Intra- and Inter-Frequency Brain Network Structure in Health and Schizophrenia

Empirical studies over the past two decades have supported the hypothesis that schizophrenia is characterized by altered connectivity patterns in functional brain networks. These alterations have been proposed as genetically-mediated diagnostic biomarkers and are thought to underlie altered cognitive functions such as working memory. In this study, we perform an extensive analysis of functional connectivity patterns extracted from MEG data in 14 subjects with schizophrenia and 14 healthy controls during a 2-back working memory task. We investigate uni-, bi- and multivariate properties of sensor time series by computing wavelet entropy of and correlation between time series, and by constructing binary networks of functional connectivity both within and between classical frequency bands (gamma, beta, alpha, and theta). Networks are based on the mutual information between wavelet time series, and estimated for 66 separate time windows. We observed decreases in entropy in prefrontal and lateral sensor time series and increases in connectivity strength in the schizophrenia group in comparison to the healthy controls. We identified an inverse relationship between entropy and strength across both subjects and sensors that varied over frequency bands and was more pronounced in controls than in patients. Brain network topology was altered in schizophrenia specifically in high frequency gamma and beta band networks as well as in the gamma-beta cross-frequency networks. Network topology varied over trials to a greater extent in patients than in controls, suggesting disease-associated alterations in dynamic network properties of brain function. Our results identify signatures of aberrant neurophysiological behavior in schizophrenia across uni-, bi- and multivariate scales and identify cross-frequency network architecture and network dynamics as candidate intermediate phenotypes.

💡 Research Summary

This study presents a comprehensive, multi-scale analysis of functional brain network alterations in schizophrenia (SZ) during a working memory task. Using magnetoencephalography (MEG) data from 14 patients with SZ and 14 healthy controls performing a visual 2-back task, the researchers investigated abnormalities across univariate (signal complexity), bivariate (pairwise connectivity), and multivariate (whole-brain network topology) scales, with a novel focus on both within-frequency and cross-frequency interactions.

The methodology involved decomposing MEG sensor time-series into four classical frequency bands (gamma, beta, alpha, theta) using wavelet transforms. Functional connectivity was quantified as the normalized mutual information between sensor pairs, both within the same frequency band and across different bands, creating 10 distinct network types per subject. For each of the 66 task trials, binary networks were constructed over a wide range of connection densities, and their topology was characterized using 12 graph theory metrics (e.g., efficiency, modularity, clustering). Temporal variability in network organization was assessed by calculating the coefficient of variation (CV) of these metrics across trials. Statistical comparisons employed Functional Data Analysis (FDA) to compare groups across the entire density spectrum.

Key findings revealed that individuals with SZ exhibited significantly impaired working memory performance. At the univariate level, wavelet entropy was decreased in prefrontal and lateral sensors in SZ, indicating reduced signal complexity. At the bivariate level, average connectivity strength was increased in SZ. An inverse relationship between entropy and strength was observed across subjects and sensors, but this relationship was more pronounced in controls, suggesting a decoupling of local processing and long-range communication in SZ.

Multivariate network analysis showed that the topological organization of functional networks was specifically altered in SZ for high-frequency gamma and beta band networks, as well as for the cross-frequency gamma-beta network. Crucially, the study found that network topology varied significantly more from trial to trial in patients than in controls, indicating increased temporal variability and dynamic instability in the functional networks of the SZ brain.

In conclusion, the research identifies signatures of aberrant neurophysiology in SZ across multiple analytical scales. It highlights that dysconnectivity in SZ is not only frequency-specific but also involves disrupted integration across frequencies. Furthermore, it establishes altered network dynamics—increased temporal variability—as a core feature of the disease. The findings propose cross-frequency network architecture and network dynamics as promising candidate intermediate phenotypes for schizophrenia, potentially bridging genetic risk, neurophysiological dysfunction, and cognitive symptoms.