights into the relationship between functional connectivity and cognitive dysfunction in schizophrenia. The proposed imaging workflow shows promise for identifying biomarkers associated with specific clinical symptoms and may guide future treatment development and understanding of the underlying mechanisms of the disorder.