Global patterns of network connectivity in schizophrenia: A TMS-EEG network study

Background: Schizophrenia has been long considered a disease of brain dysconnectivity. Until recently, it was difficult to quantify global and emergent features of the brain. To address this, topological measures which apply Network Analysis (based on Graph theory) to neuroimaging data are now able to identify global network connectivity properties of the brain. In the current study, we administer Transcranial Magnetic Stimulation (TMS) pulses to elicit a transient change in network organization (i.e. perturb the network), while electroencephalography (EEG) simultaneously measures network response. This novel approach allows us to examine global and emergent features of the schizophrenia network and assess whether network anomalies relate to clinical aspects of schizophrenia.

Methods: One hundred TMS pulses (perturbations) were applied at a threshold of 80% to the frontal regions of schizophrenia patients (n = 14) and healthy controls (n = 12) while EEG was simultaneously recorded. For each participant, Pearson cross-correlation was applied to averaged EEG data to construct correlation matrices which represent the network response to the TMS stimuli. Global network measures were then extracted from these matrices for comparison between the groups. The relationship between the network metrics and clinical aspects (negative/positive/neurological symptoms) of schizophrenia were then examined.

Results: Following the network perturbation (TMS pulse), schizophrenia patients presented with a more random network response, altered connectivity (number of links) and reduced segregation (clustering co-efficient) when compared to healthy controls. These altered network connectivity metrics were correlated with both clinical and neurological features of schizophrenia.

Discussion: Schizophrenia patients presented with altered patterns of global network connectivity in response to the TMS perturbation of the frontal regions. Moreover, these network abnormalities were related to clinical aspects of schizophrenia. We anticipate that future combined TMS and network analysis studies will contribute to further mapping global network features of schizophrenia and identifying potential network biomarkers.